v7.3.1 release notes

It's redundant now

.docker/Dockerfile

@@ -3,37 +3,60 @@
 FROM      ubuntu:18.04
 
 RUN apt-get update && apt-get install -y --no-install-recommends \
+  build-essential autoconf automake libtool \
+  libleptonica-dev \
+  zlib1g-dev \
+  libexempi3 \
   ocrmypdf \
+  pngquant \
+  python3-pip \
+  python3-venv \
+  tesseract-ocr \
+  tesseract-ocr-chi-sim \
+  tesseract-ocr-deu \
   tesseract-ocr-eng \
   tesseract-ocr-fra \
-  tesseract-ocr-deu \
-  tesseract-ocr-spa \
   tesseract-ocr-por \
-  tesseract-ocr-chi-sim \
-  python3-venv \
-  python3-pip
+  tesseract-ocr-spa \
+  unpaper \
+  wget
+
 
 ENV LANG=C.UTF-8
 
+# Compile and install jbig2
+# Needs libleptonica-dev, zlib1g-dev
+RUN \
+  mkdir jbig2 \
+  && wget -q https://github.com/agl/jbig2enc/archive/0.29.tar.gz -O - | \
+      tar xz -C jbig2 --strip-components=1 \
+  && cd jbig2 \
+  && ./autogen.sh && ./configure && make && make install \
+  && cd .. \
+  && rm -rf jbig2
+
+RUN apt-get remove -y autoconf automake libtool
+
 RUN python3 -m venv --system-site-packages /appenv
 
 # This installs the latest binary wheel instead of the code in the current
 # folder. Installing from source will fail, apparently because cffi needs
-# build-essentials (gcc) to do a source installation 
+# build-essentials (gcc) to do a source installation
 # (i.e. "pip install ."). It's unclear to me why this is the case.
 RUN . /appenv/bin/activate; \
   pip install --upgrade pip \
-  && pip install ocrmypdf[fitz]
+  && pip install --upgrade ocrmypdf
 
 # Now copy the application in, mainly to get the test suite.
 # Do this now to make the best use of Docker cache.
 COPY . /application
 RUN . /appenv/bin/activate; \
-  pip install -r /application/test_requirements.txt
+  pip install -r /application/requirements/test.txt
 
 # Remove the junk, including the source version of application since it was
 # already installed
 RUN rm -rf /tmp/* /var/tmp/* /root/* /application/ocrmypdf \
+  && apt-get remove -y build-essential \
   && apt-get autoremove -y \
   && apt-get autoclean -y
 

.github/ISSUE_TEMPLATE

@@ -2,11 +2,4 @@ Please include the command line and a test file with your issue report.
 
 If possible, please use a test file that we can include in future test cases (no personal information, no copyrighted material).
 
-If you wish to encrypt a test file so that only the maintainer of OCRmyPDF can view it, you may use:
-
-```bash
-
-gpg --recv-keys 4434eb74c4a35f7f --keyserver pgp.mit.edu
-gpg --output test.pdf.gpg --encrypt --receipient 4434eb74c4a35f7f test.pdf
-
-```
+If you wish to encrypt a test file for the OCRmyPDF maintainer only, see the [Wiki](https://github.com/jbarlow83/OCRmyPDF/wiki).

.gitignore

@@ -1,44 +1,41 @@
 # Development environment
+.bash_history
+.pylintrc
+.pytest_cache/
+.ruffus_history.sqlite
+.venv/
 *.pyc
 *.sublime-*
-venv*/
-pyvenv.cfg
-tasks.py
-.bash_history
-.ruffus_history.sqlite
-.idea/
-.pytest_cache/
 
 # Package building
-*.egg-info/
-.cache/
 .eggs/
+*.egg-info/
 build/
 dist/
 wheelhouse/
 
 # Automatically generated files
-ocrmypdf/lib/_*.py
-ocrmypdf/version.py
 docs/_build/
 docs/_static/
 docs/_templates/
 docs/Makefile
+ocrmypdf/lib/_*.py
 
 # Code coverage
-.coverage
+.coverage*
 htmlcov/
 
 # Testing
-log/
+.ipynb_checkpoints/
+.vscode/
+*.ipynb
+*.profile
 /*.pdf
 /*.qdf
-*.ipynb
-.ipynb_checkpoints/
+/*.png
+/scratch.py
+IDEAS
+log/
 tests/output/
 tests/resources/private/
 tmp/
-pdfbox-app*.jar
-.vscode/
-IDEAS
-_Dockerfile.local

.readthedocs.yml

@@ -0,0 +1,10 @@
+build:
+  image: latest
+
+python:
+  version: 3.6
+
+formats:
+  - pdf
+
+requirements_file: requirements/main.txt

.travis.yml

@@ -1,56 +1,99 @@
-dist: trusty
-language: python
 cache:
-  ccache: true
   pip: true
   directories:
   - $HOME/Library/Caches/Homebrew
 
-env:
-  global:
-    - secure: "hsf6MT+n2x3OiDM2fQyJZdV0/PWYmv81LdVqC6cfnHBE/8N3DloJRqQ7WfO14TxhiK9PEC7MpyCj0lSabUHEO7gSH6Vks6I1asoSkt8S9/bSMlhT4hei+pwVpeGEiU5xHVATNjY+D919VC3IFvc3XmjT74h/2SLhaZ+jhEmDggM="  # HOMEBREW_OCRMYPDF_TOKEN
-
 matrix:
   include:
     - os: linux
+      dist: trusty
       sudo: required
       language: python
-      python: 3.5
-      env: EXTRAS=
+      python: "3.5"
+      env:
+        - DIST=trusty
+      addons: &trusty_apt
+        apt:
+          update: true
+          sources:
+          - sourceline: 'ppa:alex-p/tesseract-ocr'
+          - sourceline: 'ppa:heyarje/libav-11'
+          - sourceline: 'ppa:vshn/ghostscript'
+          packages:
+          - ghostscript
+          - libavcodec56
+          - libavformat56
+          - libavutil54
+          - libexempi3
+          - libffi-dev
+          - pngquant
+          - poppler-utils
+          - qpdf
+          - tesseract-ocr
+          - tesseract-ocr-deu
+          - tesseract-ocr-eng
+          - tesseract-ocr-fra
     - os: linux
+      dist: trusty
       sudo: required
       language: python
-      python: 3.6
-      env: EXTRAS=
+      python: "3.6"
+      env:
+        - DIST=trusty
+      addons: *trusty_apt
     - os: linux
+      dist: xenial
       sudo: required
       language: python
-      python: 3.6
-      env: EXTRAS=[fitz]
+      python: "3.7"
+      env:
+        - DIST=xenial
+      addons:
+        apt:
+          update: true
+          sources:
+          - sourceline: 'ppa:alex-p/tesseract-ocr'
+          packages:
+          - ghostscript
+          - libexempi3
+          - libffi-dev
+          - pngquant
+          - poppler-utils
+          - qpdf
+          - tesseract-ocr
+          - tesseract-ocr-deu
+          - tesseract-ocr-eng
+          - tesseract-ocr-fra
+          - unpaper
     - os: osx
-      osx_image: xcode8
+      osx_image: xcode9.2
       language: generic
-      env: EXTRAS=
-    - os: osx
-      osx_image: xcode8
-      language: generic
-      env: EXTRAS=[fitz]
 
 before_cache:
 - rm -f $HOME/.cache/pip/log/debug.log
 
 before_install: |
+  mkdir -p bin
   if [[ "$TRAVIS_OS_NAME" == "linux" ]]; then
-    bash .travis/linux_before_install.sh
+    pip3 install --upgrade pip
+    pip3 install --upgrade wheel
+    if [[ "$DIST" == "trusty" ]]; then
+      mkdir -p packages
+      wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O packages/unpaper_6.1-1.deb
+      sudo dpkg -i packages/unpaper_6.1-1.deb
+    fi
   elif [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
-    brew update && brew bundle --file=.travis/Brewfile
+    brew update --quiet && brew bundle --file=.travis/Brewfile --quiet
     pip3 install --upgrade pip
     pip3 install wheel
   fi
 
 install:
-- pip3 install ".$EXTRAS"
-- pip3 install -r test_requirements.txt
+- export PATH=$PWD/bin:$PATH
+- pip3 install pycparser  # py3.7 workaround for https://github.com/eliben/pycparser/issues/251
+- pip3 install -r requirements/main.txt
+- pip3 install --no-deps .
+- pip3 install -r requirements/test.txt
 
 script:
 - tesseract --version
@@ -59,7 +102,7 @@ script:
 
 deploy:
   # release for main pypi
-  # 3.6 w/o fitz is considered the build leader and does the deploy, otherwise there is
+  # 3.6 is considered the build leader and does the deploy, otherwise there is
   # a race and all versions will try to deploy
   # OTOH if we ever need separate binary wheels then each version needs its
   # own deploy
@@ -71,33 +114,5 @@ deploy:
   on:
     branch: master
     tags: true
-    condition: $TRAVIS_PYTHON_VERSION == "3.6" && $TRAVIS_OS_NAME == "linux" && $EXTRAS == ""
+    condition: $TRAVIS_PYTHON_VERSION == "3.6" && $TRAVIS_OS_NAME == "linux"
   skip_upload_docs: true
-
-  # test pypi
-- provider: pypi
-  server: https://testpypi.pypi.org/legacy/
-  user: ocrmypdf-travis
-  password:
-    secure: "DTFOmmNL6olA0+yXvp4u9jXZlZeqrJsJ0526jzqf4a3gZ6jnGTq5UI6WzRsslSyoMMfXKtHQebqHM6ogSgCZinyZ3ufHJo8fn9brxbEc2gsiWkbj5o3bGwdWMT1vNNE7XW0VCpw87rZ1EEwjl4FJHFudMlPR1yfU5+uq0k0PACo="
-  distributions: "sdist"
-  on:
-    branch: develop
-    tags: false
-    condition: $TRAVIS_OS_NAME == "osx"
-  skip_upload_docs: true
-
-  # null deploy for osx
-  # we really just want to run after_deploy *after* pypi upload is done, but
-  # after_deploy on runs if a given box deployed
-- provider: script
-  script: /usr/bin/true
-  on:
-    branch: master
-    tags: true
-    condition: $TRAVIS_OS_NAME == "osx"
-
-after_deploy: |
-  if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
-    bash .travis/osx_brew.sh
-  fi

.travis/Brewfile

@@ -1,13 +1,11 @@
 brew 'python'
 
+brew 'exempi'
 brew 'ghostscript'
-brew 'jbig2dec'
+brew 'jbig2enc'
 brew 'leptonica'
-brew 'libffi'
-brew 'libtiff'
-brew 'libxml2'
-brew 'mupdf-tools'
 brew 'openjpeg'
+brew 'pngquant'
 brew 'qpdf'
 brew 'tesseract'
 brew 'unpaper'

.travis/autobrew.py

@@ -1,93 +0,0 @@
-#!/usr/bin/env python3
-# © 2017-18 James R. Barlow: github.com/jbarlow83
-
-from string import Template
-from subprocess import run, PIPE
-import re
-
-recipe_template = Template("""
-class Ocrmypdf < Formula
-  include Language::Python::Virtualenv
-
-  desc "Adds an OCR text layer to scanned PDF files"
-  homepage "https://github.com/jbarlow83/OCRmyPDF"
-  ${ocrmypdf_url}
-  ${ocrmypdf_sha256}
-
-  depends_on "pkg-config" => :build
-  depends_on "mupdf-tools" => :build  # statically links libmupdf.a
-  depends_on "freetype"
-  depends_on "ghostscript"
-  depends_on "jpeg"
-  depends_on "libpng"
-  depends_on "python"
-  depends_on "qpdf"
-  depends_on "tesseract"
-  depends_on "unpaper"
-
-${resources}
-  def install
-    venv = virtualenv_create(libexec, "python3")
-
-    resource("Pillow").stage do
-      inreplace "setup.py" do |s|
-        sdkprefix = MacOS::CLT.installed? ? "" : MacOS.sdk_path
-        s.gsub! "openjpeg.h", "probably_not_a_header_called_this_eh.h"
-        s.gsub! "ZLIB_ROOT = None", "ZLIB_ROOT = ('#{sdkprefix}/usr/lib', '#{sdkprefix}/usr/include')"
-        s.gsub! "JPEG_ROOT = None", "JPEG_ROOT = ('#{Formula["jpeg"].opt_prefix}/lib', '#{Formula["jpeg"].opt_prefix}/include')"
-        s.gsub! "FREETYPE_ROOT = None", "FREETYPE_ROOT = ('#{Formula["freetype"].opt_prefix}/lib', '#{Formula["freetype"].opt_prefix}/include')"
-      end
-
-      # avoid triggering "helpful" distutils code that doesn't recognize Xcode 7 .tbd stubs
-      ENV.append "CFLAGS", "-I#{MacOS.sdk_path}/System/Library/Frameworks/Tk.framework/Versions/8.5/Headers" unless MacOS::CLT.installed?
-      venv.pip_install Pathname.pwd
-    end
-
-    res = resources.map(&:name).to_set - ["Pillow"]
-
-    res.each do |r|
-      venv.pip_install resource(r)
-    end
-
-    venv.pip_install_and_link buildpath
-  end
-
-  test do
-    # Since we use Python 3, we require a UTF-8 locale
-    ENV["LC_ALL"] = "en_US.UTF-8"
-
-    system "#{bin}/ocrmypdf", "-f", "-q", "--deskew",
-                              test_fixtures("test.pdf"), "ocr.pdf"
-    assert_predicate testpath/"ocr.pdf", :exist?
-  end
-end
-""")
-
-def main():
-    p = run(['poet', '--single', 'ocrmypdf'],
-            encoding='utf-8', stdout=PIPE, check=True)
-
-    ocrmypdf_lines = p.stdout.splitlines()
-    ocrmypdf_url = ocrmypdf_lines[1].strip()
-    ocrmypdf_sha256 = ocrmypdf_lines[2].strip()
-
-    ocrmypdf_version = re.search(
-        r'ocrmypdf-(.+)\.tar.*', ocrmypdf_url).group(1)
-    print(f"Autobrewing {ocrmypdf_version}")
-
-    p = run(['poet', '--resources', 'ocrmypdf'],
-            encoding='utf-8', stdout=PIPE, check=True)
-
-    poet_resources = p.stdout
-
-    # Remove the duplicate "ocrmypdf" resource block
-    all_resources = poet_resources.split('resource')
-    kept_resources = [block for block in all_resources if 'ocrmypdf' not in block]
-    resources = 'resource'.join(kept_resources)
-
-    with open('ocrmypdf.rb', 'w') as out:
-        out.write(recipe_template.substitute(**locals()))
-
-
-if __name__ == '__main__':
-    main()

.travis/linux_before_install.sh

@@ -1,42 +0,0 @@
-#!/bin/bash
-# © 2017 James R. Barlow: github.com/jbarlow83
-set -euo pipefail
-set -x
-
-sudo add-apt-repository ppa:vshn/ghostscript -y
-sudo add-apt-repository ppa:heyarje/libav-11 -y
-sudo apt-get update -qq
-sudo apt-get install -y \
-	ghostscript \
-	poppler-utils \
-	libavformat56 \
-	libavcodec56 \
-	libavutil54 \
-	libffi-dev
-
-sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
-
-sudo apt-get update
-sudo apt-get autoremove -y
-sudo apt-get install -y --no-install-recommends \
-		tesseract-ocr \
-		tesseract-ocr-eng \
-		tesseract-ocr-fra \
-		tesseract-ocr-deu
-
-pip install --upgrade pip
-mkdir -p packages
-wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O packages/unpaper_6.1-1.deb
-sudo dpkg -i packages/unpaper_6.1-1.deb
-
-if [ ! -f /usr/local/bin/qpdf ]; then
-	export QPDF_RELEASE='https://github.com/qpdf/qpdf/releases/download/release-qpdf-8.0.2/qpdf-8.0.2.tar.gz'
-	mkdir qpdf
-    wget -q $QPDF_RELEASE -O - | tar xz -C qpdf --strip-components=1
-    cd qpdf/
-	export PATH="/usr/local/opt/ccache/libexec:$PATH"
-    ./configure --prefix=/usr
-    make -j 2
-    sudo make install
-    cd ..
-fi

.travis/osx_brew.sh

@@ -1,23 +0,0 @@
-#!/bin/bash
-# © 2017 James R. Barlow: github.com/jbarlow83
-set -uo pipefail
-set -x
-
-pip3 install homebrew-pypi-poet
-python3 .travis/autobrew.py
-cat ocrmypdf.rb
-
-# brew audit crashes Travis
-#brew audit ocrmypdf.rb
-
-# Important: disable debug output so token is hidden
-set +x
-git clone https://$HOMEBREW_OCRMYPDF_TOKEN@github.com/jbarlow83/homebrew-ocrmypdf.git
-set -x
-
-pushd homebrew-ocrmypdf
-cp ../ocrmypdf.rb Formula/ocrmypdf.rb
-git add Formula/ocrmypdf.rb
-git commit -m "homebrew-ocrmypdf: automatic release $TRAVIS_BUILD_NUMBER $TRAVIS_TAG"
-git push origin master
-popd

MANIFEST.in

@@ -1,7 +1,5 @@
 # requirements
-include requirements.txt
-include test_requirements.txt
-include dev_requirements.txt
+recursive-include requirements *
 
 # git
 include .git_archival.txt
@@ -12,21 +10,20 @@ recursive-include .docker *
 
 # tests
 include .coveragerc
+recursive-include tests *.bin
 recursive-include tests *.jpg
+recursive-include tests *.jsonl
 recursive-include tests *.png
 recursive-include tests *.pdf
 recursive-include tests *.py
 recursive-include tests *.rst
 recursive-include tests *.txt
-recursive-include tests/cache *
-recursive-exclude tests/output *
-recursive-exclude tests/output_pageinfo *
 recursive-exclude tests/resources/private *
 
 # documentation
 include LICENSE
 include *.rst
-recursive-exclude .github * 
+recursive-exclude .github *
 recursive-include docs *.py
 recursive-include docs *.rst
 recursive-include docs *.svg
@@ -42,8 +39,5 @@ exclude .travis*
 
 
 # code
-recursive-include src/ocrmypdf *.py
-exclude ocrmypdf/lib/_leptonica.py
-
-
-
+exclude src/ocrmypdf/lib/_leptonica.py
+exclude scratch.py

README.md

@@ -0,0 +1,142 @@
+OCRmyPDF
+========
+
+[![Travis build status][travis]](https://travis-ci.org/jbarlow83/OCRmyPDF) [![PyPI version][pypi]](https://pypi.org/project/ocrmypdf/) ![Homebrew version][homebrew] ![ReadTheDocs][docs]
+
+[travis]: https://travis-ci.org/jbarlow83/OCRmyPDF.svg?branch=master "Travis build status"
+
+[pypi]: https://img.shields.io/pypi/v/ocrmypdf.svg "PyPI version"
+
+[homebrew]: https://img.shields.io/homebrew/v/ocrmypdf.svg "Homebrew version"
+
+[docs]: https://readthedocs.org/projects/ocrmypdf/badge/?version=latest "RTD"
+
+OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to be searched or copy-pasted.
+
+```bash
+ocrmypdf                      # it's a scriptable command line program
+   -l eng+fra                 # it supports multiple languages
+   --rotate-pages             # it can fix pages that are misrotated
+   --deskew                   # it can deskew crooked PDFs!
+   --title "My PDF"           # it can change output metadata
+   --jobs 4                   # it uses multiple cores by default
+   --output-type pdfa         # it produces PDF/A by default
+   input_scanned.pdf          # takes PDF input (or images)
+   output_searchable.pdf      # produces validated PDF output
+```
+
+Main features
+-------------
+
+-   Generates a searchable [PDF/A](https://en.wikipedia.org/?title=PDF/A) file from a regular PDF
+-   Places OCR text accurately below the image to ease copy / paste
+-   Keeps the exact resolution of the original embedded images
+-   When possible, inserts OCR information as a "lossless" operation without disrupting any other content
+-   Optimizes PDF images, often producing files smaller than the input file
+-   If requested deskews and/or cleans the image before performing OCR
+-   Validates input and output files
+-   Distributes work across all available CPU cores
+-   Uses [Tesseract OCR](https://github.com/tesseract-ocr/tesseract) engine
+-   Supports more than [100 languages](https://github.com/tesseract-ocr/tessdata) recognized by Tesseract
+-   Battle-tested on thousands of PDFs, a test suite and continuous integration
+
+For details: please consult the [documentation](https://ocrmypdf.readthedocs.io/en/latest/).
+
+Motivation
+----------
+
+I searched the web for a free command line tool to OCR PDF files on Linux/UNIX: I found many, but none of them were really satisfying.
+
+-   Either they produced PDF files with misplaced text under the image (making copy/paste impossible)
+-   Or they did not handle accents and multilingual characters
+-   Or they changed the resolution of the embedded images
+-   Or they generated ridiculously large PDF files
+-   Or they crashed when trying to OCR
+-   Or they did not produce valid PDF files
+-   On top of that none of them produced PDF/A files (format dedicated for long time storage)
+
+...so I decided to develop my own tool.
+
+Installation
+------------
+
+Linux, UNIX, and macOS are supported. Windows is not directly supported but there is a Docker image available that runs on Windows.
+
+Users of Debian 9 or later or Ubuntu 16.10 or later may simply
+
+```bash
+apt-get install ocrmypdf
+```
+
+and users of Fedora 29 or later may simply
+
+```bash
+dnf install ocrmypdf
+```
+
+and macOS users with Homebrew may simply
+
+```bash
+brew install ocrmypdf
+```
+
+For everyone else, [see our documentation](https://ocrmypdf.readthedocs.io/en/latest/installation.html) for installation steps.
+
+Languages
+---------
+
+OCRmyPDF uses Tesseract for OCR, and relies on its language packs. For Linux users, you can often find packages that provide language packs:
+
+```bash
+# Display a list of all Tesseract language packs
+apt-cache search tesseract-ocr
+
+# Debian/Ubuntu users
+apt-get install tesseract-ocr-chi-sim  # Example: Install Chinese Simplified language back
+```
+
+You can then pass the `-l LANG` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple languages can be requested.
+
+Documentation and support
+-------------------------
+
+Once ocrmypdf is installed, the built-in help which explains the command syntax and options can be accessed via:
+
+```bash
+ocrmypdf --help
+```
+
+Our [documentation is served on Read the Docs](https://ocrmypdf.readthedocs.io/en/latest/index.html).
+
+If you detect an issue, please:
+
+-   Check whether your issue is already known
+-   If no problem report exists on github, please create one here: <https://github.com/jbarlow83/OCRmyPDF/issues>
+-   Describe your problem thoroughly
+-   Append the console output of the script when running the debug mode (`-v 1` option)
+-   If possible provide your input PDF file as well as the content of the temporary folder (using a file sharing service like Dropbox)
+
+Requirements
+------------
+
+Runs on CPython 3.5, 3.6 and 3.7. Requires external program installations of Ghostscript, Tesseract OCR, QPDF, and Leptonica. ocrmypdf is pure Python, but uses CFFI to portably generate library bindings.
+
+Press & Media
+-------------
+
+-   [c't 1-2014, page 59](http://heise.de/-2279695): Detailed presentation of OCRmyPDF v1.0 in the leading German IT magazine c't
+-   [heise Open Source, 09/2014: Texterkennung mit OCRmyPDF](http://heise.de/-2356670)
+
+License
+-------
+
+The OCRmyPDF software is licensed under the GNU GPLv3. Certain files are covered by other licenses, as noted in their source files.
+
+The license for each test file varies, and is noted in tests/resources/README.rst. The documentation is licensed under Creative Commons Attribution-ShareAlike 4.0 (CC-BY-SA 4.0).
+
+OCRmyPDF versions prior to 6.0 were licensed under the MIT License.
+
+Disclaimer
+----------
+
+The software is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

README.rst

@@ -1,155 +0,0 @@
-OCRmyPDF
-========
-
-.. image:: https://travis-ci.org/jbarlow83/OCRmyPDF.svg?branch=master
-    :target: https://travis-ci.org/jbarlow83/OCRmyPDF
-
-.. image:: https://img.shields.io/pypi/v/ocrmypdf.svg   
-    :target: https://pypi.org/project/ocrmypdf/
-
-.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg   
-    :alt: homebrew
-    :target: http://brewformulas.org/Ocrmypdf
-
-
-OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to
-be searched or copy-pasted.
-
-.. code-block:: bash
-
-   ocrmypdf                      # it's a scriptable command line program
-      -l eng+fra                 # it supports multiple languages
-      --rotate-pages             # it can fix pages that are misrotated
-      --deskew                   # it can deskew crooked PDFs!
-      --title "My PDF"           # it can change output metadata
-      --jobs 4                   # it uses multiple cores by default
-      --output-type pdfa         # it produces PDF/A by default
-      input_scanned.pdf          # takes PDF input (or images)
-      output_searchable.pdf      # produces validated PDF output
-
-
-Main features
--------------
-
--  Generates a searchable
-   `PDF/A <https://en.wikipedia.org/?title=PDF/A>`_ file from a regular PDF
--  Places OCR text accurately below the image to ease copy / paste
--  Keeps the exact resolution of the original embedded images
--  When possible, inserts OCR information as a "lossless" operation without rendering vector information
--  Keeps file size about the same
--  If requested deskews and/or cleans the image before performing OCR
--  Validates input and output files
--  Provides debug mode to enable easy verification of the OCR results
--  Processes pages in parallel when more than one CPU core is
-   available
--  Uses `Tesseract OCR <https://github.com/tesseract-ocr/tesseract>`_ engine
--  Supports more than `100 languages <https://github.com/tesseract-ocr/tessdata>`_ recognized by Tesseract
--  Battle-tested on thousands of PDFs, a test suite and continuous integration
-
-For details: please consult the `documentation <https://ocrmypdf.readthedocs.io/en/latest/>`_.
-
-Motivation
-----------
-
-I searched the web for a free command line tool to OCR PDF files on
-Linux/UNIX: I found many, but none of them were really satisfying.
-
--  Either they produced PDF files with misplaced text under the image (making copy/paste impossible) 
--  Or they did not handle accents and multilingual characters
--  Or they changed the resolution of the embedded images
--  Or they generated ridiculously large PDF files
--  Or they crashed when trying to OCR some of my PDF files
--  Or they did not produce valid PDF files (even though they were readable with my current PDF reader)
--  On top of that none of them produced PDF/A files (format dedicated for long time storage)
-
-...so I decided to develop my own tool (using various existing scripts
-as an inspiration). 
-
-Installation
-------------
-
-Linux, UNIX, and macOS are supported. Windows is not directly supported but there is a Docker image available that runs on Windows.
-
-Users of Debian 9 or later or Ubuntu 16.10 or later may simply
-
-.. code-block:: bash
-
-   apt-get install ocrmypdf
-
-and macOS users with Homebrew may simply
-
-.. code-block:: bash
-
-   brew install ocrmypdf
-
-For everyone else, `see our documentation <https://ocrmypdf.readthedocs.io/en/latest/installation.html>`_ for installation steps.
-
-Languages
----------
-
-OCRmyPDF uses Tesseract for OCR, and relies on its language packs. For Linux users,
-you can often find packages that provide language packs:
-
-.. code-block:: bash
-
-   # Display a list of all Tesseract language packs
-   apt-cache search tesseract-ocr
-
-   # Debian/Ubuntu users
-   apt-get install tesseract-ocr-chi-sim  # Example: Install Chinese Simplified language back
-   
-You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple
-languages can be requested.
-
-Documentation and support
--------------------------
-
-Once ocrmypdf is installed, the built-in help which explains the command syntax and options can be accessed via:
-
-.. code-block:: bash
-
-   ocrmypdf --help
-
-Our `documentation is served on Read the Docs <https://ocrmypdf.readthedocs.io/en/latest/index.html>`_.
-
-If you detect an issue, please:
-
--  Check whether your issue is already known
--  If no problem report exists on github, please create one here:
-   https://github.com/jbarlow83/OCRmyPDF/issues
--  Describe your problem thoroughly
--  Append the console output of the script when running the debug mode
-   (``-v 1`` option)
--  If possible provide your input PDF file as well as the content of the
-   temporary folder (using a file sharing service like Dropbox)
-
-Requirements
-------------
-
-Runs on CPython 3.6, and requires external program installations of Ghostscript, Tesseract OCR, QPDF, and Leptonica. ocrmypdf is pure Python, but uses CFFI to portably generate library bindings.
-
-Python 3.5 is also supported.
-
-Press & Media
--------------
-
--  `c't 1-2014, page 59 <http://heise.de/-2279695>`_:
-   Detailed presentation of OCRmyPDF v1.0 in the leading German IT
-   magazine c't
--  `heise Open Source, 09/2014: Texterkennung mit
-   OCRmyPDF <http://heise.de/-2356670>`_
-
-License
--------
-
-The OCRmyPDF software is licensed under the GNU GPLv3. Certain files are covered by other licenses, as noted in their source files. 
-
-The license for each test file varies, and is noted in tests/resources/README.rst. The documentation is licensed under Creative Commons Attribution-ShareAlike 4.0 (CC-BY-SA 4.0).
-
-OCRmyPDF versions prior to 6.0 were licensed under the MIT License.
-
-Disclaimer
-----------
-
-The software is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
-CONDITIONS OF ANY KIND, either express or implied.

docs/advanced.rst

@@ -9,17 +9,19 @@ OCRmyPDF provides many features to control the behavior of the OCR engine, Tesse
 When OCR is skipped
 """""""""""""""""""
 
-If a page in a PDF seems to have text, by default OCRmyPDF will exit without modifying the PDF. This is to ensure that PDFs that were previously OCRed or were "born digital" rather than scanned are not processed. 
+If a page in a PDF seems to have text, by default OCRmyPDF will exit without modifying the PDF. This is to ensure that PDFs that were previously OCRed or were "born digital" rather than scanned are not processed.
 
 If ``--skip-text`` is issued, then no OCR will be performed on pages that already have text. The page will be copied to the output. This may be useful for documents that contain both "born digital" and scanned content, or to use OCRmyPDF to normalize and convert to PDF/A regardless of their contents.
 
-If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information.
+If ``--redo-ocr`` is issued, then a detailed text analysis is performed. Text is categorized as either visible or invisible. Invisible text (OCR) is stripped out. Then an image of each page is created with visible text masked out. The page image is sent for OCR, and any additional text is inserted as OCR. If a file contains a mix of text and bitmap images that contain text, OCRmyPDF will locate the additional text in images without disrupting the existing text.
+
+If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information. Any forms and vector graphics will be rasterized as well.
 
 
 Time and image size limits
 """"""""""""""""""""""""""
 
-By default, OCRmyPDF permits tesseract to run for three minutes (180 seconds) per page. This is usually more than enough time to find all text on a reasonably sized page with modern hardware. 
+By default, OCRmyPDF permits tesseract to run for three minutes (180 seconds) per page. This is usually more than enough time to find all text on a reasonably sized page with modern hardware.
 
 If a page is skipped, it will be inserted without OCR. If preprocessing was requested, the preprocessed image layer will be inserted.
 
@@ -27,32 +29,32 @@ If you want to adjust the amount of time spent on OCR, change ``--tesseract-time
 
 .. code-block:: bash
 
-	# Allow 300 seconds for OCR; skip any page larger than 50 megapixels
-	ocrmypdf --tesseract-timeout 300 --skip-big 50 bigfile.pdf output.pdf
+    # Allow 300 seconds for OCR; skip any page larger than 50 megapixels
+    ocrmypdf --tesseract-timeout 300 --skip-big 50 bigfile.pdf output.pdf
 
 Overriding default tesseract
 """"""""""""""""""""""""""""
 
-OCRmyPDF checks the system ``PATH`` for the ``tesseract`` binary.  
+OCRmyPDF checks the system ``PATH`` for the ``tesseract`` binary.
 
 Some relevant environment variables that influence Tesseract's behavior include:
 
 .. envvar:: TESSDATA_PREFIX
 
-	Overrides the path to Tesseract's data files. This can allow simultaneous installation of the "best" and "fast" training data sets. OCRmyPDF does not manage this environment variable.
+    Overrides the path to Tesseract's data files. This can allow simultaneous installation of the "best" and "fast" training data sets. OCRmyPDF does not manage this environment variable.
 
 .. envvar:: OMP_THREAD_LIMIT
 
-	Controls the number of threads Tesseract will use. OCRmyPDF will manage this environment if it is not already set. (Currently, it will set it to 1 because this gives the best results in testing.)
+    Controls the number of threads Tesseract will use. OCRmyPDF will manage this environment if it is not already set. (Currently, it will set it to 1 because this gives the best results in testing.)
 
 For example, if you are testing tesseract 4.00 and don't wish to use an existing tesseract 3.04 installation, you can launch OCRmyPDF as follows:
 
 .. code-block:: bash
 
-	env \
-		PATH=/home/user/src/tesseract4/api:$PATH \
-		TESSDATA_PREFIX=/home/user/src/tesseract4 \
-		ocrmypdf --tesseract-oem 2 input.pdf output.pdf
+    env \
+        PATH=/home/user/src/tesseract4/api:$PATH \
+        TESSDATA_PREFIX=/home/user/src/tesseract4 \
+        ocrmypdf --tesseract-oem 2 input.pdf output.pdf
 
 In this example ``TESSDATA_PREFIX`` directs Tesseract 4.0 to use LSTM training data. ``--tesseract-oem 1`` requests tesseract 4.0's new LSTM engine. (Tesseract 4.0 only.)
 
@@ -80,19 +82,19 @@ Create a file named "no-dict.cfg" with these contents:
 
 ::
 
-	load_system_dawg 0
-	language_model_penalty_non_dict_word 0
-	language_model_penalty_non_freq_dict_word 0
+    load_system_dawg 0
+    language_model_penalty_non_dict_word 0
+    language_model_penalty_non_freq_dict_word 0
 
 then run ocrmypdf as follows (along with any other desired arguments):
 
 .. code-block:: bash
 
-	ocrmypdf --tesseract-config no-dict.cfg input.pdf output.pdf
+    ocrmypdf --tesseract-config no-dict.cfg input.pdf output.pdf
 
 .. warning::
 
-	Some combinations of control parameters will break Tesseract or break assumptions that OCRmyPDF makes about Tesseract's output.
+    Some combinations of control parameters will break Tesseract or break assumptions that OCRmyPDF makes about Tesseract's output.
 
 
 Changing the PDF renderer
@@ -105,7 +107,7 @@ rendering
   Creating a new PDF from other data (such as an existing PDF).
 
 
-OCRmyPDF has three PDF renderers: ``sandwich``, ``hocr``, ``tesseract``. The renderer may be selected using ``--pdf-renderer``. The default is ``auto`` which lets OCRmyPDF select the renderer to use. Currently, ``auto`` selects ``sandwich`` for Tesseract 3.05.01 or newer, or ``hocr`` for older versions of Tesseract.
+OCRmyPDF has these PDF renderers: ``sandwich`` and ``hocr``. The renderer may be selected using ``--pdf-renderer``. The default is ``auto`` which lets OCRmyPDF select the renderer to use. Currently, ``auto`` selects ``sandwich`` for Tesseract 3.05.01 or newer, or ``hocr`` for older versions of Tesseract.
 
 The ``sandwich`` renderer
 """""""""""""""""""""""""
@@ -116,7 +118,7 @@ Currently this is the best renderer for most uses, however it is implemented in
 
 When image preprocessing features like ``--deskew`` are used, the original PDF will be rendered as a full page and the OCR layer will be placed on top.
 
-This renderer requires Tesseract 3.05.01 or newer.
+If a PDF created with this renderer using Tesseract versions older than 3.05.00 is then passed through Ghostscript's pdfwrite feature, the OCR text *may* be corrupted. The ``--output-type=pdfa`` argument will produce a warning in this situation.  For this reason, OCRmyPDF automatically selects the ``hocr`` for older Tesseract versions.
 
 The ``hocr`` renderer
 """""""""""""""""""""
@@ -132,8 +134,60 @@ This works in all versions of Tesseract.
 The ``tesseract`` renderer
 """"""""""""""""""""""""""
 
-The ``tesseract`` renderer creates a PDF with the image and text layers precomposed, meaning that it always transcodes, loses image quality and rasterizes any vector objects. It does a better job on non-Latin text and document structure than ``hocr``.
+The ``tesseract`` renderer was removed. OCRmyPDF's new approach to text layer grafting makes it functionally equivalent to ``sandwich``.
 
-If a PDF created with this renderer using Tesseract versions older than 3.05.00 is then passed through Ghostscript's pdfwrite feature, the OCR text *may* be corrupted. The ``--output-type=pdfa`` argument will produce a warning in this situation.
+Return code policy
+------------------
 
-*This renderer is deprecated and will be removed whenever support for older versions of Tesseract is dropped.*
+OCRmyPDF writes all messages to ``stderr``.  ``stdout`` is reserved for piping
+output files.  ``stdin`` is reserved for piping input files.
+
+The return codes generated by the OCRmyPDF are considered part of the stable
+user interface.  They may be imported from ``ocrmypdf.exceptions``.
+
+.. list-table:: Return codes
+    :widths: 5 35 60
+    :header-rows: 1
+
+    *	- Code
+        - Name
+        - Interpretation
+    *	- 0
+        - ``ExitCode.ok``
+        - Everything worked as expected.
+    *	- 1
+        - ``ExitCode.bad_args``
+        - Invalid arguments, exited with an error.
+    *	- 2
+        - ``ExitCode.input_file``
+        - The input file does not seem to be a valid PDF.
+    *	- 3
+        - ``ExitCode.missing_dependency``
+        - An external program required by OCRmyPDF is missing.
+    *	- 4
+        - ``ExitCode.invalid_output_pdf``
+        - An output file was created, but it does not seem to be a valid PDF. The file will be available.
+    *	- 5
+        - ``ExitCode.file_access_error``
+        - The user running OCRmyPDF does not have sufficient permissions to read the input file and write the output file.
+    *	- 6
+        - ``ExitCode.already_done_ocr``
+        - The file already appears to contain text so it may not need OCR. See output message.
+    *	- 7
+        - ``ExitCode.child_process_error``
+        - An error occurred in an external program (child process) and OCRmyPDF cannot continue.
+    *	- 8
+        - ``ExitCode.encrypted_pdf``
+        - The input PDF is encrypted. OCRmyPDF does not read encrypted PDFs. Use another program such as ``qpdf`` to remove encryption.
+    *	- 9
+        - ``ExitCode.invalid_config``
+        - A custom configuration file was forwarded to Tesseract using ``--tesseract-config``, and Tesseract rejected this file.
+    *   - 10
+        - ``ExitCode.pdfa_conversion_failed``
+        - A valid PDF was created, PDF/A conversion failed. The file will be available.
+    *	- 15
+        - ``ExitCode.other_error``
+        - Some other error occurred.
+    *	- 130
+        - ``ExitCode.ctrl_c``
+        - The program was interrupted by pressing Ctrl+C.

docs/conf.py

@@ -60,6 +60,24 @@ author = 'James R. Barlow'
 # built documents.
 #
 # The short X.Y version.
+
+import os
+on_rtd = os.environ.get('READTHEDOCS') == 'True'
+
+if on_rtd:
+    # Help ReadTheDocs avoid having to install any binary extension modules
+    import sys
+    from unittest.mock import MagicMock
+
+    class Mock(MagicMock):
+        @classmethod
+        def __getattr__(cls, name):
+            return MagicMock()
+
+    MOCK_MODULES = ['pikepdf', 'libxmp', 'libxmp.utils', 'ocrmypdf.leptonica']
+    sys.modules.update((mod_name, Mock()) for mod_name in MOCK_MODULES)
+
+
 from ocrmypdf import __version__ as OCRMYPDF_VERSION
 
 _version_parts = OCRMYPDF_VERSION.split('.')

docs/cookbook.rst

@@ -57,20 +57,18 @@ You can increase (decrease) the parameter ``--rotate-pages-threshold`` to make p
 
 If the page is "just a little off horizontal", like a crooked picture, then you want ``--deskew``. ``--rotate-pages`` is for when the cardinal angle is wrong.
 
-
 OCR languages other than English
 """"""""""""""""""""""""""""""""
 
-By default OCRmyPDF assumes the document is English. 
+By default OCRmyPDF assumes the document is English.
 
 .. code-block:: bash
 
-    ocrmypdf -l fre LeParisien.pdf LeParisien.pdf
-    ocrmypdf -l eng+fre Bilingual-English-French.pdf Bilingual-English-French.pdf
+    ocrmypdf -l fra LeParisien.pdf LeParisien.pdf
+    ocrmypdf -l eng+fra Bilingual-English-French.pdf Bilingual-English-French.pdf
 
 Language packs must be installed for all languages specified. See :ref:`Installing additional language packs <lang-packs>`.
 
-
 Produce PDF and text file containing OCR text
 """""""""""""""""""""""""""""""""""""""""""""
 
@@ -94,7 +92,7 @@ If you are starting with images, you can just use Tesseract 3.04 or later direct
     # When there are multiple images
     tesseract text-file-containing-list-of-image-filenames.txt output-prefix pdf
 
-Tesseract's PDF output is quite good – OCRmyPDF uses it by internally by default. However, OCRmyPDF has many features not available in Tesseract like like image processing, metadata control, and PDF/A generation.
+Tesseract's PDF output is quite good – OCRmyPDF uses it internally by default. However, OCRmyPDF has many features not available in Tesseract like like image processing, metadata control, and PDF/A generation.
 
 Use a program like `img2pdf <https://gitlab.mister-muffin.de/josch/img2pdf>`_ to convert your images to PDFs, and then pipe the results to run ocrmypdf.  The `-` tells ocrmypdf to read standard input.
 
@@ -116,7 +114,6 @@ If you have multiple images, you must use ``img2pdf`` to convert the images to P
 
     ImageMagick ``convert`` can also convert a group of images to PDF, but in the author's experience it takes a long time, transcodes unnecessarily and gives poor results.
 
-
 Image processing
 ----------------
 
@@ -127,11 +124,13 @@ OCRmyPDF perform some image processing on each page of a PDF, if desired.  The s
 * ``--remove-background`` attempts to detect and remove a noisy background from grayscale or color images.  Monochrome images are ignored. This should not be used on documents that contain color photos as it may remove them.
 
 * ``--deskew`` will correct pages were scanned at a skewed angle by rotating them back into place.  Skew determination and correction is performed using `Postl's variance of line sums <http://www.leptonica.com/skew-measurement.html>`_ algorithm as implemented in `Leptonica <http://www.leptonica.com/index.html>`_.
-  
+
 * ``--clean`` uses `unpaper <https://www.flameeyes.eu/projects/unpaper>`_ to clean up pages before OCR, but does not alter the final output.  This makes it less likely that OCR will try to find text in background noise.
 
 * ``--clean-final`` uses unpaper to clean up pages before OCR and inserts the page into the final output.  You will want to review each page to ensure that unpaper did not remove something important.
 
+* ``-mask-barcodes`` will "cover up" any barcodes detected in the image of a page. Barcodes are known to confuse Tesseract OCR and interfere with the recognition of text on the same baseline as a barcode. The output file will contain the unaltered image of the barcode.
+
 .. note::
 
     In many cases image processing will rasterize PDF pages as images, potentially losing quality.
@@ -140,7 +139,6 @@ OCRmyPDF perform some image processing on each page of a PDF, if desired.  The s
 
     ``--clean-final`` and ``-remove-background`` may leave undesirable visual artifacts in some images where their algorithms have shortcomings. Files should be visually reviewed after using these options.
 
-
 OCR and correct document skew (crooked scan)
 """"""""""""""""""""""""""""""""""""""""""""
 
@@ -167,28 +165,22 @@ If you set ``--tesseract-timeout 0`` OCRmyPDF will apply its image processing wi
     ocrmypdf --tesseract-timeout=0 --remove-background input.pdf output.pdf
 
 
-Redo OCR
-""""""""
+Redo existing OCR
+"""""""""""""""""
 
-To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--force-ocr`` argument. Normally, OCRmyPDF does not modify files that already appear to contain OCR text.
+To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--redo-ocr`` argument. (Normally, OCRmyPDF will exit with an error if asked to modify a file with OCR.)
+
+This may be helpful for users who want to take advantage of accuracy improvements in Tesseract 4.0 for files they previously OCRed with an earlier version of Tesseract and OCRmyPDF.
 
 .. code-block:: bash
 
-    ocrmypdf --force-ocr input.pdf output.pdf
+    ocrmypdf --redo-ocr input.pdf output.pdf
 
-Note that the method above will force rasterization of all pages, potentially reducing quality or losing vector content. 
+This method will replace OCR without rasterizing, reducing quality or removing vector content. If a file contains a mix of pure digital text and OCR, digital text will be ignored and OCR will be replaced. As such this mode is incompatible with image processing options, since they alter the appearance of the file.
 
-To ensure quality is preserved, one could extract all of the images and rebuild the PDF for a lossless transformation. This recipe does not work when PDFs contain multiple images per page, as many do in practice. It will also lose any page rotation information.
-
-.. code-block:: bash
-
-    pdfimages -all old-ocr.pdf prefix  # extract all images
-    img2pdf -o temp.pdf prefix*        # construct new PDF from the images
-    # review the new PDF to ensure it visually matches the old one
-    ocrmypdf --output-type pdf temp.pdf new-ocr.pdf
-
-``--output-type pdf`` is used here to avoid using Ghostscript which will also rasterize images.
+In some cases, existing OCR cannot be detected or replaced. Files produced by OCRmyPDF v2.2 or earlier, for example, are internally represented as having visible text with an opaque image drawn on top. This situation cannot be detected.
 
+If ``--redo-ocr`` does not work, you can use ``--force-ocr``, which will force rasterization of all pages, potentially reducing quality or losing vector content.
 
 Improving OCR quality
 ---------------------
@@ -199,3 +191,19 @@ Rotating pages and deskewing helps to ensure that the page orientation is correc
 
 OCR quality will suffer if the resolution of input images is not correct (since the range of pixel sizes that will be checked for possible fonts will also be incorrect).
 
+PDF optimization
+----------------
+
+By default OCRmyPDF will attempt to perform lossless optimizations on the images inside PDFs after OCR is complete. Optimization is performed even if no OCR text is found.
+
+The ``--optimize N`` (short form ``-O``) argument controls optimization, where ``N`` ranges from 0 to 3. ``--optimize 0`` disables optimizations. ``1`` enables lossless optimizations that can be performed safely with no quality loss. ``2`` enables lossy optimizations such as image color quantizations. ``3`` enables more aggressive optimizations and targets a lower JPEG quality.
+
+Optimization is improved when a JBIG2 encoder is available and when ``pngquant`` is installed. If either of these components are missing, then some types of images will not be optimized.
+
+Currently optimization attempts to find more efficient encodings for images. The types of optimization available may expand over time. By default, OCRmyPDF compresses data streams inside PDFs, and will change inefficient encodings to more modern versions. A program like ``qpdf`` can be used to change encodings, e.g. to inspect the internals fo a PDF.
+
+.. code-block:: bash
+
+    ocrmypdf --optimize 3 in.pdf out.pdf  # Make it small
+
+Some users may consider enabling lossy JBIG2. See: :ref:`jbig2-lossy`.

docs/index.rst

@@ -18,6 +18,7 @@ PDF is the best format for storing and exchanging scanned documents.  Unfortunat
    release_notes
    installation
    languages
+   jbig2
 
 .. toctree::
    :caption: Usage
@@ -36,4 +37,3 @@ Indices and tables
 * :ref:`genindex`
 * :ref:`modindex`
 * :ref:`search`
-

docs/installation.rst

@@ -1,16 +1,53 @@
 Installation
 ============
 
-The easiest way to install OCRmyPDF to follow the steps for your operating system/platform. 
+.. |latest| image:: https://img.shields.io/pypi/v/ocrmypdf.svg
+    :alt: OCRmyPDF latest released version on PyPI
+
+|latest|
+
+The easiest way to install OCRmyPDF to follow the steps for your operating system/platform.
 
 If you want to use the latest version of OCRmyPDF, your best bet is to install the most recent version your platform provides, and then upgrade that version by installing the Python binary wheels.
 
 .. contents:: Platform-specific steps
-    :depth: 1
+    :depth: 2
     :local:
 
-Installing on Debian and Ubuntu 16.10 or newer
-----------------------------------------------
+Installing on Linux
+-------------------
+
+Debian and Ubuntu 16.10 or newer
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. |deb-stable| image:: https://repology.org/badge/version-for-repo/debian_stable/ocrmypdf.svg
+    :alt: Debian 9 stable ("stretch")
+
+.. |deb-testing| image:: https://repology.org/badge/version-for-repo/debian_testing/ocrmypdf.svg
+    :alt: Debian 10 testing ("buster")
+
+.. |deb-unstable| image:: https://repology.org/badge/version-for-repo/debian_unstable/ocrmypdf.svg
+    :alt: Debian unstable
+
+.. |ubu-1710| image:: https://repology.org/badge/version-for-repo/ubuntu_17_10/ocrmypdf.svg
+    :alt: Ubuntu 17.10
+
+.. |ubu-1804| image:: https://repology.org/badge/version-for-repo/ubuntu_18_04/ocrmypdf.svg
+    :alt: Ubuntu 18.04 LTS
+
+.. |ubu-1810| image:: https://repology.org/badge/version-for-repo/ubuntu_18_10/ocrmypdf.svg
+    :alt: Ubuntu 18.10
+
+
++-------------------------------------------+
+| **OCRmyPDF versions in Debian & Ubuntu**  |
++-------------------------------------------+
+| |latest|                                  |
++-------------------------------------------+
+| |deb-stable| |deb-testing| |deb-unstable| |
++-------------------------------------------+
+| |ubu-1710| |ubu-1804| |ubu-1810|          |
++-------------------------------------------+
 
 Users of Debian 9 ("stretch") or later or Ubuntu 16.10 or later may simply
 
@@ -18,12 +55,219 @@ Users of Debian 9 ("stretch") or later or Ubuntu 16.10 or later may simply
 
     apt-get install ocrmypdf
 
-To see what versions are available, check the `Debian Package Tracker <https://tracker.debian.org/pkg/ocrmypdf>`_ or `Ubuntu launchpad.net <https://launchpad.net/ocrmypdf>`_.
+As indicated in the table above, Debian and Ubuntu releases may lag behind the latest version. If the version available for your platform is out of date, you could opt to install the latest version from source. See `Installing HEAD revision from sources`_.
 
-Installing on macOS with Homebrew
----------------------------------
+For full details on version availability for your platform, check the `Debian Package Tracker <https://tracker.debian.org/pkg/ocrmypdf>`_ or `Ubuntu launchpad.net <https://launchpad.net/ocrmypdf>`_.
 
-.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg   
+.. note::
+
+    OCRmyPDF for Debian and Ubuntu currently omit the JBIG2 encoder. OCRmyPDF works fine without it but will produce larger output files. If you build jbig2enc from source, ocrmypdf 7.0.0 and later will automatically detect it (specifically the ``jbig2`` binary) on the ``PATH``. To add JBIG2 encoding, see :ref:`jbig2`.
+
+Fedora 29 or newer
+^^^^^^^^^^^^^^^^^^
+
+.. |fedora-29| image:: https://repology.org/badge/version-for-repo/fedora29/ocrmypdf.svg
+    :alt: Fedora 29
+
+.. |fedora-rawhide| image:: https://repology.org/badge/version-for-repo/fedora_rawhide/ocrmypdf.svg
+    :alt: Fedore Rawhide
+
+
++------------------------------+
+| **OCRmyPDF version**         |
++------------------------------+
+| |latest|                     |
++------------------------------+
+| |fedora-29| |fedora-rawhide| |
++------------------------------+
+
+Users of Fedora 29 later may simply
+
+.. code-block:: bash
+
+    dnf install ocrmypdf
+
+For full details on version availability, check the `Fedora Package Tracker
+<https://apps.fedoraproject.org/packages/ocrmypdf>`_.
+
+If the version available for your platform is out of date, you could opt to
+install the latest version from source. See `Installing HEAD revision from
+sources`_.
+
+.. note::
+
+    OCRmyPDF for Fedora currently omits the JBIG2 encoder due to patent issues.
+    OCRmyPDF works fine without it but will produce larger output files. If you
+    build jbig2enc from source, ocrmypdf 7.0.0 and later will automatically
+    detect it on the ``PATH``. To add JBIG2 encoding, see `Installing the JBIG2
+    encoder`_.
+
+Installing the latest version on Ubuntu 18.04 LTS
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Ubuntu 18.04 includes ocrmypdf 6.1.2. To install a more recent version, first
+install the system version to get most of the dependencies:
+
+.. code-block:: bash
+
+    sudo apt-get update
+    sudo apt-get install \
+        ocrmypdf \
+        python3-pip
+
+There are a few dependency changes between ocrmypdf 6.1.2 and 7.x. Let's get
+these, too.
+
+.. code-block:: bash
+
+    sudo apt-get install \
+        libexempi3 \
+        pngquant
+
+Then install the most recent ocrmypdf for the local user and set the user's ``PATH`` to check for the user's Python packages.
+
+.. code-block:: bash
+
+    export PATH=$HOME/.local/bin:$PATH
+    pip3 install --user ocrmypdf
+
+To add JBIG2 encoding, see :ref:`jbig2`.
+
+Ubuntu 16.04 LTS
+^^^^^^^^^^^^^^^^
+
+No package is currently available for Ubuntu 16.04, but you can install the dependencies manually:
+
+.. code-block:: bash
+
+    sudo apt-get update
+    sudo apt-get install \
+        ghostscript \
+        libexempi3 \
+        pngquant \
+        python3-cffi \
+        python3-pip \
+        qpdf \
+        tesseract-ocr \
+        unpaper
+
+If you wish install OCRmyPDF for the current user, and ensure that the ``PATH``
+environment variable contains ``$HOME/.local/bin``.
+
+.. code-block:: bash
+
+    export PATH=$HOME/.local/bin:$PATH
+    pip3 install --user ocrmypdf
+
+Alternately, you can install ocrmypdf system-wide. (Not recommended.)
+
+.. code-block:: bash
+
+    sudo pip3 install ocrmypdf
+
+At your option, you may upgrade Ubuntu 16.04 LTS to Tesseract 4.0 for improved OCR results.
+
+.. code-block:: bash
+
+    sudo apt-get install -y software-properties-common python-software-properties
+    sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
+    sudo apt-get update
+    sudo apt-get upgrade tesseract-ocr
+
+To add JBIG2 encoding, see :ref:`jbig2`.
+
+Ubuntu 14.04 LTS
+^^^^^^^^^^^^^^^^
+
+Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other options, because it is older and does not provide ``pip``.
+
+Update apt-get:
+
+.. code-block:: bash
+
+    sudo apt-get update
+
+Install system dependencies:
+
+.. code-block:: bash
+
+    sudo apt-get install \
+        software-properties-common python-software-properties \
+        zlib1g-dev \
+        libexempi3 \
+        libjpeg-dev \
+        libffi-dev \
+        pngquant \
+        qpdf
+
+We will need backports of Ghostscript 9.16, libav-11 (for unpaper 6.1), Tesseract 4.00 (alpha), and Python 3.6. This will replace Ghostscript and Tesseract 3.x on your system. Python 3.6 will be installed alongside the system Python 3.4.
+
+If you prefer to not modify your system in this matter, consider using a Docker container.
+
+.. code-block:: bash
+
+    sudo add-apt-repository ppa:vshn/ghostscript -y
+    sudo add-apt-repository ppa:heyarje/libav-11 -y
+    sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
+    sudo add-apt-repository ppa:jonathonf/python-3.6 -y
+
+    sudo apt-get update
+
+    sudo apt-get install \
+        python3.6-dev \
+        ghostscript \
+        tesseract-ocr \
+        tesseract-ocr-eng \
+        libavformat56 libavcodec56 libavutil54 \
+        wget
+
+Now we need to install ``pip`` and let it install ocrmypdf:
+
+.. code-block:: bash
+
+    curl https://bootstrap.pypa.io/ez_setup.py -o - | python3.6 && python3.6 -m easy_install pip
+    pip3.6 install ocrmypdf
+
+The ``wget`` command will download a program and run it.
+
+These installation instructions omit the optional dependency ``unpaper``, which is only available at version 0.4.2 in Ubuntu 14.04. The author could not find a backport of ``unpaper``, and created a .deb package to do the job of installing unpaper 6.1 (for x86 64-bit only):
+
+.. code-block:: bash
+
+    wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O unpaper_6.1-1.deb
+    sudo dpkg -i unpaper_6.1-1.deb
+
+To add JBIG2 encoding, see :ref:`jbig2`.
+
+ArchLinux
+^^^^^^^^^
+
+.. image:: https://repology.org/badge/version-for-repo/aur/ocrmypdf.svg
+    :alt: ArchLinux
+    :target: https://repology.org/metapackage/ocrmypdf
+
+The author is aware of an `ArchLinux User Repository package for ocrmypdf <https://aur.archlinux.org/packages/ocrmypdf/>`_. You can use the following command.
+
+.. code-block:: bash
+
+    yaourt -S ocrmypdf
+
+If you have any difficulties with installation, check the repository package page.
+
+Other Linux packages
+^^^^^^^^^^^^^^^^^^^^
+
+See the `Repology <https://repology.org/metapackage/ocrmypdf/versions>`_ page.
+
+In general, first install the OCRmyPDF package for your system, then optionally use the procedure `Installing with Python pip`_ to install a more recent version.
+
+Installing on macOS
+-------------------
+
+Homebrew
+^^^^^^^^
+
+.. image:: https://img.shields.io/homebrew/v/ocrmypdf.svg
     :alt: homebrew
     :target: http://brewformulas.org/Ocrmypdf
 
@@ -41,12 +285,64 @@ OCRmyPDF is now a standard `Homebrew <https://brew.sh>`_ formula. To install on
 
     Users who previously installed OCRmyPDF from the private tap should switch to the mainline version (``brew untap jbarlow83/ocrmypdf``) and install from there.
 
-.. _Docker-install:
+Manual installation on macOS
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+These instructions probably work on all macOS supported by Homebrew.
+
+If it's not already present, `install Homebrew <http://brew.sh/>`_.
+
+Update Homebrew:
+
+.. code-block:: bash
+
+    brew update
+
+Install or upgrade the required Homebrew packages, if any are missing. To do this, download the ``Brewfile`` that lists all of the dependencies to the current directory, and run ``brew bundle`` to process them (installing or upgrading as needed). ``Brewfile`` is a plain text file.
+
+.. code-block:: bash
+
+    wget https://github.com/jbarlow83/OCRmyPDF/raw/master/.travis/Brewfile
+    brew bundle
+
+This will include the English, French, German and Spanish language packs. If you need other languages you can optionally install them all:
+
+.. _macos-all-languages:
+
+.. code-block:: bash
+
+    brew install tesseract --with-all-languages  # Option 2: for all language packs
+
+Update the homebrew pip:
+
+.. code-block:: bash
+
+    pip3 install --upgrade pip
+
+You can then install OCRmyPDF from PyPI, for the current user:
+
+.. code-block:: bash
+
+    pip3 install --user ocrmypdf
+
+or system-wide:
+
+.. code-block:: bash
+
+    pip3 install ocrmypdf
+
+The command line program should now be available:
+
+.. code-block:: bash
+
+    ocrmypdf --help
+
+.. _docker-install:
 
 Installing the Docker image
 ---------------------------
 
-For many users, installing the Docker image will be easier than installing all of OCRmyPDF's dependencies. For Windows, it is the only option.
+For some users, installing the Docker image will be easier than installing all of OCRmyPDF's dependencies. For Windows, it is the only option.
 
 If you have `Docker <https://docs.docker.com/>`_ installed on your system, you can install a Docker image of the latest release.
 
@@ -55,7 +351,7 @@ Follow the Docker installation instructions for your platform.  If you can run t
 .. code-block:: bash
 
     docker run hello-world
-   
+
 OCRmyPDF will use all available CPU cores.  By default, the VirtualBox machine instance on Windows and macOS has only a single CPU core enabled. Use the VirtualBox Manager to determine the name of your Docker engine host, and then follow these optional steps to enable multiple CPUs:
 
 .. code-block:: bash
@@ -103,14 +399,14 @@ You can then run ocrmypdf using the command:
 .. code-block:: bash
 
     docker run --rm ocrmypdf --help
-  
+
 To execute the OCRmyPDF on a local file, you must `provide a writable volume to the Docker image <https://docs.docker.com/userguide/dockervolumes/>`_, and both the input and output file must be inside the writable volume.  This example command uses the current working directory as the writable volume:
 
 .. code-block:: bash
 
     docker run --rm -v "$(pwd):/home/docker" <other docker arguments>   ocrmypdf <your arguments to ocrmypdf>
 
-In this worked example, the current working directory contains an input file called ``test.pdf`` and the output will go to ``output.pdf``: 
+In this worked example, the current working directory contains an input file called ``test.pdf`` and the output will go to ``output.pdf``:
 
 .. code-block:: bash
 
@@ -142,206 +438,12 @@ Or in the wonderful `fish shell <https://fishshell.com/>`_:
 
 .. note::
 
-    The ocrmypdf Docker containers are designed to be used for a single OCR job. The ``docker run --rm`` argument tells Docker to delete temporary storage associated with container when it is done executing.
-
-Manual installation on macOS
-----------------------------
-
-These instructions probably work on all macOS supported by Homebrew.
-
-If it's not already present, `install Homebrew <http://brew.sh/>`_.
-
-Update Homebrew:
-
-.. code-block:: bash
-
-    brew update
-
-Install or upgrade the required Homebrew packages, if any are missing:
-
-.. code-block:: bash
-
-    brew install libpng openjpeg jbig2dec libtiff     # image libraries
-    brew install qpdf
-    brew install ghostscript
-    brew install python3
-    brew install libxml2 libffi leptonica
-    brew install unpaper   # optional
-   
-Python 3.5 and 3.6 are supported.
-
-Install the required Tesseract OCR engine with the language packs you plan to use:
-   
-.. code-block:: bash
-
-    brew install tesseract                       # Option 1: for English, French, German, Spanish
-
-.. _macos-all-languages:
-
-.. code-block:: bash
-
-    brew install tesseract --with-all-languages  # Option 2: for all language packs
-
-Update the homebrew pip:
-
-.. code-block:: bash
-
-    pip3 install --upgrade pip
-
-You can then install OCRmyPDF from PyPI, for the current user:
-
-.. code-block:: bash
-
-    pip3 install --user ocrmypdf[fitz]
-
-or system-wide:
-
-.. code-block:: bash
-
-    pip3 install ocrmypdf[fitz]
-
-``[fitz]`` includes the optional dependency on PyMuPDF, which improves OCRmyPDF's output in many cases.
-
-The command line program should now be available:
-
-.. code-block:: bash
-
-    ocrmypdf --help
-
-Installing the latest version on Ubuntu 18.04 LTS
--------------------------------------------------
-
-Ubuntu 18.04 includes ocrmypdf 6.1.2. To install a more recent version, first
-install the system version to get all the dependencies:
-
-.. code-block:: bash
-
-    sudo apt-get update
-    sudo apt-get install \
-        ocrmypdf \
-        python3-pip
-
-Then install ocrmypdf 6.1.5 for the local user and set the user's ``PATH`` to check for the user's Python packages.
-
-.. code-block:: bash
-
-    export PATH=$HOME/.local/bin:$PATH
-    pip3 install --user ocrmypdf[fitz]
-
-
-Installing on Ubuntu 16.04 LTS
-------------------------------
-
-No package is currently available for Ubuntu 16.04, but you can install the dependencies manually:
-
-.. code-block:: bash
-
-    sudo apt-get update
-    sudo apt-get install \
-        unpaper \
-        ghostscript \
-        tesseract-ocr \
-        qpdf \
-        python3-pip \
-        python3-cffi
-
-If you wish install OCRmyPDF for the current user:
-
-.. code-block:: bash
-
-    pip3 install --user ocrmypdf[fitz]
-
-Alternately, system-wide. Note that this may modify the system Python environment:
-
-.. code-block:: bash
-
-    sudo pip3 install ocrmypdf[fitz]
-
-If you wish to install OCRmyPDF to a virtual environment to isolate the system Python, you can follow these steps.
-
-.. code-block:: bash
-
-    python3 -m venv venv-ocrmypdf
-    source venv-ocrmypdf/bin/activate
-    pip3 install ocrmypdf[fitz]
-
-``[fitz]`` includes the optional dependency on PyMuPDF, which improves OCRmyPDF's output in many cases.
-
-
-Installing on Ubuntu 14.04 LTS
-------------------------------
-
-Installing on Ubuntu 14.04 LTS (trusty) is more difficult than some other options, because it is older and does not provide ``pip``.
-
-Update apt-get:
-
-.. code-block:: bash
-
-    sudo apt-get update
-
-Install system dependencies:
-
-.. code-block:: bash
-
-    sudo apt-get install \
-        software-properties-common python-software-properties \
-        zlib1g-dev \
-        libjpeg-dev \
-        libffi-dev \
-        qpdf
-
-We will need backports of Ghostscript 9.16, libav-11 (for unpaper 6.1), Tesseract 4.00 (alpha), and Python 3.6. This will replace Ghostscript and Tesseract 3.x on your system. Python 3.6 will be installed alongside the system Python 3.
-
-If you prefer to not modify your system in this matter, consider using a Docker container.
-
-.. code-block:: bash
-
-    sudo add-apt-repository ppa:vshn/ghostscript -y
-    sudo add-apt-repository ppa:heyarje/libav-11 -y
-    sudo add-apt-repository ppa:alex-p/tesseract-ocr -y
-    sudo add-apt-repository ppa:jonathonf/python-3.6 -y
-
-    sudo apt-get update
-
-    sudo apt-get install \
-        python3.6-dev \
-        ghostscript \
-        tesseract-ocr \
-        tesseract-ocr-eng \
-        libavformat56 libavcodec56 libavutil54 \
-        wget
-
-Now we need to install ``pip`` and let it install ocrmypdf:
-
-.. code-block:: bash
-
-    curl https://bootstrap.pypa.io/ez_setup.py -o - | python3.6 && python3.6 -m easy_install pip
-    pip3.6 install ocrmypdf[fitz]
-
-The ``wget`` command will download a program and run it.
-
-These installation instructions omit the optional dependency ``unpaper``, which is only available at version 0.4.2 in Ubuntu 14.04. The author could not find a backport of ``unpaper``, and created a .deb package to do the job of installing unpaper 6.1 (for x86 64-bit only):
-
-.. code-block:: bash
-
-    wget -q 'https://www.dropbox.com/s/vaq0kbwi6e6au80/unpaper_6.1-1.deb?raw=1' -O unpaper_6.1-1.deb
-    sudo dpkg -i unpaper_6.1-1.deb
-
-
-Installing on ArchLinux
------------------------
-
-The author is aware of an `ArchLinux package for ocrmypdf <https://aur.archlinux.org/packages/ocrmypdf/>`_. It seems like the following command might work.
-
-.. code-block:: bash
-
-    pacman -S ocrmypdf
-
+    The ocrmypdf Docker images are designed for application delivery, to enable use of OCRmyPDF without fussing with dependencies. ``docker run --rm`` argument tells Docker to delete the container after it runs, because each container is only good for a single job. The Docker image is not designed for use as a persistent web service or for use on Amazon EC2 Container Service (AWS ECS).
 
 Installing on Windows
 ---------------------
 
-Direct installation on Windows is not possible.  Install the _`Docker` container as described above.  Ensure that your command prompt can run the docker "hello world" container.
+Direct installation on Windows is not possible.  `Install the Docker <docker-install_>`_ container as described above.  Ensure that your command prompt can run the docker "hello world" container.
 
 It would probably not be too difficult to run on Windows.  The main reason this has been avoided is the difficulty of packaging and installing the various non-Python dependencies: Tesseract, QPDF, Ghostscript, Leptonica.  Pull requests to add or improve Windows support would be quite welcome.
 
@@ -363,7 +465,11 @@ where /c/Users/sampleuser is a Unix representation of the Windows path C:\\Users
 Installing with Python pip
 --------------------------
 
-First, install `your platform's version <https://repology.org/metapackage/ocrmypdf/versions>`_ of ``ocrmypdf``, if available, as a way of ensuring that external dependencies are (mostly) satisified, even though the platform version may be out of date. Use ``ocrmypdf --version`` to confirm what version was installed.
+OCRmyPDF is delivered by PyPI because it is a convenient way to install the latest version. However, PyPI and ``pip`` cannot address the fact that ``ocrmypdf`` depends on certain non-Python system libraries and programs being instsalled.
+
+For best results, first install `your platform's version <https://repology.org/metapackage/ocrmypdf/versions>`_ of ``ocrmypdf``, using the instructions elsewhere in this document. Then you can use ``pip`` to get the latest version if your platform version is out of date. Chances are that this will satisfy most dependencies.
+
+Use ``ocrmypdf --version`` to confirm what version was installed.
 
 Then you can install the latest OCRmyPDF from the Python wheels. First try:
 
@@ -382,27 +488,36 @@ Since ``pip3 install --user`` does not work correctly on some platforms, notably
 Requirements for pip and HEAD install
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-OCRmyPDF currently requires these external programs to be installed:
+OCRmyPDF currently requires these external programs and libraries to be installed, and must be satisfied using the operating system package manager. ``pip`` cannot provide them.
 
 - Python 3.5 or newer
-- Tesseract 3.04 or newer
 - Ghostscript 9.15 or newer
-- qpdf 7.0.0 or newer
+- libexempi3 2.2.0 or newer
+- qpdf 8.1.0 or newer
+- Tesseract 3.04 or newer
 
-The following dependencies are recommended:
+Using Python 3.5 in production is discouraged. Python 3.6 and 3.7 give much better performance.
 
-- Python 3.6
-- Tesseract 4.00 or newer
-- Ghostscript 9.22 or newer
-- qpdf 8.0.2 or newer
+As of ocrmypdf 7.2.1, the following versions are recommended:
+
+- Python 3.7
+- Ghostscript 9.23 or newer
+- libexempi3 2.4.5 or newer
+- qpdf 8.2.1
+- Tesseract 4.0.0-rc1
+- jbig2enc 0.29 or newer
+- pngquant 2.5 or newer
 - unpaper 6.1
-- PyMuPDF 1.12.5 or newer
 
-These are in addition to the Python packaging dependencies, meaning that unfortunately, the ``pip install`` command cannot satisfy all of them. 
+jbig2enc, pngquant, and unpaper are optional. If missing certain features are disabled. OCRmyPDF will discover them as soon as they are available.
 
-Python 3.6 and Tesseract 4.0.0-beta.1 are recommended for best OCR results and best performance.
+**jbig2enc**, if present, will be used to optimize the encoding of monochrome images.  This can significantly reduce the file size of the output file.  It is not required.  `jbig2enc <https://github.com/agl/jbig2enc>`_ is not generally available for Ubuntu or Debian due to lingering concerns about patent issues, but can easily be built from source. To add JBIG2 encoding, see :ref:`jbig2`.
 
-The library PyMuPDF is not widely available in platform distributions, and it improves OCRmyPDF in certain conditions. Consider installing OCRmyPDF from the Python binary wheels, which include a precompiled version of this library.
+**pngquant**, if present, is optionally used to optimize the encoding of PNG-style images in PDFs (actually, any that are that losslessly encoded) by lossily quantizing to a smaller color palette. It is only activated then the ``--optimize`` argument is ``2`` or ``3``.
+
+**unpaper**, if present, enables the ``--clean`` and ``--clean-final`` command line options.
+
+These are in addition to the Python packaging dependencies, meaning that unfortunately, the ``pip install`` command cannot satisfy all of them.
 
 
 Installing HEAD revision from sources
@@ -410,6 +525,8 @@ Installing HEAD revision from sources
 
 If you have ``git`` and Python 3.5 or newer installed, you can install from source. When the ``pip`` installer runs, it will alert you if dependencies are missing.
 
+If you prefer to build every from source, you will need to `build pikepdf from source <https://pikepdf.readthedocs.io/en/latest/installation.html#building-from-source>`_. First ensure you can build and install pikepdf.
+
 To install the HEAD revision from sources in the current Python 3 environment:
 
 .. code-block:: bash
@@ -432,8 +549,8 @@ You may find it easiest to install in a virtual environment, rather than system-
     cd OCRmyPDF
     pip3 install .
 
-However, ``ocrmypdf`` will only be accessible on the system PATH after
-you activate the virtual environment.
+However, ``ocrmypdf`` will only be accessible on the system PATH
+when you activate the virtual environment.
 
 To run the program:
 
@@ -446,8 +563,18 @@ need to be installed. The script requires specific versions of the
 dependencies. Older version than the ones mentioned in the release notes
 are likely not to be compatible to OCRmyPDF.
 
+For development
+~~~~~~~~~~~~~~~
 
-Other Linux packages
---------------------
+To install all of the development and test requirements:
 
-See the `Repology <https://repology.org/metapackage/ocrmypdf/versions>`_ page.
+.. code-block:: bash
+
+    git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
+    python3 -m venv
+    source venv/bin/activate
+    cd OCRmyPDF
+    pip install -e .
+    pip install -r requirements/dev.txt -r requirements/test.txt
+
+To add JBIG2 encoding, see :ref:`jbig2`.

docs/introduction.rst

@@ -1,9 +1,7 @@
 Introduction
 ============
-
 OCRmyPDF is a Python 3 package that adds OCR layers to PDFs.
 
-
 About OCR
 ---------
 
@@ -66,6 +64,8 @@ In the case of a PDF that is nothing other than a container of images (no rotati
 
 OCRmyPDF uses several strategies depending on input options and the input PDF itself, but generally speaking it rasterizes a page for OCR and then grafts the OCR back onto the original. As such it can handle complex PDFs and still preserve their contents as much as possible.
 
+OCRmyPDF also supports a many, many edge cases that have cropped over several years of development. We support PDF features like images inside of Form XObjects, and pages with UserUnit scaling. We support rare image formats like non-monochrome 1-bit images. We warn about files you may not to OCR. Thanks to pikepdf and QPDF, we auto-repair PDFs that are damaged. (Not that you need to know what any of these are! You should be able to throw any PDF at it.)
+
 
 Limitations
 -----------
@@ -76,24 +76,25 @@ OCRmyPDF is limited by the Tesseract OCR engine.  As such it experiences these l
 * It is not capable of recognizing handwriting.
 * It may find gibberish and report this as OCR output.
 * If a document contains languages outside of those given in the ``-l LANG`` arguments, results may be poor.
-* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns and join text across the columns.
+* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns, and may try to join text across columns.
 * Poor quality scans may produce poor quality OCR. Garbage in, garbage out.
-* PDFs that use transparent layers are not currently checked in the test suite, so they may not work correctly.
 * It does not expose information about what font family text belongs to.
 
 OCRmyPDF is also limited by the PDF specification:
 
 * PDF encodes the position of text glyphs but does not encode document structure.  There is no markup that divides a document in sections, paragraphs, sentences, or even words (since blank spaces are not represented). As such all elements of document structure including the spaces between words must be derived heuristically.  Some PDF viewers do a better job of this than others.
-* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround. While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones.
+* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround (when using ``--pdf-renderer hocr``). While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones.
 
 Ghostscript also imposes some limitations:
 
 * PDFs containing JBIG2-encoded content will be converted to CCITT Group4 encoding, which has lower compression ratios, if Ghostscript PDF/A is enabled.
 * PDFs containing JPEG 2000-encoded content will be converted to JPEG encoding, which may introduce compression artifacts, if Ghostscript PDF/A is enabled.
-* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format.
-  
-OCRmyPDF is currently not designed to be used as a Python API; it is designed to be run as a command line tool. ``import ocrmypf`` currently attempts to process the command line on ``sys.argv`` at import time so it has side effects that will interfere with its use as a package. The API it presents should not be considered stable.
+* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format. (Ghostscript 9.25+ can copy JPEG images without transcoding them; earlier versions will transcode.)
 
+Regarding OCRmyPDF itself:
+
+* PDFs that use transparency are not currently represented in the test suite
+* The Python API exported by ``import ocrmypdf`` is design to help scripts that use OCRmyPDF but is not currently capable of running OCRmyPDF jobs due to limitations in an underlying library.
 
 Similar programs
 ----------------
@@ -112,4 +113,3 @@ Web front-ends
 * `OCRmyPDF-web <https://github.com/sseemayer/OCRmyPDF-web>`_, a micro web-frontend for OCRmyPDF (third-party, not actively maintained)
 
 Bear in mind that OCRmyPDF is not designed to be secure against malware-bearing PDFs (see `Using OCRmyPDF online`_).
-

docs/jbig2.rst

@@ -0,0 +1,35 @@
+.. _jbig2:
+
+Installing the JBIG2 encoder
+============================
+
+Most Linux distributions do not include a JBIG2 encoder since JBIG2 encoding was patented for a long time. All known JBIG2 US patents have expired as of 2017, but it is possible that unknown patents exist.
+
+JBIG2 encoding is recommended for OCRmyPDF and is used to losslessly create smaller PDFs. If JBIG2 encoding not available, lower quality encodings will be used.
+
+JBIG2 decoding is not patented and is performed automatically by most PDF viewers. It is widely supported has been part of the PDF specification since 2001.
+
+On macOS, Homebrew packages jbig2enc and OCRmyPDF includes it by default. The Docker image for OCRmyPDF also builds its own JBIG2 encoder from source.
+
+For all other Linux, you must build a JBIG2 encoder from source:
+
+.. code-block:: bash
+
+    git clone https://github.com/agl/jbig2enc
+    cd jbig2enc
+    ./autogen.sh
+    ./configure && make
+    [sudo] make install
+
+.. _jbig2-lossy:
+
+Lossy mode JBIG2
+----------------
+
+OCRmyPDF provides lossy mode JBIG2 as an advanced feature. Users should `review the technical concerns with JBIG2 in lossy mode <https://abbyy.technology/en:kb:tip:jbig2_compression_and_ocr>`_ and decide if this feature is acceptable for their use case.
+
+JBIG2 lossy mode does achieve higher compression ratios than any other monochrome (bitonal) compression technology; for large text documents the savings are considerable. JBIG2 lossless still gives great compression ratios and is a major improvement over the older CCITT G4 standard. As explained above, there is some risk of substitution errors.
+
+To turn on JBIG2 lossy mode, add the argument ``--jbig2-lossy``. ``--optimize {1,2,3}`` are necessary for the argument to take effect also required. Also, a JBIG2 encoder must be installed as described in the previous section.
+
+*ocrmypdf v7.0 and v7.1 used lossy mode by default.*

docs/languages.rst

@@ -17,12 +17,27 @@ Debian and Ubuntu users
    # Display a list of all Tesseract language packs
    apt-cache search tesseract-ocr
 
-   # Debian/Ubuntu users
-   apt-get install tesseract-ocr-chi-sim  # Example: Install Chinese Simplified language back
-   
+   # Install Chinese Simplified language pack
+   apt-get install tesseract-ocr-chi-sim
+
 You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to what languages it should search for. Multiple
 languages can be requested using either ``-l eng+fre`` (English and French) or ``-l eng -l fre``.
 
+Fedora users
+------------
+
+.. code-block:: bash
+
+   # Display a list of all Tesseract language packs
+   dnf search tesseract
+
+   # Install Chinese Simplified language pack
+   dnf install tesseract-langpack-chi_sim
+
+You can then pass the ``-l LANG`` argument to OCRmyPDF to give a hint as to
+what languages it should search for. Multiple languages can be requested using
+either ``-l eng+fre`` (English and French) or ``-l eng -l fre``.
+
 macOS users
 -----------
 
@@ -59,18 +74,3 @@ If you wish to add a single language pack, you could do the following:
 .. code-block:: bash
 
     host$ docker commit <container_id> ocrmypdf-tess4-heb
-
-
-Known limitations
------------------
-
-As of v4.2, users of ocrmypdf working languages outside the Latin alphabet should use the following syntax:
-
-.. code-block:: bash
-
-	ocrmypdf -l eng+gre --output-type pdf --pdf-renderer tesseract
-
-The reasons for this are:
-
-* The latest version of Ghostscript (9.19 as of this writing) has unfixed bugs in Unicode handling that generate invalid character maps, so Ghostscript cannot be used for PDF/A conversion
-* The default "hocr" PDF renderer does not handle Asian fonts properly

docs/pipeline.svg

@@ -4,550 +4,389 @@
 <!-- Generated by graphviz version 2.40.1 (20161225.0304)
  -->
 <!-- Title: Pipeline: Pages: 1 -->
-<svg width="1952pt" height="1277pt"
- viewBox="0.00 0.00 1952.00 1277.18" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
-<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 1273.1825)">
+<svg width="1484pt" height="1277pt"
+ viewBox="0.00 0.00 1484.00 1277.20" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
+<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 1273.1977)">
 <title>Pipeline:</title>
-<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-1273.1825 1948,-1273.1825 1948,4 -4,4"/>
+<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-1273.1977 1480,-1273.1977 1480,4 -4,4"/>
 <g id="clust1" class="cluster">
 <title>clustertasks</title>
-<polygon fill="none" stroke="#000000" points="8,-8 8,-1261.1825 1936,-1261.1825 1936,-8 8,-8"/>
-<text text-anchor="middle" x="972" y="-1233.1825" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
+<polygon fill="none" stroke="#000000" points="8,-8 8,-1261.1977 1468,-1261.1977 1468,-8 8,-8"/>
+<text text-anchor="middle" x="738" y="-1233.1977" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
 </g>
 <!-- t0 -->
 <g id="node1" class="node">
 <title>t0</title>
-<polygon fill="#efa03b" stroke="#000000" points="876.7053,-1215.1825 667.2947,-1215.1825 663.2947,-1211.1825 663.2947,-1179.1825 872.7053,-1179.1825 876.7053,-1183.1825 876.7053,-1215.1825"/>
-<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 663.2947,-1211.1825 "/>
-<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 872.7053,-1179.1825 "/>
-<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 876.7053,-1215.1825 "/>
-<text text-anchor="middle" x="770" y="-1191.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.triage</text>
+<polygon fill="#efa03b" stroke="#000000" points="1215.7053,-1215.1977 1006.2947,-1215.1977 1002.2947,-1211.1977 1002.2947,-1179.1977 1211.7053,-1179.1977 1215.7053,-1183.1977 1215.7053,-1215.1977"/>
+<polyline fill="none" stroke="#000000" points="1211.7053,-1211.1977 1002.2947,-1211.1977 "/>
+<polyline fill="none" stroke="#000000" points="1211.7053,-1211.1977 1211.7053,-1179.1977 "/>
+<polyline fill="none" stroke="#000000" points="1211.7053,-1211.1977 1215.7053,-1215.1977 "/>
+<text text-anchor="middle" x="1109" y="-1191.1977" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.triage</text>
 </g>
 <!-- t1 -->
 <g id="node2" class="node">
 <title>t1</title>
-<polygon fill="#efa03b" stroke="#000000" points="941.0433,-1157.1825 602.9567,-1157.1825 598.9567,-1153.1825 598.9567,-1121.1825 937.0433,-1121.1825 941.0433,-1125.1825 941.0433,-1157.1825"/>
-<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 598.9567,-1153.1825 "/>
-<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 937.0433,-1121.1825 "/>
-<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 941.0433,-1157.1825 "/>
-<text text-anchor="middle" x="770" y="-1133.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.repair_and_parse_pdf</text>
+<polygon fill="#efa03b" stroke="#000000" points="1280.0433,-1157.1977 941.9567,-1157.1977 937.9567,-1153.1977 937.9567,-1121.1977 1276.0433,-1121.1977 1280.0433,-1125.1977 1280.0433,-1157.1977"/>
+<polyline fill="none" stroke="#000000" points="1276.0433,-1153.1977 937.9567,-1153.1977 "/>
+<polyline fill="none" stroke="#000000" points="1276.0433,-1153.1977 1276.0433,-1121.1977 "/>
+<polyline fill="none" stroke="#000000" points="1276.0433,-1153.1977 1280.0433,-1157.1977 "/>
+<text text-anchor="middle" x="1109" y="-1133.1977" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.repair_and_parse_pdf</text>
 </g>
 <!-- t0&#45;&gt;t1 -->
 <g id="edge1" class="edge">
 <title>t0&#45;&gt;t1</title>
-<path fill="none" stroke="#0044a0" d="M770,-1179.1488C770,-1175.4743 770,-1171.5242 770,-1167.5888"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="773.5001,-1167.4047 770,-1157.4047 766.5001,-1167.4047 773.5001,-1167.4047"/>
+<path fill="none" stroke="#0044a0" d="M1109,-1179.1641C1109,-1175.4895 1109,-1171.5395 1109,-1167.604"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1112.5001,-1167.4199 1109,-1157.4199 1105.5001,-1167.42 1112.5001,-1167.4199"/>
 </g>
 <!-- t2 -->
 <g id="node3" class="node">
 <title>t2</title>
-<polygon fill="#efa03b" stroke="#000000" points="1043.9447,-1078.2585 787,-1095.158 530.0553,-1078.2585 530.2951,-1050.9147 1043.7049,-1050.9147 1043.9447,-1078.2585"/>
-<polygon fill="none" stroke="#000000" points="1047.9814,-1081.9991 787,-1099.164 526.0186,-1081.9991 526.3263,-1046.9173 1047.6737,-1046.9173 1047.9814,-1081.9991"/>
-<text text-anchor="middle" x="787" y="-1064.7009" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.pre_split_pages</text>
+<polygon fill="#efa03b" stroke="#000000" points="1030.0926,-1078.2732 785,-1095.1731 539.9074,-1078.2732 540.1362,-1050.9285 1029.8638,-1050.9285 1030.0926,-1078.2732"/>
+<polygon fill="none" stroke="#000000" points="1034.1249,-1082.0021 785,-1099.1801 535.8751,-1082.0021 536.1685,-1046.931 1033.8315,-1046.931 1034.1249,-1082.0021"/>
+<text text-anchor="middle" x="785" y="-1064.7153" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.marker_pages</text>
 </g>
 <!-- t1&#45;&gt;t2 -->
 <g id="edge2" class="edge">
 <title>t1&#45;&gt;t2</title>
-<path fill="none" stroke="#0044a0" d="M774.5557,-1120.8306C775.4962,-1117.0421 776.5194,-1112.92 777.5618,-1108.7212"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="780.986,-1109.4542 779.9984,-1098.9055 774.1922,-1107.7677 780.986,-1109.4542"/>
+<path fill="none" stroke="#0044a0" d="M1023.449,-1121.1152C984.5695,-1112.8974 938.0675,-1103.0685 896.4957,-1094.2816"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="897.1066,-1090.8335 886.5989,-1092.1898 895.6589,-1097.6822 897.1066,-1090.8335"/>
 </g>
-<!-- t20 -->
-<g id="node22" class="node">
-<title>t20</title>
-<polygon fill="#efa03b" stroke="#000000" points="649.5156,-1020.2193 286.4844,-1020.2193 282.4844,-1016.2193 282.4844,-984.2193 645.5156,-984.2193 649.5156,-988.2193 649.5156,-1020.2193"/>
-<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 282.4844,-1016.2193 "/>
-<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 645.5156,-984.2193 "/>
-<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 649.5156,-1020.2193 "/>
-<text text-anchor="middle" x="466" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.generate_postscript_stub</text>
+<!-- t16 -->
+<g id="node17" class="node">
+<title>t16</title>
+<polygon fill="#00cc66" stroke="#000000" points="1215.5059,-328.1861 942.4941,-328.1861 938.4941,-324.1861 938.4941,-292.1861 1211.5059,-292.1861 1215.5059,-296.1861 1215.5059,-328.1861"/>
+<polyline fill="none" stroke="#000000" points="1211.5059,-324.1861 938.4941,-324.1861 "/>
+<polyline fill="none" stroke="#000000" points="1211.5059,-324.1861 1211.5059,-292.1861 "/>
+<polyline fill="none" stroke="#000000" points="1211.5059,-324.1861 1215.5059,-328.1861 "/>
+<text text-anchor="middle" x="1077" y="-304.1861" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.weave_layers</text>
 </g>
-<!-- t1&#45;&gt;t20 -->
-<g id="edge31" class="edge">
-<title>t1&#45;&gt;t20</title>
-<path fill="none" stroke="#0044a0" d="M598.785,-1123.5822C563.105,-1117.7214 531.8209,-1109.8012 517,-1099.1825 493.9382,-1082.6593 480.4369,-1052.4741 473.1842,-1030.1561"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="476.5199,-1029.095 470.2797,-1020.5329 469.8185,-1031.1177 476.5199,-1029.095"/>
+<!-- t1&#45;&gt;t16 -->
+<g id="edge24" class="edge">
+<title>t1&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M1097.2373,-1120.9576C1077.987,-1088.7184 1043.7317,-1019.7372 1063,-963.2656 1106.87,-834.6914 1482.8098,-573.3183 1409,-459.2656 1362.7595,-387.8137 1273.2613,-350.2799 1198.8305,-330.7528"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1199.4555,-327.3005 1188.9015,-328.2301 1197.7316,-334.0849 1199.4555,-327.3005"/>
 </g>
-<!-- t24 -->
-<g id="node24" class="node">
-<title>t24</title>
-<polygon fill="#efa03b" stroke="#000000" points="20.4213,-156.4247 242,-129.112 463.5787,-156.4247 463.3719,-200.6177 20.6281,-200.6177 20.4213,-156.4247"/>
-<polygon fill="none" stroke="#000000" points="16.4083,-152.8864 242,-125.079 467.5917,-152.8864 467.3496,-204.6205 16.6504,-204.6205 16.4083,-152.8864"/>
-<text text-anchor="middle" x="242" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_mupdf</text>
+<!-- t17 -->
+<g id="node18" class="node">
+<title>t17</title>
+<polygon fill="#efa03b" stroke="#000000" points="1439.5156,-1009.7492 1076.4844,-1009.7492 1072.4844,-1005.7492 1072.4844,-973.7492 1435.5156,-973.7492 1439.5156,-977.7492 1439.5156,-1009.7492"/>
+<polyline fill="none" stroke="#000000" points="1435.5156,-1005.7492 1072.4844,-1005.7492 "/>
+<polyline fill="none" stroke="#000000" points="1435.5156,-1005.7492 1435.5156,-973.7492 "/>
+<polyline fill="none" stroke="#000000" points="1435.5156,-1005.7492 1439.5156,-1009.7492 "/>
+<text text-anchor="middle" x="1256" y="-985.7492" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.generate_postscript_stub</text>
 </g>
-<!-- t1&#45;&gt;t24 -->
-<g id="edge41" class="edge">
-<title>t1&#45;&gt;t24</title>
-<path fill="none" stroke="#0044a0" d="M598.6617,-1124.8567C376.4578,-1103.9215 17,-1060.9746 17,-1002.2193 17,-1002.2193 17,-1002.2193 17,-335.7123 17,-276.266 68.2524,-235.5065 121.8346,-209.1377"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="123.6227,-212.1629 131.1479,-204.7049 120.6143,-205.8423 123.6227,-212.1629"/>
+<!-- t1&#45;&gt;t17 -->
+<g id="edge25" class="edge">
+<title>t1&#45;&gt;t17</title>
+<path fill="none" stroke="#0044a0" d="M1127.0251,-1121.1176C1153.0342,-1095.0292 1201.2007,-1046.7157 1230.7429,-1017.0834"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1233.2458,-1019.5302 1237.8275,-1009.9772 1228.2885,-1014.5879 1233.2458,-1019.5302"/>
 </g>
-<!-- t23 -->
-<g id="node25" class="node">
-<title>t23</title>
-<polygon fill="#efa03b" stroke="#000000" points="1006.4902,-156.4239 1218,-129.1092 1429.5098,-156.4239 1429.3124,-200.6199 1006.6876,-200.6199 1006.4902,-156.4239"/>
-<polygon fill="none" stroke="#000000" points="1002.4742,-152.9063 1218,-125.073 1433.5258,-152.9063 1433.2948,-204.623 1002.7052,-204.623 1002.4742,-152.9063"/>
-<text text-anchor="middle" x="1218" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_qpdf</text>
+<!-- t18 -->
+<g id="node19" class="node">
+<title>t18</title>
+<polygon fill="#efa03b" stroke="#000000" points="1070.189,-214.4288 1263,-187.1099 1455.811,-214.4288 1455.631,-258.6317 1070.369,-258.6317 1070.189,-214.4288"/>
+<polygon fill="none" stroke="#000000" points="1066.1715,-210.9543 1263,-183.0662 1459.8285,-210.9543 1459.6181,-262.6353 1066.3819,-262.6353 1066.1715,-210.9543"/>
+<text text-anchor="middle" x="1263" y="-220.6461" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.metadata_fixup</text>
 </g>
-<!-- t1&#45;&gt;t23 -->
-<g id="edge46" class="edge">
-<title>t1&#45;&gt;t23</title>
-<path fill="none" stroke="#0044a0" d="M941.3419,-1136.2202C1263.4143,-1128.3561 1927,-1100.2573 1927,-1002.2193 1927,-1002.2193 1927,-1002.2193 1927,-335.7123 1927,-234.7734 1651.3069,-194.7548 1443.7522,-178.9295"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1443.8443,-175.4267 1433.6111,-178.1712 1443.3223,-182.4072 1443.8443,-175.4267"/>
+<!-- t1&#45;&gt;t18 -->
+<g id="edge28" class="edge">
+<title>t1&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1241.1115,-1121.1192C1311.0076,-1105.2558 1393.5851,-1075.5895 1449,-1020.2329 1458.4922,-1010.7507 1459,-1005.1661 1459,-991.7492 1459,-991.7492 1459,-991.7492 1459,-393.7258 1459,-341.9727 1441.8377,-325.2216 1402,-292.1861 1390.9869,-283.0534 1378.4885,-274.871 1365.6104,-267.6415"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1367.1079,-264.4723 1356.6476,-262.7888 1363.775,-270.6279 1367.1079,-264.4723"/>
 </g>
 <!-- t3 -->
 <g id="node4" class="node">
 <title>t3</title>
-<polygon fill="#efa03b" stroke="#000000" points="1055.3851,-1020.2193 808.6149,-1020.2193 804.6149,-1016.2193 804.6149,-984.2193 1051.3851,-984.2193 1055.3851,-988.2193 1055.3851,-1020.2193"/>
-<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 804.6149,-1016.2193 "/>
-<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1051.3851,-984.2193 "/>
-<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1055.3851,-1020.2193 "/>
-<text text-anchor="middle" x="930" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.split_page</text>
+<polygon fill="#efa03b" stroke="#000000" points="1015.5194,-999.3074 785,-1016.2081 554.4806,-999.3074 554.6958,-971.9616 1015.3042,-971.9616 1015.5194,-999.3074"/>
+<polygon fill="none" stroke="#000000" points="1019.5491,-1003.0204 785,-1020.2165 550.4509,-1003.0204 550.7267,-967.9639 1019.2733,-967.9639 1019.5491,-1003.0204"/>
+<text text-anchor="middle" x="785" y="-985.7492" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_or_skip</text>
 </g>
 <!-- t2&#45;&gt;t3 -->
 <g id="edge3" class="edge">
 <title>t2&#45;&gt;t3</title>
-<path fill="none" stroke="#0044a0" d="M837.0145,-1046.7494C851.946,-1039.5987 868.2558,-1031.7881 882.9879,-1024.733"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="884.9035,-1027.6964 892.4109,-1020.2204 881.88,-1021.383 884.9035,-1027.6964"/>
+<path fill="none" stroke="#0044a0" d="M785,-1046.5938C785,-1041.5635 785,-1036.139 785,-1030.7273"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="788.5001,-1030.4427 785,-1020.4427 781.5001,-1030.4428 788.5001,-1030.4427"/>
 </g>
 <!-- t4 -->
 <g id="node5" class="node">
 <title>t4</title>
-<polygon fill="#efa03b" stroke="#000000" points="1194.5194,-941.2939 964,-958.1945 733.4806,-941.2939 733.6958,-913.948 1194.3042,-913.948 1194.5194,-941.2939"/>
-<polygon fill="none" stroke="#000000" points="1198.5491,-945.0069 964,-962.203 729.4509,-945.0069 729.7267,-909.9503 1198.2733,-909.9503 1198.5491,-945.0069"/>
-<text text-anchor="middle" x="964" y="-927.7357" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_or_skip</text>
+<polygon fill="#efa03b" stroke="#000000" points="889.3166,-941.2656 582.6834,-941.2656 578.6834,-937.2656 578.6834,-905.2656 885.3166,-905.2656 889.3166,-909.2656 889.3166,-941.2656"/>
+<polyline fill="none" stroke="#000000" points="885.3166,-937.2656 578.6834,-937.2656 "/>
+<polyline fill="none" stroke="#000000" points="885.3166,-937.2656 885.3166,-905.2656 "/>
+<polyline fill="none" stroke="#000000" points="885.3166,-937.2656 889.3166,-941.2656 "/>
+<text text-anchor="middle" x="734" y="-917.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_preview</text>
 </g>
 <!-- t3&#45;&gt;t4 -->
 <g id="edge4" class="edge">
 <title>t3&#45;&gt;t4</title>
-<path fill="none" stroke="#0044a0" d="M939.1114,-983.867C941.197,-979.6659 943.4863,-975.0549 945.8053,-970.3839"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="948.9439,-971.9327 950.2559,-961.4194 942.6741,-968.8199 948.9439,-971.9327"/>
+<path fill="none" stroke="#0044a0" d="M767.1627,-967.797C762.8396,-961.9919 758.1927,-955.7519 753.7858,-949.8343"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="756.3465,-947.4129 747.5666,-941.483 750.7323,-951.5939 756.3465,-947.4129"/>
 </g>
 <!-- t5 -->
 <g id="node6" class="node">
 <title>t5</title>
-<polygon fill="#efa03b" stroke="#000000" points="1119.3166,-883.252 812.6834,-883.252 808.6834,-879.252 808.6834,-847.252 1115.3166,-847.252 1119.3166,-851.252 1119.3166,-883.252"/>
-<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 808.6834,-879.252 "/>
-<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1115.3166,-847.252 "/>
-<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1119.3166,-883.252 "/>
-<text text-anchor="middle" x="964" y="-859.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_preview</text>
+<polygon fill="#efa03b" stroke="#000000" points="1028.0824,-883.2656 769.9176,-883.2656 765.9176,-879.2656 765.9176,-847.2656 1024.0824,-847.2656 1028.0824,-851.2656 1028.0824,-883.2656"/>
+<polyline fill="none" stroke="#000000" points="1024.0824,-879.2656 765.9176,-879.2656 "/>
+<polyline fill="none" stroke="#000000" points="1024.0824,-879.2656 1024.0824,-847.2656 "/>
+<polyline fill="none" stroke="#000000" points="1024.0824,-879.2656 1028.0824,-883.2656 "/>
+<text text-anchor="middle" x="897" y="-859.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.orient_page</text>
+</g>
+<!-- t3&#45;&gt;t5 -->
+<g id="edge6" class="edge">
+<title>t3&#45;&gt;t5</title>
+<path fill="none" stroke="#0044a0" d="M869.1177,-967.8063C880.553,-961.0264 890.81,-952.3527 898,-941.2656 907.042,-927.3226 906.9457,-908.7053 904.5139,-893.387"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="907.914,-892.5383 902.5408,-883.4071 901.0469,-893.8961 907.914,-892.5383"/>
 </g>
 <!-- t4&#45;&gt;t5 -->
 <g id="edge5" class="edge">
 <title>t4&#45;&gt;t5</title>
-<path fill="none" stroke="#0044a0" d="M964,-909.7834C964,-904.5184 964,-898.8955 964,-893.4799"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="967.5001,-893.4695 964,-883.4695 960.5001,-893.4695 967.5001,-893.4695"/>
+<path fill="none" stroke="#0044a0" d="M784.6807,-905.2319C801.0865,-899.3943 819.4462,-892.8614 836.4913,-886.7963"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="838.1026,-889.938 846.3506,-883.288 835.7559,-883.343 838.1026,-889.938"/>
 </g>
 <!-- t6 -->
 <g id="node7" class="node">
 <title>t6</title>
-<polygon fill="#efa03b" stroke="#000000" points="1269.0824,-825.252 1010.9176,-825.252 1006.9176,-821.252 1006.9176,-789.252 1265.0824,-789.252 1269.0824,-793.252 1269.0824,-825.252"/>
-<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1006.9176,-821.252 "/>
-<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1265.0824,-789.252 "/>
-<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1269.0824,-825.252 "/>
-<text text-anchor="middle" x="1138" y="-801.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.orient_page</text>
-</g>
-<!-- t4&#45;&gt;t6 -->
-<g id="edge7" class="edge">
-<title>t4&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M1072.6871,-909.8532C1097.768,-902.2558 1119.5887,-893.2079 1128,-883.252 1138.9559,-870.2843 1141.4763,-851.3696 1141.2475,-835.6545"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1144.7291,-835.2052 1140.7199,-825.3983 1137.7384,-835.5649 1144.7291,-835.2052"/>
+<polygon fill="#efa03b" stroke="#000000" points="1064.7325,-825.2656 689.2675,-825.2656 685.2675,-821.2656 685.2675,-789.2656 1060.7325,-789.2656 1064.7325,-793.2656 1064.7325,-825.2656"/>
+<polyline fill="none" stroke="#000000" points="1060.7325,-821.2656 685.2675,-821.2656 "/>
+<polyline fill="none" stroke="#000000" points="1060.7325,-821.2656 1060.7325,-789.2656 "/>
+<polyline fill="none" stroke="#000000" points="1060.7325,-821.2656 1064.7325,-825.2656 "/>
+<text text-anchor="middle" x="875" y="-801.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
 </g>
 <!-- t5&#45;&gt;t6 -->
-<g id="edge6" class="edge">
+<g id="edge7" class="edge">
 <title>t5&#45;&gt;t6</title>
-<path fill="none" stroke="#0044a0" d="M1018.1009,-847.2184C1035.9411,-841.2716 1055.9459,-834.6034 1074.4267,-828.4431"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1075.5526,-831.7572 1083.9326,-825.2745 1073.3389,-825.1165 1075.5526,-831.7572"/>
-</g>
-<!-- t7 -->
-<g id="node8" class="node">
-<title>t7</title>
-<polygon fill="#efa03b" stroke="#000000" points="917.7325,-767.252 542.2675,-767.252 538.2675,-763.252 538.2675,-731.252 913.7325,-731.252 917.7325,-735.252 917.7325,-767.252"/>
-<polyline fill="none" stroke="#000000" points="913.7325,-763.252 538.2675,-763.252 "/>
-<polyline fill="none" stroke="#000000" points="913.7325,-763.252 913.7325,-731.252 "/>
-<polyline fill="none" stroke="#000000" points="913.7325,-763.252 917.7325,-767.252 "/>
-<text text-anchor="middle" x="728" y="-743.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
-</g>
-<!-- t6&#45;&gt;t7 -->
-<g id="edge8" class="edge">
-<title>t6&#45;&gt;t7</title>
-<path fill="none" stroke="#0044a0" d="M1010.5209,-789.2184C964.4341,-782.6988 912.2171,-775.312 865.3222,-768.6781"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="865.792,-765.2098 855.4003,-767.2745 864.8114,-772.1408 865.792,-765.2098"/>
-</g>
-<!-- t17 -->
-<g id="node15" class="node">
-<title>t17</title>
-<polygon fill="#ff69b4" stroke="#000000" points="1224.2696,-447.252 843.7304,-447.252 839.7304,-443.252 839.7304,-411.252 1220.2696,-411.252 1224.2696,-415.252 1224.2696,-447.252"/>
-<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 839.7304,-443.252 "/>
-<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1220.2696,-411.252 "/>
-<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1224.2696,-447.252 "/>
-<text text-anchor="middle" x="1032" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
-</g>
-<!-- t6&#45;&gt;t17 -->
-<g id="edge16" class="edge">
-<title>t6&#45;&gt;t17</title>
-<path fill="none" stroke="#0044a0" d="M1158.4397,-788.9107C1179.7261,-767.6923 1210,-730.5987 1210,-691.252 1210,-691.252 1210,-691.252 1210,-575.252 1210,-532.3545 1227.3937,-512.2638 1200,-479.252 1190.5284,-467.8379 1163.9108,-457.8666 1134.5783,-449.8944"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1135.1079,-446.4151 1124.5475,-447.2681 1133.3349,-453.1869 1135.1079,-446.4151"/>
-</g>
-<!-- t14 -->
-<g id="node17" class="node">
-<title>t14</title>
-<polygon fill="#00cc66" stroke="#000000" points="1565,-457.252 1242.4141,-429.252 1565,-401.252 1887.5859,-429.252 1565,-457.252"/>
-<text text-anchor="middle" x="1565" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_image_layer</text>
-</g>
-<!-- t6&#45;&gt;t14 -->
-<g id="edge22" class="edge">
-<title>t6&#45;&gt;t14</title>
-<path fill="none" stroke="#0044a0" d="M1211.6459,-789.1984C1287.5889,-768.3738 1395,-731.7602 1395,-691.252 1395,-691.252 1395,-691.252 1395,-575.252 1395,-519.4485 1448.3697,-480.3436 1495.3026,-456.7437"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1497.1229,-459.7505 1504.5857,-452.2301 1494.0619,-453.4552 1497.1229,-459.7505"/>
-</g>
-<!-- t21 -->
-<g id="node20" class="node">
-<title>t21</title>
-<polygon fill="#efa03b" stroke="#000000" points="1414.7716,-270.1725 1169.2284,-270.1725 1165.2284,-266.1725 1165.2284,-234.1725 1410.7716,-234.1725 1414.7716,-238.1725 1414.7716,-270.1725"/>
-<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1165.2284,-266.1725 "/>
-<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1410.7716,-234.1725 "/>
-<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1414.7716,-270.1725 "/>
-<text text-anchor="middle" x="1290" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.skip_page</text>
-</g>
-<!-- t6&#45;&gt;t21 -->
-<g id="edge28" class="edge">
-<title>t6&#45;&gt;t21</title>
-<path fill="none" stroke="#0044a0" d="M1269.322,-800.6326C1488.9576,-787.7851 1907,-754.9633 1907,-691.252 1907,-691.252 1907,-691.252 1907,-429.252 1907,-329.7211 1606.6136,-282.8737 1425.0278,-263.647"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1425.1684,-260.1428 1414.8592,-262.586 1424.4419,-267.105 1425.1684,-260.1428"/>
-</g>
-<!-- t19 -->
-<g id="node21" class="node">
-<title>t19</title>
-<polygon fill="#66ccff" stroke="#000000" points="465.6739,-447.252 60.3261,-447.252 56.3261,-443.252 56.3261,-411.252 461.6739,-411.252 465.6739,-415.252 465.6739,-447.252"/>
-<polyline fill="none" stroke="#000000" points="461.6739,-443.252 56.3261,-443.252 "/>
-<polyline fill="none" stroke="#000000" points="461.6739,-443.252 461.6739,-411.252 "/>
-<polyline fill="none" stroke="#000000" points="461.6739,-443.252 465.6739,-447.252 "/>
-<text text-anchor="middle" x="261" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_and_render_pdf</text>
-</g>
-<!-- t6&#45;&gt;t19 -->
-<g id="edge30" class="edge">
-<title>t6&#45;&gt;t19</title>
-<path fill="none" stroke="#0044a0" d="M1006.6775,-802.4744C883.1032,-797.0687 693.0454,-786.4114 529,-767.252 327.1018,-743.6717 78,-894.5226 78,-691.252 78,-691.252 78,-691.252 78,-575.252 78,-532.3545 60.4757,-512.1551 88,-479.252 97.684,-467.6756 124.6714,-457.7191 154.5327,-449.8074"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="155.4872,-453.176 164.3106,-447.311 153.7556,-446.3936 155.4872,-453.176"/>
-</g>
-<!-- t8 -->
-<g id="node9" class="node">
-<title>t8</title>
-<polygon fill="#efa03b" stroke="#000000" points="942.0801,-709.252 517.9199,-709.252 513.9199,-705.252 513.9199,-673.252 938.0801,-673.252 942.0801,-677.252 942.0801,-709.252"/>
-<polyline fill="none" stroke="#000000" points="938.0801,-705.252 513.9199,-705.252 "/>
-<polyline fill="none" stroke="#000000" points="938.0801,-705.252 938.0801,-673.252 "/>
-<polyline fill="none" stroke="#000000" points="938.0801,-705.252 942.0801,-709.252 "/>
-<text text-anchor="middle" x="728" y="-685.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_remove_background</text>
-</g>
-<!-- t7&#45;&gt;t8 -->
-<g id="edge9" class="edge">
-<title>t7&#45;&gt;t8</title>
-<path fill="none" stroke="#0044a0" d="M728,-731.2184C728,-727.5438 728,-723.5938 728,-719.6583"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="731.5001,-719.4742 728,-709.4743 724.5001,-719.4743 731.5001,-719.4742"/>
+<path fill="none" stroke="#0044a0" d="M890.1597,-847.2319C888.6839,-843.3412 887.0909,-839.1417 885.5115,-834.9776"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="888.731,-833.5965 881.9119,-825.4878 882.186,-836.0791 888.731,-833.5965"/>
 </g>
 <!-- t13 -->
 <g id="node16" class="node">
 <title>t13</title>
-<polygon fill="#efa03b" stroke="#000000" points="481,-535.252 97.1874,-507.252 481,-479.252 864.8126,-507.252 481,-535.252"/>
-<text text-anchor="middle" x="481" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_visible_page_image</text>
+<polygon fill="#00cc66" stroke="#000000" points="1077,-515.2656 754.4141,-487.2656 1077,-459.2656 1399.5859,-487.2656 1077,-515.2656"/>
+<text text-anchor="middle" x="1077" y="-481.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_image_layer</text>
 </g>
-<!-- t7&#45;&gt;t13 -->
+<!-- t5&#45;&gt;t13 -->
 <g id="edge20" class="edge">
-<title>t7&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M592.7552,-731.2177C555.8805,-724.9263 521.5464,-717.3855 505,-709.252 470.5499,-692.3178 456.557,-686.3456 441,-651.252 425.2764,-615.7827 441.5685,-572.6191 457.8398,-542.8513"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="460.9365,-544.4864 462.8533,-534.0664 454.8568,-541.0167 460.9365,-544.4864"/>
+<title>t5&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M1000.2704,-847.241C1031.2359,-840.6873 1060.5415,-832.9862 1074,-825.2656 1114.0954,-802.2644 1145,-795.4899 1145,-749.2656 1145,-749.2656 1145,-749.2656 1145,-633.2656 1145,-590.368 1154.8687,-575.2844 1135,-537.2656 1131.571,-530.7041 1126.8893,-524.6314 1121.6622,-519.1451"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1123.8989,-516.4405 1114.2626,-512.0367 1119.0494,-521.4885 1123.8989,-516.4405"/>
+</g>
+<!-- t7 -->
+<g id="node8" class="node">
+<title>t7</title>
+<polygon fill="#efa03b" stroke="#000000" points="856.0801,-767.2656 431.9199,-767.2656 427.9199,-763.2656 427.9199,-731.2656 852.0801,-731.2656 856.0801,-735.2656 856.0801,-767.2656"/>
+<polyline fill="none" stroke="#000000" points="852.0801,-763.2656 427.9199,-763.2656 "/>
+<polyline fill="none" stroke="#000000" points="852.0801,-763.2656 852.0801,-731.2656 "/>
+<polyline fill="none" stroke="#000000" points="852.0801,-763.2656 856.0801,-767.2656 "/>
+<text text-anchor="middle" x="642" y="-743.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_remove_background</text>
+</g>
+<!-- t6&#45;&gt;t7 -->
+<g id="edge8" class="edge">
+<title>t6&#45;&gt;t7</title>
+<path fill="none" stroke="#0044a0" d="M802.5546,-789.2319C777.7884,-783.067 749.9071,-776.1266 724.4138,-769.7806"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="724.95,-766.3073 714.4006,-767.288 723.259,-773.1 724.95,-766.3073"/>
+</g>
+<!-- t12 -->
+<g id="node15" class="node">
+<title>t12</title>
+<polygon fill="#efa03b" stroke="#000000" points="742,-593.2656 358.1874,-565.2656 742,-537.2656 1125.8126,-565.2656 742,-593.2656"/>
+<text text-anchor="middle" x="742" y="-559.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_visible_page_image</text>
+</g>
+<!-- t6&#45;&gt;t12 -->
+<g id="edge18" class="edge">
+<title>t6&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M875.1443,-789.1129C874.6867,-773.2956 872.5831,-750.0189 865,-731.2656 844.2991,-680.0716 804.0126,-630.4165 775.1838,-598.9641"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="777.4937,-596.3085 768.1202,-591.3701 772.3682,-601.076 777.4937,-596.3085"/>
+</g>
+<!-- t8 -->
+<g id="node9" class="node">
+<title>t8</title>
+<polygon fill="#efa03b" stroke="#000000" points="732.6759,-709.2656 413.3241,-709.2656 409.3241,-705.2656 409.3241,-673.2656 728.6759,-673.2656 732.6759,-677.2656 732.6759,-709.2656"/>
+<polyline fill="none" stroke="#000000" points="728.6759,-705.2656 409.3241,-705.2656 "/>
+<polyline fill="none" stroke="#000000" points="728.6759,-705.2656 728.6759,-673.2656 "/>
+<polyline fill="none" stroke="#000000" points="728.6759,-705.2656 732.6759,-709.2656 "/>
+<text text-anchor="middle" x="571" y="-685.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_deskew</text>
+</g>
+<!-- t7&#45;&gt;t8 -->
+<g id="edge9" class="edge">
+<title>t7&#45;&gt;t8</title>
+<path fill="none" stroke="#0044a0" d="M619.9243,-731.2319C613.9709,-726.3685 607.4267,-721.0226 601.115,-715.8666"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="603.2653,-713.1037 593.3065,-709.4878 598.8367,-718.5248 603.2653,-713.1037"/>
+</g>
+<!-- t7&#45;&gt;t12 -->
+<g id="edge17" class="edge">
+<title>t7&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M717.0346,-731.1646C726.8757,-725.6884 735.7103,-718.5557 742,-709.2656 763.0291,-678.205 759.4083,-634.0888 752.8082,-602.8898"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="756.1481,-601.7976 750.487,-592.842 749.3278,-603.3733 756.1481,-601.7976"/>
 </g>
 <!-- t9 -->
 <g id="node10" class="node">
 <title>t9</title>
-<polygon fill="#efa03b" stroke="#000000" points="773.6759,-651.252 454.3241,-651.252 450.3241,-647.252 450.3241,-615.252 769.6759,-615.252 773.6759,-619.252 773.6759,-651.252"/>
-<polyline fill="none" stroke="#000000" points="769.6759,-647.252 450.3241,-647.252 "/>
-<polyline fill="none" stroke="#000000" points="769.6759,-647.252 769.6759,-615.252 "/>
-<polyline fill="none" stroke="#000000" points="769.6759,-647.252 773.6759,-651.252 "/>
-<text text-anchor="middle" x="612" y="-627.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_deskew</text>
+<polygon fill="#efa03b" stroke="#000000" points="712.8828,-651.2656 411.1172,-651.2656 407.1172,-647.2656 407.1172,-615.2656 708.8828,-615.2656 712.8828,-619.2656 712.8828,-651.2656"/>
+<polyline fill="none" stroke="#000000" points="708.8828,-647.2656 407.1172,-647.2656 "/>
+<polyline fill="none" stroke="#000000" points="708.8828,-647.2656 708.8828,-615.2656 "/>
+<polyline fill="none" stroke="#000000" points="708.8828,-647.2656 712.8828,-651.2656 "/>
+<text text-anchor="middle" x="560" y="-627.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_clean</text>
 </g>
 <!-- t8&#45;&gt;t9 -->
 <g id="edge10" class="edge">
 <title>t8&#45;&gt;t9</title>
-<path fill="none" stroke="#0044a0" d="M691.9327,-673.2184C680.9122,-667.7081 668.6526,-661.5783 657.1104,-655.8072"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="658.5545,-652.6162 648.045,-651.2745 655.424,-658.8772 658.5545,-652.6162"/>
+<path fill="none" stroke="#0044a0" d="M567.5798,-673.2319C566.8624,-669.4493 566.0897,-665.3747 565.3216,-661.3247"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="568.7581,-660.6605 563.4559,-651.4878 561.8807,-661.9649 568.7581,-660.6605"/>
 </g>
-<!-- t8&#45;&gt;t13 -->
-<g id="edge19" class="edge">
-<title>t8&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M763.3121,-673.1037C771.2083,-667.2594 778.5531,-659.9847 783,-651.252 790.2604,-636.9942 792.214,-628.3326 783,-615.252 759.3602,-581.6921 662.7714,-550.9641 585.1057,-530.9385"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="585.806,-527.5052 575.2515,-528.4283 584.078,-534.2886 585.806,-527.5052"/>
+<!-- t8&#45;&gt;t12 -->
+<g id="edge16" class="edge">
+<title>t8&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M670.0045,-673.2618C693.2528,-667.2338 713.6919,-659.8307 722,-651.2656 734.1971,-638.6912 739.4901,-620.2051 741.6123,-603.5914"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="745.1041,-603.8427 742.5574,-593.5585 738.1349,-603.1862 745.1041,-603.8427"/>
 </g>
 <!-- t10 -->
 <g id="node11" class="node">
 <title>t10</title>
-<polygon fill="#efa03b" stroke="#000000" points="1117.8828,-593.252 816.1172,-593.252 812.1172,-589.252 812.1172,-557.252 1113.8828,-557.252 1117.8828,-561.252 1117.8828,-593.252"/>
-<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 812.1172,-589.252 "/>
-<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1113.8828,-557.252 "/>
-<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1117.8828,-593.252 "/>
-<text text-anchor="middle" x="965" y="-569.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_clean</text>
+<polygon fill="#efa03b" stroke="#000000" points="340.1094,-583.2656 35.8906,-583.2656 31.8906,-579.2656 31.8906,-547.2656 336.1094,-547.2656 340.1094,-551.2656 340.1094,-583.2656"/>
+<polyline fill="none" stroke="#000000" points="336.1094,-579.2656 31.8906,-579.2656 "/>
+<polyline fill="none" stroke="#000000" points="336.1094,-579.2656 336.1094,-547.2656 "/>
+<polyline fill="none" stroke="#000000" points="336.1094,-579.2656 340.1094,-583.2656 "/>
+<text text-anchor="middle" x="186" y="-559.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_ocr_image</text>
 </g>
 <!-- t9&#45;&gt;t10 -->
 <g id="edge11" class="edge">
 <title>t9&#45;&gt;t10</title>
-<path fill="none" stroke="#0044a0" d="M721.7564,-615.2184C760.9379,-608.7806 805.2658,-601.4973 845.2464,-594.9283"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="846.0113,-598.3496 855.3115,-593.2745 844.8763,-591.4422 846.0113,-598.3496"/>
+<path fill="none" stroke="#0044a0" d="M460.7568,-615.2213C409.7279,-605.9434 347.4237,-594.6153 295.0537,-585.0935"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="295.5223,-581.6214 285.0575,-583.276 294.2701,-588.5085 295.5223,-581.6214"/>
 </g>
-<!-- t9&#45;&gt;t13 -->
-<g id="edge18" class="edge">
-<title>t9&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M593.2309,-615.1993C573.0988,-595.8356 540.6771,-564.6514 515.8453,-540.7673"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="518.0381,-538.0203 508.4046,-533.6106 513.1856,-543.0654 518.0381,-538.0203"/>
+<!-- t9&#45;&gt;t12 -->
+<g id="edge15" class="edge">
+<title>t9&#45;&gt;t12</title>
+<path fill="none" stroke="#0044a0" d="M608.2948,-615.2213C627.0226,-608.2242 648.8712,-600.0609 669.4134,-592.3859"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="670.9148,-595.5613 679.0573,-588.7826 668.4648,-589.004 670.9148,-595.5613"/>
 </g>
 <!-- t11 -->
 <g id="node12" class="node">
 <title>t11</title>
-<polygon fill="#efa03b" stroke="#000000" points="1191.1094,-525.252 886.8906,-525.252 882.8906,-521.252 882.8906,-489.252 1187.1094,-489.252 1191.1094,-493.252 1191.1094,-525.252"/>
-<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 882.8906,-521.252 "/>
-<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1187.1094,-489.252 "/>
-<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1191.1094,-525.252 "/>
-<text text-anchor="middle" x="1037" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_ocr_image</text>
+<polygon fill="#00cc66" stroke="#000000" points="334.0996,-505.2656 19.9004,-505.2656 15.9004,-501.2656 15.9004,-469.2656 330.0996,-469.2656 334.0996,-473.2656 334.0996,-505.2656"/>
+<polyline fill="none" stroke="#000000" points="330.0996,-501.2656 15.9004,-501.2656 "/>
+<polyline fill="none" stroke="#000000" points="330.0996,-501.2656 330.0996,-469.2656 "/>
+<polyline fill="none" stroke="#000000" points="330.0996,-501.2656 334.0996,-505.2656 "/>
+<text text-anchor="middle" x="175" y="-481.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
 </g>
 <!-- t10&#45;&gt;t11 -->
 <g id="edge12" class="edge">
 <title>t10&#45;&gt;t11</title>
-<path fill="none" stroke="#0044a0" d="M984.2947,-557.0293C992.2846,-549.4832 1001.6851,-540.6049 1010.318,-532.4517"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1012.7851,-534.9359 1017.6521,-525.5251 1007.9787,-529.8468 1012.7851,-534.9359"/>
-</g>
-<!-- t10&#45;&gt;t13 -->
-<g id="edge17" class="edge">
-<title>t10&#45;&gt;t13</title>
-<path fill="none" stroke="#0044a0" d="M836.5676,-557.2078C770.772,-547.9638 690.4915,-536.6847 622.8731,-527.1846"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="622.9531,-523.6616 612.5634,-525.7361 621.9792,-530.5935 622.9531,-523.6616"/>
-</g>
-<!-- t12 -->
-<g id="node13" class="node">
-<title>t12</title>
-<polygon fill="#00cc66" stroke="#000000" points="822.0996,-447.252 507.9004,-447.252 503.9004,-443.252 503.9004,-411.252 818.0996,-411.252 822.0996,-415.252 822.0996,-447.252"/>
-<polyline fill="none" stroke="#000000" points="818.0996,-443.252 503.9004,-443.252 "/>
-<polyline fill="none" stroke="#000000" points="818.0996,-443.252 818.0996,-411.252 "/>
-<polyline fill="none" stroke="#000000" points="818.0996,-443.252 822.0996,-447.252 "/>
-<text text-anchor="middle" x="663" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
-</g>
-<!-- t11&#45;&gt;t12 -->
-<g id="edge13" class="edge">
-<title>t11&#45;&gt;t12</title>
-<path fill="none" stroke="#0044a0" d="M950.2568,-489.1612C893.2425,-477.2705 818.5749,-461.6981 759.6569,-449.4104"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="760.1,-445.9276 749.5961,-447.3122 758.6708,-452.7802 760.1,-445.9276"/>
-</g>
-<!-- t11&#45;&gt;t17 -->
-<g id="edge15" class="edge">
-<title>t11&#45;&gt;t17</title>
-<path fill="none" stroke="#0044a0" d="M1035.8403,-489.1612C1035.2388,-479.7779 1034.4904,-468.1019 1033.8144,-457.5566"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1037.3033,-457.2705 1033.1707,-447.5149 1030.3177,-457.7184 1037.3033,-457.2705"/>
+<path fill="none" stroke="#0044a0" d="M183.4487,-547.1748C182.1254,-537.7914 180.4788,-526.1154 178.9917,-515.5701"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="182.4378,-514.9417 177.5755,-505.5285 175.5064,-515.9193 182.4378,-514.9417"/>
 </g>
 <!-- t15 -->
 <g id="node14" class="node">
 <title>t15</title>
-<polygon fill="#00cc66" stroke="#000000" points="1229.0433,-353.7123 920.9567,-353.7123 916.9567,-349.7123 916.9567,-317.7123 1225.0433,-317.7123 1229.0433,-321.7123 1229.0433,-353.7123"/>
-<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 916.9567,-349.7123 "/>
-<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1225.0433,-317.7123 "/>
-<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1229.0433,-353.7123 "/>
-<text text-anchor="middle" x="1073" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_page</text>
+<polygon fill="#ff69b4" stroke="#000000" points="736.2696,-505.2656 355.7304,-505.2656 351.7304,-501.2656 351.7304,-469.2656 732.2696,-469.2656 736.2696,-473.2656 736.2696,-505.2656"/>
+<polyline fill="none" stroke="#000000" points="732.2696,-501.2656 351.7304,-501.2656 "/>
+<polyline fill="none" stroke="#000000" points="732.2696,-501.2656 732.2696,-469.2656 "/>
+<polyline fill="none" stroke="#000000" points="732.2696,-501.2656 736.2696,-505.2656 "/>
+<text text-anchor="middle" x="544" y="-481.2656" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
 </g>
-<!-- t12&#45;&gt;t15 -->
+<!-- t10&#45;&gt;t15 -->
 <g id="edge14" class="edge">
-<title>t12&#45;&gt;t15</title>
-<path fill="none" stroke="#0044a0" d="M754.3855,-411.2269C799.2649,-402.117 854.0781,-390.6067 903,-379.252 933.2642,-372.2277 966.3927,-363.8908 995.3012,-356.4015"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="996.5416,-359.6955 1005.3389,-353.7914 994.7799,-352.9208 996.5416,-359.6955"/>
+<title>t10&#45;&gt;t15</title>
+<path fill="none" stroke="#0044a0" d="M269.0322,-547.1748C323.4947,-535.3086 394.7854,-519.776 451.1284,-507.5002"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="452.0829,-510.8744 461.1086,-505.3257 450.5927,-504.0348 452.0829,-510.8744"/>
 </g>
-<!-- t16 -->
-<g id="node19" class="node">
-<title>t16</title>
-<polygon fill="#00cc66" stroke="#000000" points="651.9202,-270.1725 284.0798,-270.1725 280.0798,-266.1725 280.0798,-234.1725 647.9202,-234.1725 651.9202,-238.1725 651.9202,-270.1725"/>
-<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 280.0798,-266.1725 "/>
-<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 647.9202,-234.1725 "/>
-<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 651.9202,-270.1725 "/>
-<text text-anchor="middle" x="466" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_debug_page</text>
+<!-- t14 -->
+<g id="node13" class="node">
+<title>t14</title>
+<polygon fill="#00cc66" stroke="#000000" points="330.0433,-411.7258 21.9567,-411.7258 17.9567,-407.7258 17.9567,-375.7258 326.0433,-375.7258 330.0433,-379.7258 330.0433,-411.7258"/>
+<polyline fill="none" stroke="#000000" points="326.0433,-407.7258 17.9567,-407.7258 "/>
+<polyline fill="none" stroke="#000000" points="326.0433,-407.7258 326.0433,-375.7258 "/>
+<polyline fill="none" stroke="#000000" points="326.0433,-407.7258 330.0433,-411.7258 "/>
+<text text-anchor="middle" x="174" y="-387.7258" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_page</text>
 </g>
-<!-- t12&#45;&gt;t16 -->
-<g id="edge27" class="edge">
-<title>t12&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M772.353,-411.2135C825.7002,-401.3017 880.0px85,-389.0648 888,-379.252 912.2169,-349.0628 914.0433,-320.8011 888,-292.1725 871.9898,-274.5731 764.4698,-264.5473 661.9779,-258.9344"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="661.9319,-255.4271 651.7593,-258.3883 661.5583,-262.4171 661.9319,-255.4271"/>
+<!-- t11&#45;&gt;t14 -->
+<g id="edge13" class="edge">
+<title>t11&#45;&gt;t14</title>
+<path fill="none" stroke="#0044a0" d="M174.8072,-469.2302C174.6642,-455.8584 174.4666,-437.3702 174.3027,-422.0357"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="177.8024,-421.987 174.1956,-412.025 170.8028,-422.0619 177.8024,-421.987"/>
 </g>
-<!-- t25 -->
-<g id="node27" class="node">
-<title>t25</title>
-<polygon fill="#efa03b" stroke="#000000" points="488.7205,-323.495 682,-296.1762 875.2795,-323.495 875.0991,-367.6976 488.9009,-367.6976 488.7205,-323.495"/>
-<polygon fill="none" stroke="#000000" points="484.7033,-320.0193 682,-292.1327 879.2967,-320.0193 879.0858,-371.7013 484.9142,-371.7013 484.7033,-320.0193"/>
-<text text-anchor="middle" x="682" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_sidecars</text>
+<!-- t19 -->
+<g id="node22" class="node">
+<title>t19</title>
+<polygon fill="#efa03b" stroke="#000000" points="351.7205,-381.5085 545,-354.1897 738.2795,-381.5085 738.0991,-425.7112 351.9009,-425.7112 351.7205,-381.5085"/>
+<polygon fill="none" stroke="#000000" points="347.7033,-378.0329 545,-350.1463 742.2967,-378.0329 742.0858,-429.7148 347.9142,-429.7148 347.7033,-378.0329"/>
+<text text-anchor="middle" x="545" y="-387.7258" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_sidecars</text>
 </g>
-<!-- t12&#45;&gt;t25 -->
-<g id="edge52" class="edge">
-<title>t12&#45;&gt;t25</title>
-<path fill="none" stroke="#0044a0" d="M666.6634,-411.2166C668.3795,-402.768 670.5105,-392.2769 672.6326,-381.8293"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="676.09,-382.3905 674.6507,-371.8939 669.2301,-380.9971 676.09,-382.3905"/>
+<!-- t11&#45;&gt;t19 -->
+<g id="edge32" class="edge">
+<title>t11&#45;&gt;t19</title>
+<path fill="none" stroke="#0044a0" d="M246.3397,-469.2302C287.966,-458.7066 342.1269,-445.0142 392.8772,-432.184"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="393.8058,-435.5594 402.6429,-429.7151 392.0901,-428.773 393.8058,-435.5594"/>
 </g>
-<!-- t18 -->
-<g id="node18" class="node">
-<title>t18</title>
-<polygon fill="#00cc66" stroke="#000000" points="1147.2989,-270.1725 856.7011,-270.1725 852.7011,-266.1725 852.7011,-234.1725 1143.2989,-234.1725 1147.2989,-238.1725 1147.2989,-270.1725"/>
-<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 852.7011,-266.1725 "/>
-<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1143.2989,-234.1725 "/>
-<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1147.2989,-270.1725 "/>
-<text text-anchor="middle" x="1000" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.combine_layers</text>
+<!-- t14&#45;&gt;t16 -->
+<g id="edge22" class="edge">
+<title>t14&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M230.6615,-375.7128C262.2207,-366.4581 302.4172,-355.9425 339,-350.1861 448.3957,-332.9725 745.7869,-320.7866 927.9215,-314.6743"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="928.248,-318.1655 938.1258,-314.3342 928.0148,-311.1693 928.248,-318.1655"/>
 </g>
-<!-- t15&#45;&gt;t18 -->
-<g id="edge24" class="edge">
-<title>t15&#45;&gt;t18</title>
-<path fill="none" stroke="#0044a0" d="M1057.1601,-317.5854C1047.0201,-305.9814 1033.7081,-290.7474 1022.4822,-277.9007"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1025.0247,-275.4911 1015.809,-270.264 1019.7536,-280.0972 1025.0247,-275.4911"/>
-</g>
-<!-- t17&#45;&gt;t18 -->
-<g id="edge23" class="edge">
-<title>t17&#45;&gt;t18</title>
-<path fill="none" stroke="#0044a0" d="M1160.5767,-411.2248C1195.7121,-403.5317 1227.1378,-393.066 1238,-379.252 1261.9222,-348.8288 1262.9959,-321.72 1238,-292.1725 1231.6408,-284.6553 1197.8632,-277.2656 1157.6598,-270.9231"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1158.0117,-267.4362 1147.596,-269.3761 1156.9481,-274.3549 1158.0117,-267.4362"/>
-</g>
-<!-- t17&#45;&gt;t25 -->
-<g id="edge51" class="edge">
-<title>t17&#45;&gt;t25</title>
-<path fill="none" stroke="#0044a0" d="M964.5165,-411.2166C925.304,-400.7368 874.3329,-387.1145 826.4989,-374.3305"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="827.2267,-370.9023 816.6621,-371.7016 825.4193,-377.6649 827.2267,-370.9023"/>
-</g>
-<!-- t13&#45;&gt;t14 -->
+<!-- t15&#45;&gt;t16 -->
 <g id="edge21" class="edge">
-<title>t13&#45;&gt;t14</title>
-<path fill="none" stroke="#0044a0" d="M672.8683,-493.1993C736.8314,-488.6166 808.4777,-483.5937 874,-479.252 1033.5051,-468.6828 1073.5671,-468.8587 1233,-457.252 1285.8429,-453.4051 1343.3743,-448.7264 1395.4741,-444.3037"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1395.81,-447.7879 1405.477,-443.4521 1395.2161,-440.8131 1395.81,-447.7879"/>
+<title>t15&#45;&gt;t16</title>
+<path fill="none" stroke="#0044a0" d="M630.0636,-469.2083C667.5329,-460.6467 711.76,-449.5912 751,-437.2656 850.4985,-406.0121 963.1653,-359.5222 1026.586,-332.2894"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1028.1566,-335.4238 1035.9553,-328.2523 1025.3866,-328.9951 1028.1566,-335.4238"/>
+</g>
+<!-- t15&#45;&gt;t19 -->
+<g id="edge31" class="edge">
+<title>t15&#45;&gt;t19</title>
+<path fill="none" stroke="#0044a0" d="M544.1928,-469.2302C544.2822,-460.8687 544.393,-450.5066 544.5035,-440.166"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="548.006,-439.9443 544.6132,-429.9075 541.0064,-439.8694 548.006,-439.9443"/>
+</g>
+<!-- t12&#45;&gt;t13 -->
+<g id="edge19" class="edge">
+<title>t12&#45;&gt;t13</title>
+<path fill="none" stroke="#0044a0" d="M833.9833,-543.8486C879.0975,-533.3444 933.4556,-520.6879 979.3116,-510.0109"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="980.3156,-513.3709 989.2614,-507.6943 978.7282,-506.5532 980.3156,-513.3709"/>
 </g>
 <!-- t13&#45;&gt;t16 -->
-<g id="edge26" class="edge">
+<g id="edge23" class="edge">
 <title>t13&#45;&gt;t16</title>
-<path fill="none" stroke="#0044a0" d="M479.4495,-479.3529C478.2545,-457.9761 476.5497,-427.7492 475,-401.252 472.546,-359.2945 469.5914,-310.7019 467.754,-280.6981"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="471.2362,-280.2988 467.1306,-270.5318 464.2493,-280.7273 471.2362,-280.2988"/>
+<path fill="none" stroke="#0044a0" d="M1077,-459.2068C1077,-426.2624 1077,-371.9187 1077,-338.7552"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1080.5001,-338.3818 1077,-328.3819 1073.5001,-338.3819 1080.5001,-338.3818"/>
 </g>
-<!-- t13&#45;&gt;t19 -->
+<!-- t16&#45;&gt;t18 -->
+<g id="edge27" class="edge">
+<title>t16&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1117.3592,-292.0592C1133.8695,-284.6437 1153.6803,-275.7459 1173.2953,-266.9361"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1174.8979,-270.0531 1182.586,-262.7632 1172.0299,-263.6676 1174.8979,-270.0531"/>
+</g>
+<!-- t17&#45;&gt;t18 -->
+<g id="edge26" class="edge">
+<title>t17&#45;&gt;t18</title>
+<path fill="none" stroke="#0044a0" d="M1326.2292,-973.7283C1377.8362,-955.768 1439,-922.4247 1439,-865.2656 1439,-865.2656 1439,-865.2656 1439,-393.7258 1439,-338.7092 1393.2056,-296.3464 1348.3356,-268.0433"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1350.0403,-264.9831 1339.6842,-262.7465 1346.3852,-270.9531 1350.0403,-264.9831"/>
+</g>
+<!-- t20 -->
+<g id="node20" class="node">
+<title>t20</title>
+<polygon fill="#efa03b" stroke="#000000" points="1400.3184,-161.1062 1129.6816,-161.1062 1125.6816,-157.1062 1125.6816,-125.1062 1396.3184,-125.1062 1400.3184,-129.1062 1400.3184,-161.1062"/>
+<polyline fill="none" stroke="#000000" points="1396.3184,-157.1062 1125.6816,-157.1062 "/>
+<polyline fill="none" stroke="#000000" points="1396.3184,-157.1062 1396.3184,-125.1062 "/>
+<polyline fill="none" stroke="#000000" points="1396.3184,-157.1062 1400.3184,-161.1062 "/>
+<text text-anchor="middle" x="1263" y="-137.1062" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.optimize_pdf</text>
+</g>
+<!-- t18&#45;&gt;t20 -->
 <g id="edge29" class="edge">
-<title>t13&#45;&gt;t19</title>
-<path fill="none" stroke="#0044a0" d="M415.2907,-483.9551C385.6007,-473.4286 350.7427,-461.0699 321.6782,-450.7652"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="322.6542,-447.3978 312.0595,-447.3549 320.315,-453.9954 322.6542,-447.3978"/>
+<title>t18&#45;&gt;t20</title>
+<path fill="none" stroke="#0044a0" d="M1263,-182.9188C1263,-179.129 1263,-175.3607 1263,-171.739"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1266.5001,-171.4394 1263,-161.4394 1259.5001,-171.4394 1266.5001,-171.4394"/>
 </g>
-<!-- t14&#45;&gt;t18 -->
-<g id="edge25" class="edge">
-<title>t14&#45;&gt;t18</title>
-<path fill="none" stroke="#0044a0" d="M1519.9776,-405.138C1460.3688,-374.2505 1351.1978,-321.2463 1252,-292.1725 1221.8888,-283.3473 1189.1477,-276.3434 1157.5373,-270.8287"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1157.7641,-267.3171 1147.3187,-269.0901 1156.59,-274.2179 1157.7641,-267.3171"/>
+<!-- t21 -->
+<g id="node21" class="node">
+<title>t21</title>
+<polygon fill="#efa03b" stroke="#000000" points="1096.0652,-47.332 1263,-20.0049 1429.9348,-47.332 1429.779,-91.5483 1096.221,-91.5483 1096.0652,-47.332"/>
+<polygon fill="none" stroke="#000000" points="1092.0538,-43.9306 1263,-15.9468 1433.9462,-43.9306 1433.7643,-95.5532 1092.2357,-95.5532 1092.0538,-43.9306"/>
+<text text-anchor="middle" x="1263" y="-53.5531" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.copy_final</text>
 </g>
-<!-- t22 -->
-<g id="node23" class="node">
-<title>t22</title>
-<polygon fill="#efa03b" stroke="#000000" points="489.918,-156.4263 735,-129.1169 980.082,-156.4263 979.8532,-200.6137 490.1468,-200.6137 489.918,-156.4263"/>
-<polygon fill="none" stroke="#000000" points="485.9027,-152.8467 735,-125.09 984.0973,-152.8467 983.8293,-204.6159 486.1707,-204.6159 485.9027,-152.8467"/>
-<text text-anchor="middle" x="735" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_ghostscript</text>
-</g>
-<!-- t18&#45;&gt;t22 -->
-<g id="edge35" class="edge">
-<title>t18&#45;&gt;t22</title>
-<path fill="none" stroke="#0044a0" d="M942.8252,-234.1499C918.2191,-226.3936 888.4232,-217.0013 859.1708,-207.7804"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="859.8692,-204.3309 849.2796,-204.6625 857.7647,-211.007 859.8692,-204.3309"/>
-</g>
-<!-- t18&#45;&gt;t24 -->
-<g id="edge39" class="edge">
-<title>t18&#45;&gt;t24</title>
-<path fill="none" stroke="#0044a0" d="M852.4716,-245.1051C747.4645,-239.0474 603.1919,-228.6387 477,-212.1725 463.5912,-210.4229 449.788,-208.4132 435.908,-206.238"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="436.2183,-202.7434 425.7921,-204.6263 435.1168,-209.6562 436.2183,-202.7434"/>
-</g>
-<!-- t18&#45;&gt;t23 -->
-<g id="edge44" class="edge">
-<title>t18&#45;&gt;t23</title>
-<path fill="none" stroke="#0044a0" d="M1047.0344,-234.1499C1066.9243,-226.5285 1090.9364,-217.3275 1114.596,-208.2617"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1115.9033,-211.509 1123.9889,-204.6625 1113.3986,-204.9724 1115.9033,-211.509"/>
-</g>
-<!-- t16&#45;&gt;t22 -->
-<g id="edge36" class="edge">
-<title>t16&#45;&gt;t22</title>
-<path fill="none" stroke="#0044a0" d="M524.0378,-234.1499C549.1239,-226.3599 579.5241,-216.9196 609.3422,-207.6602"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="610.4833,-210.9707 618.9955,-204.6625 608.4073,-204.2857 610.4833,-210.9707"/>
-</g>
-<!-- t16&#45;&gt;t24 -->
-<g id="edge40" class="edge">
-<title>t16&#45;&gt;t24</title>
-<path fill="none" stroke="#0044a0" d="M417.6711,-234.1499C397.2337,-226.5285 372.5607,-217.3275 348.2499,-208.2617"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="349.1912,-204.8773 338.5986,-204.6625 346.7453,-211.4361 349.1912,-204.8773"/>
-</g>
-<!-- t16&#45;&gt;t23 -->
-<g id="edge45" class="edge">
-<title>t16&#45;&gt;t23</title>
-<path fill="none" stroke="#0044a0" d="M651.7596,-235.3105C656.5547,-234.9215 661.308,-234.5415 666,-234.1725 811.2137,-222.7545 848.5512,-230.9322 993,-212.1725 1006.176,-210.4614 1019.7432,-208.4558 1033.3753,-206.2619"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1034.0068,-209.7051 1043.3083,-204.6325 1032.8736,-202.7974 1034.0068,-209.7051"/>
-</g>
-<!-- t21&#45;&gt;t22 -->
-<g id="edge32" class="edge">
-<title>t21&#45;&gt;t22</title>
-<path fill="none" stroke="#0044a0" d="M1165.0941,-235.3973C1085.9035,-224.7541 997.501,-212.8552 993,-212.1725 980.474,-210.2728 967.615,-208.2703 954.6442,-206.2099"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="955.1339,-202.7439 944.7072,-204.6238 954.0305,-209.6564 955.1339,-202.7439"/>
-</g>
-<!-- t21&#45;&gt;t24 -->
-<g id="edge37" class="edge">
-<title>t21&#45;&gt;t24</title>
-<path fill="none" stroke="#0044a0" d="M1165.2215,-234.9906C1162.1187,-234.699 1159.041,-234.4254 1156,-234.1725 855.1023,-209.1529 777.1841,-244.6523 477,-212.1725 462.026,-210.5524 446.5737,-208.5164 431.0798,-206.2151"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="431.3094,-202.7099 420.8968,-204.6663 430.2567,-209.6303 431.3094,-202.7099"/>
-</g>
-<!-- t21&#45;&gt;t23 -->
-<g id="edge42" class="edge">
-<title>t21&#45;&gt;t23</title>
-<path fill="none" stroke="#0044a0" d="M1274.3771,-234.0471C1268.8557,-227.6412 1262.3807,-220.0px91 1255.8248,-212.5231"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1258.4543,-210.2128 1249.2743,-204.9233 1253.152,-214.783 1258.4543,-210.2128"/>
-</g>
-<!-- t19&#45;&gt;t22 -->
-<g id="edge34" class="edge">
-<title>t19&#45;&gt;t22</title>
-<path fill="none" stroke="#0044a0" d="M260.6159,-410.9505C259.8316,-364.871 259.0249,-246.8273 271,-234.1725 302.6435,-200.7331 431.3919,-218.4498 477,-212.1725 490.4888,-210.316 504.3558,-208.2974 518.3278,-206.1801"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="519.1602,-209.5936 528.5161,-204.6219 518.1019,-202.6741 519.1602,-209.5936"/>
-</g>
-<!-- t19&#45;&gt;t24 -->
-<g id="edge38" class="edge">
-<title>t19&#45;&gt;t24</title>
-<path fill="none" stroke="#0044a0" d="M259.0498,-411.2136C258.0595,-401.7112 256.8826,-389.8585 256,-379.252 251.2976,-322.7418 247.1648,-257.5963 244.6312,-214.9019"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="248.1192,-214.5935 244.0377,-204.8165 241.1313,-215.0048 248.1192,-214.5935"/>
-</g>
-<!-- t19&#45;&gt;t23 -->
-<g id="edge43" class="edge">
-<title>t19&#45;&gt;t23</title>
-<path fill="none" stroke="#0044a0" d="M282.1528,-411.0561C318.8331,-380.6455 397.8323,-320.0855 476,-292.1725 553.979,-264.3269 579.2078,-283.062 661,-270.1725 742.8817,-257.269 762.3086,-248.2304 844,-234.1725 909.9705,-222.82 927.0119,-223.4221 993,-212.1725 1003.5851,-210.368 1014.439,-208.4601 1025.3874,-206.4907"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="1026.2474,-209.892 1035.4637,-204.666 1025,-203.004 1026.2474,-209.892"/>
-</g>
-<!-- t19&#45;&gt;t25 -->
-<g id="edge50" class="edge">
-<title>t19&#45;&gt;t25</title>
-<path fill="none" stroke="#0044a0" d="M342.173,-411.2166C389.8324,-400.6274 451.9326,-386.8297 509.989,-373.9305"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="511.0179,-377.2873 520.0207,-371.7016 509.4996,-370.4539 511.0179,-377.2873"/>
-</g>
-<!-- t20&#45;&gt;t22 -->
-<g id="edge33" class="edge">
-<title>t20&#45;&gt;t22</title>
-<path fill="none" stroke="#0044a0" d="M323.4472,-984.1943C198.6099,-964.2959 37,-926.6746 37,-865.252 37,-865.252 37,-865.252 37,-429.252 37,-211.1917 260.8906,-241.2767 477,-212.1725 490.808,-210.313 505.0108,-208.2741 519.3157,-206.1255"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="519.8697,-209.5816 529.2317,-204.6214 518.8199,-202.6608 519.8697,-209.5816"/>
-</g>
-<!-- t26 -->
-<g id="node26" class="node">
-<title>t26</title>
-<polygon fill="#efa03b" stroke="#000000" points="568.0652,-47.332 735,-20.0049 901.9348,-47.332 901.779,-91.5483 568.221,-91.5483 568.0652,-47.332"/>
-<polygon fill="none" stroke="#000000" points="564.0538,-43.9306 735,-15.9468 905.9462,-43.9306 905.7643,-95.5532 564.2357,-95.5532 564.0538,-43.9306"/>
-<text text-anchor="middle" x="735" y="-53.5531" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.copy_final</text>
-</g>
-<!-- t22&#45;&gt;t26 -->
-<g id="edge49" class="edge">
-<title>t22&#45;&gt;t26</title>
-<path fill="none" stroke="#0044a0" d="M735,-124.8984C735,-118.6708 735,-112.2475 735,-105.9711"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="738.5001,-105.6074 735,-95.6074 731.5001,-105.6075 738.5001,-105.6074"/>
-</g>
-<!-- t24&#45;&gt;t26 -->
-<g id="edge47" class="edge">
-<title>t24&#45;&gt;t26</title>
-<path fill="none" stroke="#0044a0" d="M368.6395,-140.6178C427.4047,-127.6148 498.1821,-111.9539 561.7955,-97.8781"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="562.9329,-101.2112 571.9406,-95.6333 561.4205,-94.3765 562.9329,-101.2112"/>
-</g>
-<!-- t23&#45;&gt;t26 -->
-<g id="edge48" class="edge">
-<title>t23&#45;&gt;t26</title>
-<path fill="none" stroke="#0044a0" d="M1095.1818,-140.9007C1037.3475,-127.8387 967.4279,-112.0473 904.6388,-97.8663"/>
-<polygon fill="#0044a0" stroke="#0044a0" points="905.1511,-94.3939 894.6256,-95.6048 903.6089,-101.2219 905.1511,-94.3939"/>
+<!-- t20&#45;&gt;t21 -->
+<g id="edge30" class="edge">
+<title>t20&#45;&gt;t21</title>
+<path fill="none" stroke="#0044a0" d="M1263,-124.9764C1263,-119.37 1263,-112.9163 1263,-106.2938"/>
+<polygon fill="#0044a0" stroke="#0044a0" points="1266.5001,-105.8456 1263,-95.8457 1259.5001,-105.8457 1266.5001,-105.8456"/>
 </g>
 </g>
 </svg>

docs/release_notes.rst

@@ -1,14 +1,232 @@
 Release notes
 =============
 
-OCRmyPDF uses `semantic versioning <http://semver.org/>`_ for its command line interface.
+OCRmyPDF uses `semantic versioning <http://semver.org/>`_ for its command line interface and its public API.
 
-The OCRmyPDF package itself does not contain a public API, although it is fairly stable and breaking changes are usually timed with a major release. A future release will clearly define the stable public API.
+The ``ocrmypdf`` package may now be imported. The public API may be useful in scripts that launch OCRmyPDF processes or that wish to use some of its features for working with PDFs.
+
+Unfortunately, the public API does **not** expose the ability to actually OCR a PDF. This is due to a limitation in an underlying library (ruffus) that makes OCRmyPDF non-reentrant.
+
+Note that it is licensed under GPLv3, so scripts that ``import ocrmypdf`` and are released publicly should probably also be licensed under GPLv3.
 
 .. Issue regex
-   find:    [^`]\#([0-9]{1,3})[^0-9]  
+   find:    [^`]\#([0-9]{1,3})[^0-9]
    replace: `#$1 <https://github.com/jbarlow83/OCRmyPDF/issues/$1>`_
 
+
+v7.3.1
+------
+
+-   Fixed performance regression from v7.3.0; fast page analysis was not selected when it should be.
+
+-   Fixed a few exceptions related to the new ``--mask-barcodes`` feature and improved argument checking
+
+-   Added missing detection of TrueType fonts that lack a Unicode mapping
+
+
+v7.3.0
+------
+
+-   Added a new feature ``--redo-ocr`` to detect existing OCR in a file, remove it, and redo the OCR. This may be particularly helpful for anyone who wants to take advantage of OCR quality improvements in Tesseract 4.0. Note that OCR added by OCRmyPDF before version 3.0 cannot be detected since it was not properly marked as invisible text in the earliest versions. OCR that constructs a font from visible text, such as Adobe Acrobat's ClearScan.
+
+-   OCRmyPDF's content detection is generally more sophisticated. It learns more about the contents of each PDF and makes better recommendations:
+
+    -   OCRmyPDF can now detect when a PDF contains text that cannot be mapped to Unicode (meaning it is readable to human eyes but copy-pastes as gibberish). In these cases it recommends ``--force-ocr`` to make the text searchable.
+
+    -   PDFs containing vector objects are now rendered at more appropriate resolution for OCR.
+
+    -   We now exit with an error for PDFs that contain Adobe LiveCycle Designer's dynamic XFA forms. Currently the open source community does not have tools to work with these files.
+
+    -   OCRmyPDF now warns when a PDF that contains Adobe AcroForms, since such files probably do not need OCR. It can work with these files.
+
+-   Added three new **experimental** features to improve OCR quality in certain conditions. The name, syntax and behavior of these arguments is subject to change. They may also be incompatible with some other features.
+
+    -   ``--remove-vectors`` which strips out vector graphics. This can improve OCR quality since OCR will not search artwork for readable text; however, it currently removes "text as curves" as well.
+
+    -   ``--mask-barcodes`` to detect and suppress barcodes in files. We have observed that barcodes can interfere with OCR because they are "text-like" but not actually textual.
+
+    -   ``--threshold`` which uses a more sophisticated thresholding algorithm than is currently in use in Tesseract OCR. This works around a `known issue in Tesseract 4.0 <https://github.com/tesseract-ocr/tesseract/issues/1990>`_ with dark text on bright backgrounds.
+
+-   Fixed an issue where an error message was not reported when the installed Ghostscript was very old.
+
+-   The PDF optimizer now saves files with object streams enabled when the optimization level is ``--optimize 1`` or higher (the default). This makes files a little bit smaller, but requires PDF 1.5. PDF 1.5 was first released in 2003 and is broadly supported by PDF viewers, but some rudimentary PDF parsers such as PyPDF2 do not understand object streams. You can use the command line tool ``qpdf --object-streams=disable`` or `pikepdf <https://github.com/pikepdf/pikepdf>`_ library to remove them.
+
+-   New dependency: pdfminer.six 20181108. Note this is a fork of the Python 2-only pdfminer.
+
+-   Deprecation notice: At the end of 2018, we will be ending support for Python 3.5 and Tesseract 3.x. OCRmyPDF v7 will continue to work with older versions.
+
+v7.2.1
+------
+
+-   Fix compatibility with an API change in pikepdf 0.3.5.
+
+-   A kludge to support Leptonica versions older than 1.72 in the test suite was dropped. Older versions of Leptonica are likely still compatible. The only impact is that a portion of the test suite will be skipped.
+
+
+v7.2.0
+------
+
+**Lossy JBIG2 behavior change**
+
+A user reported that ocrmypdf was in fact using JBIG2 in **lossy** compression mode. This was not the intended behavior. Users should `review the technical concerns with JBIG2 in lossy mode <https://abbyy.technology/en:kb:tip:jbig2_compression_and_ocr>`_ and decide if this is a concern for their use case.
+
+JBIG2 lossy mode does achieve higher compression ratios than any other monochrome compression technology; for large text documents the savings are considerable. JBIG2 lossless still gives great compression ratios and is a major improvement over the older CCITT G4 standard.
+
+Only users who have reviewed the concerns with JBIG2 in lossy mode should opt-in. As such, lossy mode JBIG2 is only turned on when the new argument ``--jbig2-lossy`` is issued. This is independent of the setting for ``--optimize``.
+
+Users who did not install an optional JBIG2 encoder are unaffected.
+
+(Thanks to user 'bsdice' for reporting this issue.)
+
+**Other issues**
+
+-   When the image optimizer quantizes an image to 1 bit per pixel, it will now attempt to further optimize that image as CCITT or JBIG2, instead of keeping it in the "flate" encoding which is not efficient for 1 bpp images. (`#297 <https://github.com/jbarlow83/OCRmyPDF/issues/297>`_)
+
+-   Images in PDFs that are used as soft masks (i.e. transparency masks or alpha channels) are now excluded from optimization.
+
+-   Fixed handling of Tesseract 4.0-rc1 which now accepts invalid Tesseract configuration files, which broke the test suite.
+
+v7.1.0
+------
+
+-   Improve the performance of initial text extraction, which is done to determine if a file contains existing text of some kind or not. On large files, this initial processing is now about 20x times faster. (`#299 <https://github.com/jbarlow83/OCRmyPDF/issues/299>`_)
+
+-   pikepdf 0.3.3 is now required.
+
+-   Fixed issue `#231 <https://github.com/jbarlow83/OCRmyPDF/issues/231>`_, a problem with JPEG2000 images where image metadata was only available inside the JPEG2000 file.
+
+-   Fixed some additional Ghostscript 9.25 compatibility issues.
+
+-   Improved handling of KeyboardInterrupt error messages. (`#301 <https://github.com/jbarlow83/OCRmyPDF/issues/301>`_)
+
+-   README.md is now served in GitHub markdown instead of reStructuredText.
+
+v7.0.6
+------
+
+-   Blacklist Ghostscript 9.24, now that 9.25 is available and fixes many regressions in 9.24.
+
+
+v7.0.5
+------
+
+-   Improve capability with Ghostscript 9.24, and enable the JPEG passthrough feature when this version in installed.
+
+-   Ghostscript 9.24 lost the ability to set PDF title, author, subject and keyword metadata to Unicode strings. OCRmyPDF will set ASCII strings and warn when Unicode is suppressed. Other software may be used to update metadata. This is a short term work around.
+
+-   PDFs generated by Kodak Capture Desktop, or generally PDFs that contain indirect references to null objects in their table of contents, would have an invalid table of contents after processing by OCRmyPDF that might interfere with other viewers. This has been fixed.
+
+-   Detect PDFs generated by Adobe LiveCycle, which can only be displayed in Adobe Acrobat and Reader currently. When these are encountered, exit with an error instead of performing OCR on the "Please wait" error message page.
+
+v7.0.4
+------
+
+-   Fix exception thrown when trying to optimize a certain type of PNG embedded in a PDF with the ``-O2``
+
+-   Update to pikepdf 0.3.2, to gain support for optimizing some additional image types that were previously excluded from optimization (CMYK and grayscale). Fixes `#285 <https://github.com/jbarlow83/OCRmyPDF/issues/285>`_.
+
+v7.0.3
+------
+
+-   Fix issue `#284 <https://github.com/jbarlow83/OCRmyPDF/issues/284>`_, an error when parsing inline images that have are also image masks, by upgrading pikepdf to 0.3.1
+
+v7.0.2
+------
+
+-   Fix a regression with ``--rotate-pages`` on pages that already had rotations applied. (`#279 <https://github.com/jbarlow83/OCRmyPDF/issues/279>`_)
+
+-   Improve quality of page rotation in some cases by rasterizing a higher quality preview image. (`#281 <https://github.com/jbarlow83/OCRmyPDF/issues/281>`_)
+
+v7.0.1
+------
+
+-   Fix compatibility with img2pdf >= 0.3.0 by rejecting input images that have an alpha channel
+
+-   Add forward compatibility for pikepdf 0.3.0 (unrelated to img2pdf)
+
+-   Various documentation updates for v7.0.0 changes
+
+v7.0.0
+------
+
+-   The core algorithm for combining OCR layers with existing PDF pages has been rewritten and improved considerably.  PDFs are no longer split into single page PDFs for processing; instead, images are rendered and the OCR results are grafted onto the input PDF.  The new algorithm uses less temporary disk space and is much more performant especially for large files.
+
+-   New dependency: `pikepdf <https://github.com/pikepdf/pikepdf>`_. pikepdf is a powerful new Python PDF library driving the latest OCRmyPDF features, built on the QPDF C++ library (libqpdf).
+
+-   New feature: PDF optimization with ``-O`` or ``--optimize``.  After OCR, OCRmyPDF will perform image optimizations relevant to OCR PDFs.
+
+    +   If a JBIG2 encoder is available, then monochrome images will be converted, with the potential for huge savings on large black and white images, since JBIG2 is far more efficient than any other monochrome (bi-level) compression. (All known US patents related to JBIG2 have probably expired, but it remains the responsibility of the user to supply a JBIG2 encoder such as `jbig2enc <https://github.com/agl/jbig2enc>`_. OCRmyPDF does not implement JBIG2 encoding.)
+
+    +   If ``pngquant`` is installed, OCRmyPDF will optionally use it to perform lossy quantization and compression of PNG images.
+
+    +   The quality of JPEGs can also be lowered, on the assumption that a lower quality image may be suitable for storage after OCR.
+
+    +   This image optimization component will eventually be offered as an independent command line utility.
+
+    +   Optimization ranges from ``-O0`` through ``-O3``, where ``0`` disables optimization and ``3`` implements all options. ``1``, the default, performs only safe and lossless optimizations. (This is similar to GCC's optimization parameter.) The exact type of optimizations performed will vary over time.
+
+-   Small amounts of text in the margins of a page, such as watermarks, page numbers, or digital stamps, will no longer prevent the rest of a page from being OCRed when ``--skip-text`` is issued. This behavior is based on a heuristic.
+
+-   Removed features
+
+    +   The deprecated ``--pdf-renderer tesseract`` PDF renderer was removed.
+
+    +   ``-g``, the option to generate debug text pages, was removed because it was a maintenance burden and only worked in isolated cases. HOCR pages can still be previewed by running the hocrtransform.py with appropriate settings.
+
+-   Removed dependencies
+
+    +   ``PyPDF2``
+
+    +   ``defusedxml``
+
+    +   ``PyMuPDF``
+
+-   The ``sandwich`` PDF renderer can be used with all supported versions of Tesseract, including that those prior to v3.05 which don't support ``-c textonly``. (Tesseract v4.0.0 is recommended and more efficient.)
+
+-   ``--pdf-renderer auto`` option and the diagnostics used to select a PDF renderer now work better with old versions, but may make different decisions than past versions.
+
+-   If everything succeeds but PDF/A conversion fails, a distinct return code is now returned (``ExitCode.pdfa_conversion_failed (10)``) where this situation previously returned ``ExitCode.invalid_output_pdf (4)``. The latter is now returned only if there is some indication that the output file is invalid.
+
+-   Notes for downstream packagers
+
+    +   There is also a new dependency on ``python-xmp-toolkit`` which in turn depends on ``libexempi3``.
+
+    +   It may be necessary to separately ``pip install pycparser`` to avoid `another Python 3.7 issue <https://github.com/eliben/pycparser/pull/135>`_.
+
+v6.2.5
+------
+
+-   Disable a failing test due to Tesseract 4.0rc1 behavior change. Previously, Tesseract would exit with an error message if its configuration was invalid, and OCRmyPDF would intercept this message. Now Tesseract issues a warning, which OCRmyPDF v6.2.5 may relay or ignore. (In v7.x, OCRmyPDF will respond to the warning.)
+
+-   This release branch no longer supports using the optional PyMuPDF installation, since it was removed in v7.x.
+
+-   This release branch no longer supports macOS. macOS users should upgrade to v7.x.
+
+v6.2.4
+------
+
+-   Backport Ghostscript 9.25 compatibility fixes, which removes support for setting Unicode metadata
+-   Backport blacklisting Ghostscript 9.24
+-   Older versions of Ghostscript are still supported
+
+v6.2.3
+------
+
+-   Fix compatibility with img2pdf >= 0.3.0 by rejecting input images that have an alpha channel
+-   This version will be included in Ubuntu 18.10
+
+v6.2.2
+------
+
+-   Backport compatibility fixes for Python 3.7 and ruffus 2.7.0 from v7.0.0
+-   Backport fix to ignore masks when deciding what colors are on a page
+-   Backport some minor improvements from v7.0.0: better argument validation and warnings about the Tesseract 4.0.0 ``--user-words`` regression
+
+v6.2.1
+------
+
+-   Fix recent versions of Tesseract (after 4.0.0-beta1) not being detected as supporting the ``sandwich`` renderer (`#271 <https://github.com/ppjbarlow83/OCRmyPDF/issues/271>`_).
+
 v6.2.0
 ------
 
@@ -16,7 +234,7 @@ v6.2.0
 
 -   Creation of PDF/A-3 is now supported. However, there is no ability to attach files to PDF/A-3.
 
--   List more reasons why the file size might grow.
+-   Lists more reasons why the file size might grow.
 
 -   Fix issue `#262 <https://github.com/ppjbarlow83/OCRmyPDF/issues/262>`_, ``--remove-background`` error on PDFs contained colormapped (paletted) images.
 
@@ -34,7 +252,7 @@ v6.1.5
 v6.1.4
 ------
 
--   Fix issue `#248 <https://github.com/jbarlow83/OCRmyPDF/issues/248>`_ ``--clean`` argument may remove OCR from left column of text on certain documents. We now set ``--layout none`` to suppress this.  
+-   Fix issue `#248 <https://github.com/jbarlow83/OCRmyPDF/issues/248>`_ ``--clean`` argument may remove OCR from left column of text on certain documents. We now set ``--layout none`` to suppress this.
 
 -   The test cache was updated to reflect the change above.
 
@@ -77,7 +295,7 @@ v6.1.1
 v6.1.0
 ------
 
--   PyMuPDF is now an optional but recommended dependency, to alleviate installation difficulties on platforms that have less access to PyMuPDF than the author anticipated.  Install OCRmyPDF with ``pip install ocrmypdf[fitz]`` to use it to its full potential.
+-   PyMuPDF is now an optional but recommended dependency, to alleviate installation difficulties on platforms that have less access to PyMuPDF than the author anticipated.  (For version 6.x only) install OCRmyPDF with ``pip install ocrmypdf[fitz]`` to use it to its full potential.
 
 -   Fix ``FileExistsError`` that could occur if OCR timed out while it was generating the output file. (`#218 <https://github.com/jbarlow83/OCRmyPDF/issues/218>`_)
 
@@ -119,7 +337,7 @@ v6.0.0
 v5.7.0
 ------
 
--   Fixed an issue that caused poor CPU utilization on machines more than 4 cores when running Tesseract 4. (Related to issue `#217 <https://github.com/jbarlow83/OCRmyPDF/issues/217>`_.)
+-   Fixed an issue that caused poor CPU utilization on machines with more than 4 cores when running Tesseract 4. (Related to issue `#217 <https://github.com/jbarlow83/OCRmyPDF/issues/217>`_.)
 
 -   The 'hocr' renderer has been improved. The 'sandwich' and 'tesseract' renderers are still better for most use cases, but 'hocr' may be useful for people who work with the PDF.js renderer in English/ASCII languages. (`#225 <https://github.com/jbarlow83/OCRmyPDF/issues/225>`_)
 
@@ -165,7 +383,7 @@ v5.5
 -   Add new argument ``--max-image-mpixels``. Pillow 5.0 now raises an exception when images may be decompression bombs. This argument can be used to override the limit Pillow sets.
 -   Fix output page cropped when using the sandwich renderer and OCR is skipped on a rotated and image-processed page
 -   A warning is now issued when old versions of Ghostscript are used in cases known to cause issues with non-Latin characters
--   Fix a few parameter validation checks for ``-output-type pdfa-1`` and ``pdfa-2`` 
+-   Fix a few parameter validation checks for ``-output-type pdfa-1`` and ``pdfa-2``
 
 
 v5.4.4
@@ -175,7 +393,7 @@ v5.4.4
 -   Fix issue `#200 <https://github.com/jbarlow83/OCRmyPDF/issues/200>`_: an uncommon syntax for formatting decimal numbers in a PDF would cause qpdf to issue a warning, which ocrmypdf treated as an error. Now this the warning is relayed.
 -   Fix an issue where intermediate PDFs would be created at version 1.3 instead of the version of the original file. It's possible but unlikely this had side effects.
 -   A warning is now issued when older versions of qpdf are used since issues like `#200 <https://github.com/jbarlow83/OCRmyPDF/issues/200>`_ cause qpdf to infinite-loop
--   Address issue `#140 <https://github.com/jbarlow83/OCRmyPDF/issues/140>`_: if Tesseract outputs invalid UTF-8, escape it and print its message instead of aborting with a Unicode error 
+-   Address issue `#140 <https://github.com/jbarlow83/OCRmyPDF/issues/140>`_: if Tesseract outputs invalid UTF-8, escape it and print its message instead of aborting with a Unicode error
 -   Adding previously unlisted setup requirement, pytest-runner
 -   Update documentation: fix an error in the example script for Synology with Docker images, improved security guidance, advised ``pip install --user``
 
@@ -258,7 +476,6 @@ v5.0.1
 
 -   Fixed issue `#169 <https://github.com/jbarlow83/OCRmyPDF/issues/169>`_, exception due to failure to create sidecar text files on some versions of Tesseract 3.04, including the jbarlow83/ocrmypdf Docker image
 
-
 v5.0
 ----
 
@@ -274,7 +491,6 @@ v5.0
 -   Fixed issue `#164 <https://github.com/jbarlow83/OCRmyPDF/issues/164>`_, a typo
 -   Removed the command line parameters ``-n`` and ``--just-print`` since they have not worked for some time (reported as Ubuntu bug `#1687308 <https://bugs.launchpad.net/ubuntu/+source/ocrmypdf/+bug/1687308>`_)
 
-
 v4.5.6
 ------
 
@@ -282,21 +498,18 @@ v4.5.6
 -   Fixed issue `#158 <https://github.com/jbarlow83/OCRmyPDF/issues/158>`_, ocrmypdf now stops and terminates if Ghostscript fails on an intermediate step, as it is not possible to proceed.
 -   Fixed issue `#160 <https://github.com/jbarlow83/OCRmyPDF/issues/160>`_, exception thrown on certain invalid arguments instead of error message
 
-
 v4.5.5
 ------
 
 -   Automated update of macOS homebrew tap
 -   Fixed issue `#154 <https://github.com/jbarlow83/OCRmyPDF/issues/154>`_, KeyError '/Contents' when searching for text on blank pages that have no /Contents record.  Note: incomplete fix for this issue.
 
-
 v4.5.4
 ------
 
 -   Fix ``--skip-big`` raising an exception if a page contains no images (`#152 <https://github.com/jbarlow83/OCRmyPDF/issues/152>`_) (thanks to @TomRaz)
 -   Fix an issue where pages with no images might trigger "cannot write mode P as JPEG" (`#151 <https://github.com/jbarlow83/OCRmyPDF/issues/151>`_)
 
-
 v4.5.3
 ------
 
@@ -305,7 +518,6 @@ v4.5.3
 -   Fixed an issue where the ``tess4`` renderer would duplicate content onto output pages if tesseract failed or timed out
 -   Fixed ``tess4`` renderer not recognized when lossless reconstruction is possible
 
-
 v4.5.2
 ------
 
@@ -313,21 +525,18 @@ v4.5.2
 -   Make "using Tesseract 4.0" warning less ominous
 -   Set up machinery for homebrew OCRmyPDF tap
 
-
 v4.5.1
 ------
 
 -   Fix issue `#137 <https://github.com/jbarlow83/OCRmyPDF/issues/137>`_, proportions of images with a non-square pixel aspect ratio would be distorted in output for ``--force-ocr`` and some other combinations of flags
 
-
 v4.5
 ----
 
--   Exotic PDFs containing "Form XObjects" are now supported (issue `#134 <https://github.com/jbarlow83/OCRmyPDF/issues/134>`_; PDF reference manual 8.10), and images they contain are taken into account when determining the resolution for rasterizing
+-   PDFs containing "Form XObjects" are now supported (issue `#134 <https://github.com/jbarlow83/OCRmyPDF/issues/134>`_; PDF reference manual 8.10), and images they contain are taken into account when determining the resolution for rasterizing
 -   The Tesseract 4 Docker image no longer includes all languages, because it took so long to build something would tend to fail
 -   OCRmyPDF now warns about using ``--pdf-renderer tesseract`` with Tesseract 3.04 or lower due to issues with Ghostscript corrupting the OCR text in these cases
 
-
 v4.4.2
 ------
 
@@ -338,13 +547,12 @@ v4.4.2
 
 -   OCRmyPDF now prevents running the Tesseract 4 renderer with Tesseract 3.04, which was permitted in v4.4 and v4.4.1 but will not work
 
-
 v4.4.1
 ------
 
 -   To prevent a `TIFF output error <https://github.com/python-pillow/Pillow/issues/2206>`_ caused by img2pdf >= 0.2.1 and Pillow <= 3.4.2, dependencies have been tightened
 -   The Tesseract 4.00 simultaneous process limit was increased from 1 to 2, since it was observed that 1 lowers performance
--   Documentation improvements to describe the ``--tesseract-config`` feature 
+-   Documentation improvements to describe the ``--tesseract-config`` feature
 -   Added test cases and fixed error handling for ``--tesseract-config``
 -   Tweaks to setup.py to deal with issues in the v4.4 release
 
@@ -365,32 +573,27 @@ v4.4
 
     + However, OCRmyPDF's dependency "ruffus" is not re-entrant, so no Python API is available. Scripts should continue to use the command line interface.
 
-
 v4.3.5
 ------
 
 -   Update documentation to confirm Python 3.6.0 compatibility. No code changes were needed, so many earlier versions are likely supported.
 
-
 v4.3.4
 ------
 
 -   Fixed "decimal.InvalidOperation: quantize result has too many digits" for high DPI images
 
-
 v4.3.3
 ------
 
 -   Fixed PDF/A creation with Ghostscript 9.20 properly
 -   Fixed an exception on inline stencil masks with a missing optional parameter
 
-
 v4.3.2
 ------
 
 -   Fixed a PDF/A creation issue with Ghostscript 9.20 (note: this fix did not actually work)
 
-
 v4.3.1
 ------
 
@@ -401,13 +604,12 @@ v4.3.1
 -   Cleaned up debug output
 -   Tesseract caching in test cases is now more cautious about false cache hits and reproducing exact output, not that any problems were observed
 
-
 v4.3
 ----
 
 -   New feature ``--remove-background`` to detect and erase the background of color and grayscale images
 -   Better documentation
--   Fixed an issue with PDFs that draw images when the raster stack depth is zero 
+-   Fixed an issue with PDFs that draw images when the raster stack depth is zero
 -   ocrmypdf can now redirect its output to stdout for use in a shell pipeline
 
     +  This does not improve performance since temporary files are still used for buffering
@@ -419,7 +621,6 @@ v4.2.5
 -   Fixed an issue (`#100 <https://github.com/jbarlow83/OCRmyPDF/issues/100>`_) with PDFs that omit the optional /BitsPerComponent parameter on images
 -   Removed non-free file milk.pdf
 
-
 v4.2.4
 ------
 
@@ -439,20 +640,17 @@ v4.2.3
 -   Added a test case for indirect objects and linearized PDFs
 -   Deprecated the OCRmyPDF.sh shell script
 
-
 v4.2.2
 ------
 
 -   Improvements to documentation
 
-
 v4.2.1
 ------
 
 -   Fixed an issue where PDF pages that contained stencil masks would report an incorrect DPI and cause Ghostscript to abort
 -   Implemented stdin streaming
 
-
 v4.2
 ----
 
@@ -475,7 +673,7 @@ v4.2
     +  fixes issue `#82 <https://github.com/jbarlow83/OCRmyPDF/issues/82>`_
 
 -   Fixes an issue where, with certain settings, monochrome images in PDFs would be converted to 8-bit grayscale, increasing file size (`#79 <https://github.com/jbarlow83/OCRmyPDF/issues/79>`_)
--   Support for Ubuntu 12.04 LTS "precise" has been dropped in favor of (roughly) Ubuntu 14.04 LTS "trusty" 
+-   Support for Ubuntu 12.04 LTS "precise" has been dropped in favor of (roughly) Ubuntu 14.04 LTS "trusty"
 
     +  Some Ubuntu "PPAs" (backports) are needed to make it work
 
@@ -491,7 +689,6 @@ v4.1.4
 
 -   Bug fix: monochrome images with an ICC profile attached were incorrectly converted to full color images if lossless reconstruction was not possible due to other settings; consequence was increased file size for these images
 
-
 v4.1.3
 ------
 
@@ -500,14 +697,12 @@ v4.1.3
 -   Fix order of operations for matrix multiplication (no effect on most users)
 -   Add a few leptonica wrapper functions (no effect on most users)
 
-
 v4.1.2
 ------
 
 -   Replace IEC sRGB ICC profile with Debian's sRGB (from icc-profiles-free) which is more compatible with the MIT license
 -   More helpful error message for an error related to certain types of malformed PDFs
 
-
 v4.1
 ----
 
@@ -515,20 +710,17 @@ v4.1
 -   Fixed problems in error checking if ``unpaper`` is uninstalled or missing at run-time
 -   Fixed problems with "RethrownJobError" errors during error handling that suppressed the useful error messages
 
-
 v4.0.7
 ------
 
 -   Minor correction to Ghostscript output settings
 
-
 v4.0.6
 ------
 
 -   Update install instructions
 -   Provide a sRGB profile instead of using Ghostscript's
 
-
 v4.0.5
 ------
 
@@ -542,18 +734,14 @@ v4.0.4
 
 Released with verbose debug message turned on. Do not use. Skip to v4.0.5.
 
-
 v4.0.3
 ------
 
 New features
-^^^^^^^^^^^^
 
 -   Page orientations detected are now reported in a summary comment
 
-
 Fixes
-^^^^^
 
 -   Show stack trace if unexpected errors occur
 -   Treat "too few characters" error message from Tesseract as a reason to skip that page rather than
@@ -565,7 +753,7 @@ v4.0.2
 ------
 
 Fixes
-^^^^^
+
 
 -   Fixed compatibility with Tesseract 3.04.01 release, particularly its different way of outputting
     orientation information
@@ -577,7 +765,7 @@ v4.0.1
 ------
 
 Fixes
-^^^^^
+
 
 -   Fixed a KeyError if tesseract fails to find page orientation information
 
@@ -586,17 +774,15 @@ v4.0
 ----
 
 New features
-^^^^^^^^^^^^
 
 -   Automatic page rotation (``-r``) is now available. It uses ignores any prior rotation information
     on PDFs and sets rotation based on the dominant orientation of detectable text. This feature is
-    fairly reliable but some false positives occur especially if there is not much text to work with. (`#4 <https://github.com/jbarlow83/OCRmyPDF/issues/4>`_) 
+    fairly reliable but some false positives occur especially if there is not much text to work with. (`#4 <https://github.com/jbarlow83/OCRmyPDF/issues/4>`_)
 -   Deskewing is now performed using Leptonica instead of unpaper. Leptonica is faster and more reliable
     at image deskewing than unpaper.
 
 
 Fixes
-^^^^^
 
 -   Fixed an issue where lossless reconstruction could cause some pages to be appear incorrectly
     if the page was rotated by the user in Acrobat after being scanned (specifically if it a /Rotate tag)
@@ -605,7 +791,6 @@ Fixes
 
 
 Changes
-^^^^^^^
 
 -   Logging output is now much easier to read
 -   ``--deskew`` is now performed by Leptonica instead of unpaper (`#25 <https://github.com/jbarlow83/OCRmyPDF/issues/25>`_)
@@ -621,7 +806,6 @@ v3.2.1
 ------
 
 Changes
-^^^^^^^
 
 -   Fixed issue `#47 <https://github.com/jbarlow83/OCRmyPDF/issues/47>`_ "convert() got and unexpected keyword argument 'dpi'" by upgrading to img2pdf 0.2
 -   Tweaked the Dockerfiles
@@ -631,9 +815,8 @@ v3.2
 ----
 
 New features
-^^^^^^^^^^^^
 
--   Lossless reconstruction: when possible, OCRmyPDF will inject text layers without 
+-   Lossless reconstruction: when possible, OCRmyPDF will inject text layers without
     otherwise manipulating the content and layout of a PDF page. For example, a PDF containing a mix
     of vector and raster content would see the vector content preserved. Images may still be transcoded
     during PDF/A conversion.  (``--deskew`` and ``--clean-final`` disable this mode, necessarily.)
@@ -643,7 +826,6 @@ New features
     for the polyglots among us. It is much larger.
 
 Changes
-^^^^^^^
 
 -   JPEG transcoding quality is now 95 instead of the default 75. Bigger file sizes for less degradation.
 
@@ -653,7 +835,6 @@ v3.1.1
 ------
 
 Changes
-^^^^^^^
 
 -   Fixed bug that caused incorrect page size and DPI calculations on documents with mixed page sizes
 
@@ -661,17 +842,16 @@ v3.1
 ----
 
 Changes
-^^^^^^^
 
 -   Default output format is now PDF/A-2b instead of PDF/A-1b
 -   Python 3.5 and macOS El Capitan are now supported platforms - no changes were
     needed to implement support
 -   Improved some error messages related to missing input files
 -   Fixed issue `#20 <https://github.com/jbarlow83/OCRmyPDF/issues/20>`_ - uppercase .PDF extension not accepted
--   Fixed an issue where OCRmyPDF failed to text that certain pages contained previously OCR'ed text, 
+-   Fixed an issue where OCRmyPDF failed to text that certain pages contained previously OCR'ed text,
     such as OCR text produced by Tesseract 3.04
 -   Inserts /Creator tag into PDFs so that errors can be traced back to this project
--   Added new option ``--pdf-renderer=auto``, to let OCRmyPDF pick the best PDF renderer. 
+-   Added new option ``--pdf-renderer=auto``, to let OCRmyPDF pick the best PDF renderer.
     Currently it always chooses the 'hocrtransform' renderer but that behavior may change.
 -   Set up Travis CI automatic integration testing
 
@@ -679,21 +859,20 @@ v3.0
 ----
 
 New features
-^^^^^^^^^^^^
 
--   Easier installation with a Docker container or Python's ``pip`` package manager 
+-   Easier installation with a Docker container or Python's ``pip`` package manager
 -   Eliminated many external dependencies, so it's easier to setup
 -   Now installs ``ocrmypdf`` to ``/usr/local/bin`` or equivalent for system-wide
     access and easier typing
 -   Improved command line syntax and usage help (``--help``)
 -   Tesseract 3.03+ PDF page rendering can be used instead for better positioning
     of recognized text (``--pdf-renderer tesseract``)
--   PDF metadata (title, author, keywords) are now transferred to the 
+-   PDF metadata (title, author, keywords) are now transferred to the
     output PDF
 -   PDF metadata can also be set from the command line (``--title``, etc.)
 -   Automatic repairs malformed input PDFs if possible
 -   Added test cases to confirm everything is working
--   Added option to skip extremely large pages that take too long to OCR and are 
+-   Added option to skip extremely large pages that take too long to OCR and are
     often not OCRable (e.g. large scanned maps or diagrams); other pages are still
     processed (``--skip-big``)
 -   Added option to kill Tesseract OCR process if it seems to be taking too long on
@@ -702,7 +881,6 @@ New features
 -   Multiple images on the same PDF page are now supported
 
 Changes
-^^^^^^^
 
 -   New, robust rewrite in Python 3.4+ with ruffus_ pipelines
 -   Now uses Ghostscript 9.14's improved color conversion model to preserve PDF colors
@@ -712,9 +890,9 @@ Changes
     available CPUs, increasing performance
 -   The ``-o DPI`` argument has been phased out, in favor of ``--oversample DPI``, in
     case we need ``-o OUTPUTFILE`` in the future
--   Removed several dependencies, so it's easier to install.  We no 
+-   Removed several dependencies, so it's easier to install.  We no
     longer use:
-    
+
     - GNU parallel_
     - ImageMagick_
     - Python 2.7
@@ -730,7 +908,7 @@ Changes
     - qpdf_ 5.0.0+
     - Unpaper_ 6.1 (optional)
     - some automatically managed Python packages
-  
+
 .. _ruffus: http://www.ruffus.org.uk/index.html
 .. _parallel: https://www.gnu.org/software/parallel/
 .. _ImageMagick: http://www.imagemagick.org/script/index.php
@@ -739,8 +917,7 @@ Changes
 .. _Unpaper: https://github.com/Flameeyes/unpaper
 .. _JHOVE: http://jhove.sourceforge.net/
 
-Release candidates
-^^^^^^^^^^^^^^^^^^
+Release candidates^
 
 -   rc9:
 
@@ -776,7 +953,7 @@ Release candidates
     - fixed some installer issues and errors in installation instructions
     - improve performance: run Ghostscript with multithreaded rendering
     - improve performance: use multiple cores by default
-    - bug fix: checking for wrong exception on process timeout 
+    - bug fix: checking for wrong exception on process timeout
 
 -   rc3: skipping version number intentionally to avoid confusion with Tesseract
 -   rc2: first release for public testing to test-PyPI, Github
@@ -799,21 +976,20 @@ where ``settings.txt`` contains *one argument per line*, for example:
 
 ::
 
-    -l 
-    deu 
-    --author 
-    A. Merkel 
-    --pdf-renderer 
+    -l
+    deu
+    --author
+    A. Merkel
+    --pdf-renderer
     tesseract
 
 
 Fixes
-^^^^^
+
 
 -   Handling of filenames containing spaces: fixed
 
 Notes and known issues
-^^^^^^^^^^^^^^^^^^^^^^
 
 -   Some dependencies may work with lower versions than tested, so try
     overriding dependencies if they are "in the way" to see if they work.

docs/security.rst

@@ -1,7 +1,7 @@
 PDF security issues
 ===================
 
-	OCRmyPDF should only be used on PDFs you trust. It is not designed to protect you against malware. 
+	OCRmyPDF should only be used on PDFs you trust. It is not designed to protect you against malware.
 
 Recognizing that many users have an interest in handling PDFs and applying OCR to PDFs they did not generate themselves, this article discusses the security implications of PDFs and how users can protect themselves.
 
@@ -19,9 +19,7 @@ This `article <https://theinvisiblethings.blogspot.ca/2013/02/converting-untrust
 How OCRmyPDF processes PDFs
 ---------------------------
 
-OCRmyPDF must open and interpret your PDF in order to insert an OCR layer. First, it runs all PDFs through `qpdf <https://github.com/qpdf/qpdf>`_, a program that repairs PDFs with syntax errors. This is done because, in the author's experience, a significant number of PDFs in the wild especially those created by scanners are not well-formed files. qpdf makes it more likely that OCRmyPDF will succeed, but offers no security guarantees. qpdf is also used to split the PDF into single page PDFs.
-
-After qpdf, OCRmyPDF examines each page using `PyPDF2 <https://github.com/mstamy2/PyPDF2>`_. This library also has no warranties or guarantees. OCRmyPDF works with qpdf 5.0 and up, but version 7.0 is recommended because of known security vulnerabilities in early versions.
+OCRmyPDF must open and interpret your PDF in order to insert an OCR layer. First, it runs all PDFs through `pikepdf <https://github.com/pikepdf/pikepdf>`_, a library based on `qpdf <https://github.com/qpdf/qpdf>`_, a program that repairs PDFs with syntax errors. This is done because, in the author's experience, a significant number of PDFs in the wild especially those created by scanners are not well-formed files. qpdf makes it more likely that OCRmyPDF will succeed, but offers no security guarantees. qpdf is also used to split the PDF into single page PDFs.
 
 Finally, OCRmyPDF rasterizes each page of the PDF using `Ghostscript <http://ghostscript.com/>`_ in ``-dSAFER`` mode.
 
@@ -60,13 +58,13 @@ Commercial alternatives
 
 The author also provides professional services that include OCR and building databases around PDFs, and is happy to provide consultation.
 
-Abbyy Cloud OCR is a viable commercial alternative with a web services API. 
+Abbyy Cloud OCR is a viable commercial alternative with a web services API.
 
 
 Password protection, digital signatures and certification
 ---------------------------------------------------------
 
-Password protected PDFs usually have two passwords, and owner and user password. When the user password is set to empty, PDF readers will open the file automatically and marked it as "(SECURED)". While not as reliable as a digital signature, this indicates that whoever set the password approved of the file at that time. When the user password is set, the document cannot be viewed without the password. 
+Password protected PDFs usually have two passwords, and owner and user password. When the user password is set to empty, PDF readers will open the file automatically and marked it as "(SECURED)". While not as reliable as a digital signature, this indicates that whoever set the password approved of the file at that time. When the user password is set, the document cannot be viewed without the password.
 
 Either way, OCRmyPDF does not remove passwords from PDFs and exits with an error on encountering them.
 
@@ -76,4 +74,4 @@ After OCR is applied, password protection is not permitted on PDF/A documents bu
 
 Many programs exist which are capable of inserting an image of someone's signature. On its own, this offers no security guarantees. It is trivial to remove the signature image and apply it to other files. This practice offers no real security.
 
-Important documents can be digitally signed and certified to attest to their authorship. OCRmyPDF cannot do this. Open source tools such as pdfbox (Java) have this capability as does Adobe Acrobat. 
+Important documents can be digitally signed and certified to attest to their authorship. OCRmyPDF cannot do this. Open source tools such as pdfbox (Java) have this capability as does Adobe Acrobat.

requirements.txt

@@ -1,10 +0,0 @@
-# requirements.txt can be used to replicate the developer's build environment
-# setup.py lists a separate set of requirements that are looser to simplify
-# installation
-ruffus == 2.6.3
-Pillow == 5.1.0
-reportlab == 3.4.0
-PyPDF2 == 1.26.0
-img2pdf == 0.2.4
-cffi == 1.11.5
-PyMuPDF == 1.12.5

requirements/dev.txt

@@ -1,4 +1,4 @@
 check-manifest >= 0.35
 twine >= 1.8.1
-coverage >= 4.4
+coverage >= 4.5
 GitPython == 2.1.3

requirements/main.txt

@@ -0,0 +1,13 @@
+# requirements.txt can be used to replicate the developer's build environment
+# setup.py lists a separate set of requirements that are looser to simplify
+# installation
+chardet == 3.0.4
+cffi == 1.11.5
+img2pdf == 0.3.1
+pdfminer.six == 20181108
+pikepdf == 0.3.7
+Pillow >= 5.0.0, != 5.1.0 ; sys_platform == "darwin"
+pycparser == 2.19
+python-xmp-toolkit == 2.0.1
+reportlab == 3.5.9
+ruffus == 2.8.0

requirements/test.txt

@@ -0,0 +1,8 @@
+pytest == 3.9.3
+pytest-helpers-namespace
+pytest-xdist
+pytest-cov
+python-xmp-toolkit   # requires apt-get install libexempi3
+                     # or brew install exempi
+PyPDF2 >= 1.26.0
+#PyMuPDF == 1.13.4    # optional

setup.py

@@ -186,26 +186,32 @@ if not forced and command.startswith('install') or \
     )
     check_external_program(
         program='qpdf',
-        need_version='7.0.0', # test suite known to fail on 5.1.1
-        package='qpdf',
-        version_check_args=['--version']
+        need_version='8.0.2', # test suite known to fail on 5.1.1
+        package='qpdf'
+    )
+    check_external_program(
+        program='pngquant',
+        need_version='2.0.0',
+        package='pngquant',
+        optional=True
     )
 
 if 'upload' in sys.argv[1:]:
     print('Use twine to upload the package - setup.py upload is insecure')
     sys.exit(1)
 
-tests_require = open('test_requirements.txt').read().splitlines()
+tests_require = open('requirements/test.txt', encoding='utf-8').read().splitlines()
 
 
 def readme():
-    with open('README.rst') as f:
+    with open('README.md', encoding='utf-8') as f:
         return f.read()
 
 setup(
     name='ocrmypdf',
     description='OCRmyPDF adds an OCR text layer to scanned PDF files, allowing them to be searched',
     long_description=readme(),
+    long_description_content_type='text/markdown',
     url='https://github.com/jbarlow83/OCRmyPDF',
     author='James R. Barlow',
     author_email='jim@purplerock.ca',
@@ -215,6 +221,7 @@ setup(
     classifiers=[
         "Programming Language :: Python :: 3.5",
         "Programming Language :: Python :: 3.6",
+        "Programming Language :: Python :: 3.7",
         "Development Status :: 5 - Production/Stable",
         "Environment :: Console",
         "Intended Audience :: End Users/Desktop",
@@ -234,6 +241,7 @@ setup(
         'cffi >= 1.9.1',        # to build the leptonica module
         'pytest-runner',        # to enable python setup.py test
         'setuptools_scm',       # so that version will work
+        'setuptools_scm_git_archive'    # enable version from github tarballs
     ],
     use_scm_version={'version_scheme': 'post-release'},
     cffi_modules=[
@@ -241,17 +249,18 @@ setup(
     ],
     install_requires=[
         'cffi >= 1.9.1',          # must be a setup and install requirement
-        'defusedxml >= 0.5.0',    # pure Python, so track HEAD closely
-        'img2pdf >= 0.2.4',       # pure Python, so track HEAD closely
-        'Pillow >= 4.0.0, != 5.1.0 ; sys_platform == "darwin"',        
+        'img2pdf >= 0.3.0, < 0.4',       # pure Python, so track HEAD closely
+        'pdfminer.six == 20181108',
+        'pikepdf >= 0.3.7, < 0.4',
+        'Pillow >= 4.0.0, != 5.1.0 ; sys_platform == "darwin"',
                                   # Pillow < 4 has BytesIO/TIFF bug w/img2pdf 0.2.3
                                   # block 5.1.0, broken wheels
-        'PyPDF2 >= 1.26',         # pure Python, so track HEAD closely
+        'python-xmp-toolkit >= 2, < 3',
         'reportlab >= 3.3.0',     # oldest released version with sane image handling
-        'ruffus == 2.6.3',        # pinned - ocrmypdf implements a 2.6.3 workaround
+        'ruffus >= 2.7.0',
     ],
     extras_require={
-        'fitz': ['PyMuPDF >= 1.12.5']     # for table of contents bug
+        'fitz': [],  # Backward compatibility
     },
     tests_require=tests_require,
     entry_points={
@@ -261,4 +270,10 @@ setup(
     },
     package_data={'ocrmypdf': ['data/sRGB.icc']},
     include_package_data=True,
-    zip_safe=False)
+    zip_safe=False,
+    project_urls={
+        'Documentation': 'https://ocrmypdf.readthedocs.io/',
+        'Source': 'https://github.com/jbarlow83/ocrmypdf',
+        'Tracker': 'https://github.com/jbarlow83/ocrmypdf/issues'
+    }
+)

src/ocrmypdf/__init__.py

@@ -23,3 +23,16 @@ PROGRAM_NAME = 'ocrmypdf'
 __version__ = pkg_resources.get_distribution('ocrmypdf').version
 
 VERSION = __version__
+
+from .exceptions import (
+    ExitCode, BadArgsError, PdfMergeFailedError, MissingDependencyError,
+    UnsupportedImageFormatError, DpiError, OutputFileAccessError,
+    PriorOcrFoundError, InputFileError, SubprocessOutputError,
+    EncryptedPdfError, TesseractConfigError
+)
+
+from . import helpers
+from . import hocrtransform
+from . import leptonica
+from . import pdfa
+from . import pdfinfo

src/ocrmypdf/__main__.py

@@ -17,32 +17,27 @@
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
 from tempfile import mkdtemp
-from collections.abc import Sequence
 from pathlib import Path
 import sys
 import os
 import re
-import warnings
-import multiprocessing
 import atexit
 import textwrap
 import logging
 import argparse
 
-import PyPDF2 as pypdf
 import PIL
 
 import ruffus.ruffus_exceptions as ruffus_exceptions
 import ruffus.cmdline as cmdline
 import ruffus.proxy_logger as proxy_logger
 
-from .pipeline import JobContext, JobContextManager, \
-    cleanup_working_files, build_pipeline
+from ._jobcontext import JobContext, JobContextManager, cleanup_working_files
+from ._pipeline import build_pipeline
 from .pdfa import file_claims_pdfa
-from .helpers import is_iterable_notstr, re_symlink, is_file_writable, \
+from .helpers import re_symlink, is_file_writable, \
     available_cpu_count
 from .exec import tesseract, qpdf, ghostscript
-from .lib import fitz
 from . import PROGRAM_NAME, VERSION
 
 from .exceptions import ExitCode, ExitCodeException, MissingDependencyError, \
@@ -50,8 +45,6 @@ from .exceptions import ExitCode, ExitCodeException, MissingDependencyError, \
 from . import exceptions as ocrmypdf_exceptions
 from ._unicodefun import verify_python3_env
 
-warnings.simplefilter('ignore', pypdf.utils.PdfReadWarning)
-
 
 # -------------
 # External dependencies
@@ -70,7 +63,7 @@ def complain(message):
 if 'IDE_PROJECT_ROOTS' in os.environ:
     os.environ['PATH'] = '/usr/local/bin:' + os.environ['PATH']
 
-# -------- 
+# --------
 # Critical environment tests
 
 verify_python3_env()
@@ -85,6 +78,21 @@ if tesseract.version() < MINIMUM_TESS_VERSION:
 # -------------
 # Parser
 
+def numeric(basetype, min_=None, max_=None):
+    """Validator for numeric params"""
+    min_ = basetype(min_) if min_ is not None else None
+    max_ = basetype(max_) if max_ is not None else None
+    def _numeric(string):
+        value = basetype(string)
+        if (min_ is not None and value < min_
+                or max_ is not None and value > max_):
+            msg = "%r not in valid range %r" % (string, (min_, max_))
+            raise argparse.ArgumentTypeError(msg)
+        return value
+    _numeric.__name__ = basetype.__name__
+    return _numeric
+
+
 parser = argparse.ArgumentParser(
     prog=PROGRAM_NAME,
     fromfile_prefix_chars='@',
@@ -156,7 +164,7 @@ parser.add_argument(
     '--image-dpi', metavar='DPI', type=int,
     help="For input image instead of PDF, use this DPI instead of file's.")
 parser.add_argument(
-    '--output-type', choices=['pdfa', 'pdf', 'pdfa-1', 'pdfa-2', 'pdfa-3'], 
+    '--output-type', choices=['pdfa', 'pdf', 'pdfa-1', 'pdfa-2', 'pdfa-3'],
     default='pdfa',
     help="Choose output type. 'pdfa' creates a PDF/A-2b compliant file for "
          "long term archiving (default, recommended) but may not suitable "
@@ -186,7 +194,7 @@ parser.add_argument(
 jobcontrol = parser.add_argument_group(
     "Job control options")
 jobcontrol.add_argument(
-    '-j', '--jobs', metavar='N', type=int,
+    '-j', '--jobs', metavar='N', type=numeric(int, 0, 256),
     help="Use up to N CPU cores simultaneously (default: use all).")
 jobcontrol.add_argument(
     '-q', '--quiet', action='store_true', help="Suppress INFO messages")
@@ -233,37 +241,97 @@ preprocessing.add_argument(
     help="Clean page as above, and incorporate the cleaned image in the final "
          "PDF.  Might remove desired content.")
 preprocessing.add_argument(
-    '--oversample', metavar='DPI', type=int, default=0,
+    '--oversample', metavar='DPI', type=numeric(int, 0, 5000), default=0,
     help="Oversample images to at least the specified DPI, to improve OCR "
          "results slightly")
+preprocessing.add_argument(
+    '--remove-vectors', action='store_true',
+    help="EXPERIMENTAL. Mask out any vector objects in the PDF so that they "
+         "will not be included in OCR. This can eliminate false characters.")
+preprocessing.add_argument(
+    '--mask-barcodes', action='store_true',
+    help="EXPERIMENTAL. Mask out any barcodes that appear in the PDF so they are not "
+         "considered during OCR. Barcodes can introduce false characters into "
+         "OCR.")
+preprocessing.add_argument(
+    '--threshold', action='store_true',
+    help="EXPERIMENTAL. Threshold image to 1bpp before sending it to Tesseract for OCR. Can "
+         "improve OCR quality compared to Tesseract's thresholder.")
 
 ocrsettings = parser.add_argument_group(
     "OCR options",
     "Control how OCR is applied")
 ocrsettings.add_argument(
     '-f', '--force-ocr', action='store_true',
-    help="Rasterize any fonts or vector objects on each page, apply OCR, and "
+    help="Rasterize any text or vector objects on each page, apply OCR, and "
          "save the rastered output (this rewrites the PDF)")
 ocrsettings.add_argument(
     '-s', '--skip-text', action='store_true',
     help="Skip OCR on any pages that already contain text, but include the "
          "page in final output; useful for PDFs that contain a mix of "
          "images, text pages, and/or previously OCRed pages")
-# ocrsettings.add_argument(
-#     '--redo-ocr', action='store_true',
-#     help="removing any existing OCR text, but otherwise preserve mixed PDF "
-#          "pages")
-
 ocrsettings.add_argument(
-    '--skip-big', type=float, metavar='MPixels',
+    '--redo-ocr', action='store_true',
+    help="Attempt to detect and remove the hidden OCR layer from files that "
+         "were previously OCRed with OCRmyPDF or another program. Apply OCR "
+         "to text found in raster images. Existing visible text objects will "
+         "not be changed. If there is no existing OCR, OCR will be added.")
+ocrsettings.add_argument(
+    '--skip-big', type=numeric(float, 0, 5000), metavar='MPixels',
     help="Skip OCR on pages larger than the specified amount of megapixels, "
          "but include skipped pages in final output")
 
+optimizing = parser.add_argument_group(
+    "Optimization options",
+    "Control how the PDF is optimized after OCR"
+)
+optimizing.add_argument(
+    '-O', '--optimize', type=int, choices=range(0, 4), default=1,
+    help=("Control how PDF is optimized after processing:"
+        "0 - do not optimize; "
+        "1 - do safe, lossless optimizations (default); "
+        "2 - do some lossy optimizations; "
+        "3 - do aggressive lossy optimizations (including lossy JBIG2)"
+    )
+)
+optimizing.add_argument(
+    '--jpeg-quality', type=numeric(int, 0, 100), default=0, metavar='Q',
+    help=("Adjust JPEG quality level for JPEG optimization. "
+          "100 is best quality and largest output size; "
+          "1 is lowest quality and smallest output; "
+          "0 uses the default."
+    )
+)
+optimizing.add_argument(
+    '--jpg-quality', type=numeric(int, 0, 100), default=0, metavar='Q',
+    dest='jpeg_quality',
+    help=argparse.SUPPRESS  # Alias for --jpeg-quality
+)
+optimizing.add_argument(
+    '--png-quality', type=numeric(int, 0, 100), default=0, metavar='Q',
+    help=("Adjust PNG quality level to use when quantizing PNGs. "
+          "Values have same meaning as with --jpeg-quality"
+    )
+)
+optimizing.add_argument(
+    '--jbig2-lossy', action='store_true',
+    help=("Enable JBIG2 lossy mode (better compression, not suitable for some "
+          "use cases - see documentation)."
+    )
+)
+optimizing.add_argument(
+    '--jbig2-page-group-size', type=numeric(int, 1, 10000), default=0,
+    metavar='N',
+    # Adjust number of pages to consider at once for JBIG2 compression
+    help=argparse.SUPPRESS
+)
+
 advanced = parser.add_argument_group(
     "Advanced",
     "Advanced options to control Tesseract's OCR behavior")
 advanced.add_argument(
-    '--max-image-mpixels', action='store', type=float, metavar='MPixels',
+    '--max-image-mpixels', action='store', type=numeric(float, 0),
+    metavar='MPixels',
     help="Set maximum number of pixels to unpack before treating an image as a "
          "decompression bomb",
     default=128.0)
@@ -285,22 +353,18 @@ advanced.add_argument(
     )
 advanced.add_argument(
     '--pdf-renderer',
-    choices=['auto', 'tesseract', 'hocr', 'sandwich'], default='auto',
+    choices=['auto', 'hocr', 'sandwich'], default='auto',
     help="Choose OCR PDF renderer - the default option is to let OCRmyPDF "
-         "choose."
-         "auto - let OCRmyPDF choose; "
-         "sandwich - default renderer for Tesseract 3.05.01 and newer; "
-         "hocr - default renderer for older versions of Tesseract; "
-         "tesseract - gives better results for non-Latin languages and "
-         "Tesseract older than 3.05.01 but has problems with some versions "
-         " of Ghostscript; deprecated"
+         "choose.  See documentation for discussion."
     )
 advanced.add_argument(
-    '--tesseract-timeout', default=180.0, type=float, metavar='SECONDS',
+    '--tesseract-timeout', default=180.0, type=numeric(float, 0),
+    metavar='SECONDS',
     help='Give up on OCR after the timeout, but copy the preprocessed page '
          'into the final output')
 advanced.add_argument(
-    '--rotate-pages-threshold', default=14.0, type=float, metavar='CONFIDENCE',
+    '--rotate-pages-threshold', default=14.0, type=numeric(float, 0, 1000),
+    metavar='CONFIDENCE',
     help="Only rotate pages when confidence is above this value (arbitrary "
          "units reported by tesseract)")
 advanced.add_argument(
@@ -323,14 +387,6 @@ advanced.add_argument(
 advanced.add_argument(
     '--user-patterns', metavar='FILE',
     help="Specify the location of the Tesseract user patterns file.")
-advanced.add_argument(
-    '--skip-repair', action='store_true',
-    help="Normally OCRmyPDF automatically repairs PDFs using qpdf before "
-         "processing.  If you have already run qpdf or a similar program "
-         "that repairs PDF errors, you can tell OCRmyPDF to skip repair with "
-         "this option.  This may be helpful in batch processing where all "
-         "files are repaired prior to OCR occurs, since repair is single "
-         "threaded and time consuming for large files.")
 
 debugging = parser.add_argument_group(
     "Debugging",
@@ -338,9 +394,6 @@ debugging = parser.add_argument_group(
 debugging.add_argument(
     '-k', '--keep-temporary-files', action='store_true',
     help="Keep temporary files (helpful for debugging)")
-debugging.add_argument(
-    '-g', '--debug-rendering', action='store_true',
-    help="Render each page twice with debug information on second page")
 debugging.add_argument(
     '--flowchart', type=str,
     help="Generate the pipeline execution flowchart")
@@ -365,35 +418,58 @@ def check_options_languages(options, _log):
 
 
 def check_options_output(options, log):
+    # We have these constraints to check for.
+    # 1. Ghostscript < 9.20 mangles multibyte Unicode
+    # 2. Tesseract < 3.05 embeds an older version of GlyphlessFont with which
+    #    no version of Ghostscript handles correctly.
+    # 3. hocr doesn't work on non-Latin languages (so don't select it)
+
+    languages = set(options.language)
+    is_latin = languages.issubset(HOCR_OK_LANGS)
+
+    if options.pdf_renderer == 'hocr' and not is_latin:
+        msg = (
+            "The 'hocr' PDF renderer is known to cause problems with one "
+            "or more of the languages in your document.  Use "
+            "--pdf-renderer auto (the default) to avoid this issue.")
+        log.warning(msg)
+
+    if ghostscript.version() < '9.20' \
+            and options.output_type != 'pdf' \
+            and not is_latin:
+        # https://bugs.ghostscript.com/show_bug.cgi?id=696874
+        # Ghostscript < 9.20 fails to encode multibyte characters properly
+        msg = (
+            "The installed version of Ghostscript does not work correctly "
+            "with the OCR languages you specified. Use --output-type pdf or "
+            "upgrade to Ghostscript 9.20 or later to avoid this issue.")
+        msg += "Found Ghostscript {}".format(ghostscript.version())
+        log.warning(msg)
+
+    # Decide on what renderer to use
     if options.pdf_renderer == 'auto':
-        if tesseract.has_textonly_pdf():
-            options.pdf_renderer = 'sandwich'
-        else:
+        if tesseract.version() < '3.05' \
+                and options.output_type.startswith('pdfa') \
+                and is_latin:
             options.pdf_renderer = 'hocr'
+        else:
+            options.pdf_renderer = 'sandwich'
 
-    if options.pdf_renderer == 'sandwich' and not tesseract.has_textonly_pdf():
-        raise MissingDependencyError(
-            "The 'sandwich' renderer requires Tesseract 3.05.01 or newer; "
-            "or Tesseract 4.00 alpha newer than February 2017.")
+    if options.pdf_renderer == 'sandwich' \
+            and tesseract.version() < '3.05':
+        msg = (
+            "Ghostscript will corrupt the OCR text of PDFs produced by "
+            "Tesseract 3.04.xx and older.  For best results, upgrade to a "
+            "newer release of Tesseract. "
+        )
 
-    if options.pdf_renderer == 'tesseract':
-        if tesseract.version() < '3.05' and \
-                options.output_type.startswith('pdfa'):
-            log.warning(
-                "For best results use --pdf-renderer=tesseract "
-                "--output-type=pdf to disable PDF/A generation via "
-                "Ghostscript, which is known to corrupt the OCR text of "
-                "some PDFs produced your version of Tesseract.")
-        elif tesseract.has_textonly_pdf():
-            log.warning(
-                "The argument --pdf-renderer=tesseract provides support for "
-                "versions of tesseract older than your version. For best "
-                "results omit this argument and let OCRmyPDF choose the "
-                "best available renderer.")
-
-    if options.debug_rendering and options.pdf_renderer != 'hocr':
-        log.info(
-            "Ignoring --debug-rendering because it requires --pdf-renderer=hocr")
+        if options.output_type.startswith('pdfa'):
+            msg += (
+                "The argument --output-type=pdfa* requires Ghostscript, so "
+                "the PDF will be invalid.  If you cannot upgrade Tesseract, "
+                "use --output-type=pdf.")
+            raise MissingDependencyError(msg)
+        log.warning(msg)
 
     if options.output_type == 'pdfa':
         options.output_type = 'pdfa-2'
@@ -404,12 +480,18 @@ def check_options_output(options, log):
         )
 
     lossless_reconstruction = False
-    if options.pdf_renderer in ('hocr', 'sandwich'):
-        if not any((options.deskew, options.clean_final, options.force_ocr,
-                   options.remove_background)):
-            lossless_reconstruction = True
+    if not any((options.deskew, options.clean_final, options.force_ocr,
+                options.remove_background)):
+        lossless_reconstruction = True
     options.lossless_reconstruction = lossless_reconstruction
 
+    if not options.lossless_reconstruction and options.redo_ocr:
+        raise argparse.ArgumentError(
+            None,
+            "--redo-ocr is not currently compatible with --deskew, "
+            "--clean-final, and --remove-background"
+        )
+
 
 def check_options_sidecar(options, log):
     if options.sidecar == '\0':
@@ -421,69 +503,86 @@ def check_options_sidecar(options, log):
         options.sidecar = options.output_file + '.txt'
 
 
+def _optional_program_required(name, version_fn, min_version, for_argument):
+    try:
+        if version_fn() < min_version:
+            raise MissingDependencyError(
+                "The installed '{}' is not supported. "
+                "Install version {} or newer.".format(name, min_version))
+    except (FileNotFoundError, MissingDependencyError):
+        raise MissingDependencyError(
+            "Install the '{}' program to use {}.".format(name, for_argument))
+
+
+def _optional_program_recommended(name, version_fn, min_version, for_argument):
+    try:
+        if version_fn() < min_version:
+            raise MissingDependencyError(
+                "The installed '{}' is not supported. "
+                "Install version {} or newer.".format(name, min_version))
+    except (FileNotFoundError, MissingDependencyError):
+        complain(
+            "For best results, install the optional program '{}' to use the "
+            "argument {}.".format(name, for_argument)
+        )
+
+
 def check_options_preprocessing(options, log):
     if any((options.clean, options.clean_final)):
         from .exec import unpaper
-        try:
-            if unpaper.version() < '6.1':
-                raise MissingDependencyError(
-                    "The installed 'unpaper' is not supported. "
-                    "Install version 6.1 or newer.")
-        except FileNotFoundError:
-            raise MissingDependencyError(
-                "Install the 'unpaper' program to use --clean, --clean-final.")
-
-    if options.clean and \
-            not options.clean_final and \
-            options.pdf_renderer == 'tesseract':
-        log.info(
-            "Tesseract PDF renderer cannot render --clean pages without "
-            "also performing --clean-final, so --clean-final is assumed.")
+        _optional_program_required(
+            'unpaper', unpaper.version, '6.1', '--clean, --clean-final'
+        )
 
 
 def check_options_ocr_behavior(options, log):
-    if options.force_ocr and options.skip_text:
+    exclusive_options = sum(
+        [(1 if opt else 0)
+         for opt in (options.force_ocr, options.skip_text, options.redo_ocr)
+        ]
+    )
+    if exclusive_options >= 2:
         raise argparse.ArgumentError(
             None,
-            "Error: --force-ocr and --skip-text are mutually incompatible.")
+            "Error: choose only one of --force-ocr, --skip-text, --redo-ocr.")
+    if options.force_ocr and any((options.mask_barcodes, options.threshold)):
+        raise argparse.ArgumentError(
+            '--force-ocr',
+            'Error: --force-ocr currently may not be used with --threshold or --mask-barcodes'
+        )
 
-    # if options.redo_ocr and (options.skip_text or options.force_ocr):
-    #     raise argparse.ArgumentError(
-    #         "Error: --redo-ocr and other OCR options are incompatible.")
-    languages = set(options.language)
-    if options.pdf_renderer == 'hocr' and \
-            not languages.issubset(HOCR_OK_LANGS):
-        msg = (
-            "The 'hocr' PDF renderer is known to cause problems with one "
-            "or more of the languages in your document. ")
 
-        if tesseract.has_textonly_pdf():
-            msg += (
-                "Use --pdf-renderer auto (the default) to avoid this issue.")
-        else:
-            msg += (
-                "Use --pdf-renderer tesseract --output-type pdf to avoid "
-                "this issue")
-        log.warning(msg)
-    elif ghostscript.version() < '9.20' and \
-            not languages.issubset(HOCR_OK_LANGS) \
-            and options.output_type != 'pdf':
-        msg = (
-            "The installed version of Ghostscript does not work correctly "
-            "with the OCR languages you specified. Use --output-type pdf or "
-            "upgrade to Ghostscript 9.20 or later to avoid this issue.")
-        msg += "Found Ghostscript {}".format(ghostscript.version())
-        log.warning(msg)
+def check_options_optimizing(options, log):
+    if options.optimize >= 2:
+        from .exec import pngquant, jbig2enc
+        _optional_program_required(
+            'pngquant', pngquant.version, '2.0.1', '--optimize {2,3}'
+        )
+
+    if options.jbig2_lossy:
+        _optional_program_required(
+            'jbig2', jbig2enc.version, '0.28', '--jbig2-lossy'
+        )
+    elif options.optimize >= 2:
+        # Although we use JBIG2 for optimize=1, don't nag about it unless the
+        # user is asking for more optimization
+        _optional_program_recommended(
+            'jbig2', jbig2enc.version, '0.28', '--optimize {2,3}'
+        )
+
+    if options.optimize == 0 and any([
+            options.jbig2_lossy, options.png_quality, options.jpeg_quality
+            ]):
+        log.warning(
+            "The arguments --jbig2-lossy, --png-quality, and --jpeg-quality "
+            "will be ignored because --optimize=0."
+        )
 
 
 def check_options_advanced(options, log):
     if options.tesseract_oem and not tesseract.v4():
         log.warning(
             "--tesseract-oem requires Tesseract 4.x -- argument ignored")
-    if options.pdf_renderer == 'sandwich' and not tesseract.has_textonly_pdf():
-        raise MissingDependencyError(
-            "--pdf-renderer sandwich requires Tesseract 4.x "
-            "commit 3d9fb3b or later")
     if options.pdfa_image_compression != 'auto' and \
             options.output_type.startswith('pdfa'):
         log.warning(
@@ -491,6 +590,10 @@ def check_options_advanced(options, log):
             "--output-type is not 'pdfa', 'pdfa-1', or 'pdfa-2'"
         )
 
+    if tesseract.v4() and (options.user_words or options.user_patterns):
+        log.warning(
+            'Tesseract 4.x ignores --user-words, so this has no effect')
+
 
 def check_options_metadata(options, log):
     import unicodedata
@@ -520,6 +623,7 @@ def check_options(options, log):
         check_options_sidecar(options, log)
         check_options_preprocessing(options, log)
         check_options_ocr_behavior(options, log)
+        check_options_optimizing(options, log)
         check_options_advanced(options, log)
         check_options_pillow(options, log)
     except ValueError as e:
@@ -569,60 +673,47 @@ def do_ruffus_exception(ruffus_five_tuple, options, log):
     description of the error message that occurred."""
     exit_code = None
 
-    task_name, job_name, exc_name, exc_value, exc_stack = ruffus_five_tuple
-    job_name = job_name  # unused
-    if exc_name == 'builtins.SystemExit':
+    _task_name, _job_name, exc_name, exc_value, exc_stack = ruffus_five_tuple
+
+    if isinstance(exc_name, type):
+        # ruffus is full of mystery... sometimes (probably when the process
+        # group leader is killed) exc_name is the class object of the exception,
+        # rather than a str. So reach into the object and get its name.
+        exc_name = exc_name.__name__
+
+    if exc_name.startswith('ocrmypdf.exceptions.'):
+        base_exc_name = exc_name.replace('ocrmypdf.exceptions.', '')
+        exc_class = getattr(ocrmypdf_exceptions, base_exc_name)
+        exit_code = getattr(exc_class, 'exit_code', ExitCode.other_error)
+        try:
+            if isinstance(exc_value, exc_class):
+                exc_msg = str(exc_value)
+            elif isinstance(exc_value, str):
+                exc_msg = exc_value
+            else:
+                exc_msg = str(exc_class())
+        except Exception:
+            exc_msg = "Unknown"
+
+    if exc_name in ('builtins.SystemExit', 'SystemExit'):
         match = re.search(r"\.(.+?)\)", exc_value)
         exit_code_name = match.groups()[0]
-        exit_code = getattr(ExitCode, exit_code_name, 'other_error')        
+        exit_code = getattr(ExitCode, exit_code_name, 'other_error')
     elif exc_name == 'ruffus.ruffus_exceptions.MissingInputFileError':
         log.error(cleanup_ruffus_error_message(exc_value))
         exit_code = ExitCode.input_file
-    elif exc_name == 'builtins.TypeError':
-        # Even though repair_pdf will fail, ruffus will still try
-        # to call split_pages with no input files, likely due to a bug
-        if task_name == 'split_pages':
-            log.error("Input file '{0}' is not a valid PDF".format(
-                options.input_file))
-            exit_code = ExitCode.input_file
-    elif exc_name == 'builtins.KeyboardInterrupt':
-        log.error("Interrupted by user")
+    elif exc_name in ('builtins.KeyboardInterrupt', 'KeyboardInterrupt'):
+        # We have to print in this case because the log daemon might be toast
+        print("Interrupted by user", file=sys.stderr)
         exit_code = ExitCode.ctrl_c
     elif exc_name == 'subprocess.CalledProcessError':
         # It's up to the subprocess handler to report something useful
         msg = "Error occurred while running this command:"
         log.error(msg + '\n' + exc_value)
         exit_code = ExitCode.child_process_error
-    elif (exc_name == 'PyPDF2.utils.PdfReadError' and \
-            'not been decrypted' in exc_value) or \
-            (exc_name == 'ocrmypdf.exceptions.EncryptedPdfError'):
-        log.error(textwrap.dedent("""\
-            Input PDF is encrypted. The encryption must be removed to
-            perform OCR. 
-
-            For information about this PDF's security use
-                qpdf --show-encryption infilename
-
-            You can remove the encryption using
-                qpdf --decrypt [--password=[password]] infilename
-
-            """))
-        exit_code = ExitCode.encrypted_pdf        
-    elif exc_name == 'ocrmypdf.exceptions.PdfMergeFailedError':
-        log.error(textwrap.dedent("""\
-            Failed to merge PDF image layer with OCR layer
-
-            Usually this happens because the input PDF file is mal-formed and
-            ocrmypdf cannot automatically correct the problem on its own.
-
-            Try using
-                ocrmypdf --pdf-renderer tesseract  [..other args..]
-            """))
-        exit_code = ExitCode.input_file
     elif exc_name.startswith('ocrmypdf.exceptions.'):
-        base_exc_name = exc_name.replace('ocrmypdf.exceptions.', '')
-        exc_class = getattr(ocrmypdf_exceptions, base_exc_name)
-        exit_code = exc_class.exit_code
+        if exc_msg:
+            log.error(exc_msg)
     elif exc_name == 'PIL.Image.DecompressionBombError':
         msg = cleanup_ruffus_error_message(exc_value)
         msg += ("\nUse the --max-image-mpixels argument to set increase the "
@@ -638,33 +729,30 @@ def do_ruffus_exception(ruffus_five_tuple, options, log):
     return ExitCode.other_error
 
 
-def traverse_ruffus_exception(e_args, options, log):
-    """Walk through a RethrownJobError and find the first exception.
+def traverse_ruffus_exception(exceptions, options, log):
+    """Traverse a RethrownJobError and output the exceptions
 
-    Ruffus flattens exception to 5 element tuples. Because of a bug
-    in <= 2.6.3 it may present either the single:
-      (task, job, exc, value, stack)
-    or something like:
-      [[(task, job, exc, value, stack)]]
-    
-    Generally cross-process exception marshalling doesn't work well
-    and ruffus doesn't support because BaseException has its own
-    implementation of __reduce__ that attempts to reconstruct the
-    exception based on e.__init__(e.args).
-    
-    Attempting to log the exception directly marshalls it to the logger
-    which is probably in another process, so it's better to log only
-    data from the exception at this point.
+    Ruffus presents exceptions as 5 element tuples. The RethrownJobException
+    has a list of exceptions like
+        e.job_exceptions = [(5-tuple), (5-tuple), ...]
+
+    ruffus < 2.7.0 had a bug with exception marshalling that would give
+    different output whether the main or child process raised the exception.
+    We no longer support this.
+
+    Attempting to log the exception itself will re-marshall it to the logger
+    which is normally running in another process. It's better to avoid re-
+    marshalling.
 
     The exit code will be based on this, even if multiple exceptions occurred
     at the same time."""
 
-    if isinstance(e_args, Sequence) and isinstance(e_args[0], str) and \
-            len(e_args) == 5:
-        return do_ruffus_exception(e_args, options, log)
-    elif is_iterable_notstr(e_args):
-        for exc in e_args:
-            return traverse_ruffus_exception(exc, options, log)
+    exit_codes = []
+    for exc in exceptions:
+        exit_code = do_ruffus_exception(exc, options, log)
+        exit_codes.append(exit_code)
+
+    return exit_codes[0]  # Multiple codes are rare so take the first one
 
 
 def check_closed_streams(options):
@@ -733,11 +821,7 @@ def preamble(_log):
     _log.debug('ocrmypdf ' + VERSION)
     _log.debug('tesseract ' + tesseract.version())
     _log.debug('qpdf ' + qpdf.version())
-    if fitz:
-        _log.debug('PyMuPDF ' + fitz.version[0])
-        _log.debug('libmupdf ' + fitz.version[1])
-    else:
-        _log.debug('PyMuPDF not installed')
+    _log.debug('gs ' + ghostscript.version())
 
 
 def check_environ(options, _log):
@@ -749,7 +833,7 @@ def check_environ(options, _log):
     for k in old_envvars:
         if k in os.environ:
             _log.warning(textwrap.dedent("""\
-                OCRmyPDF no longer uses the environment variable {}. 
+                OCRmyPDF no longer uses the environment variable {}.
                 Change PATH to select alternate programs.""".format(k)))
 
 
@@ -792,14 +876,12 @@ def report_output_file_size(options, _log, input_file, output_file):
     ratio = output_size / input_size
     if ratio < 1.35 or input_size < 25000:
         return  # Seems fine
-    
+
     reasons = []
-    if not fitz:
-        reasons.append("The optional dependency PyMuPDF is not installed.")
     image_preproc = {
-        'deskew', 
-        'clean_final', 
-        'remove_background', 
+        'deskew',
+        'clean_final',
+        'remove_background',
         'oversample',
         'force_ocr'
     }
@@ -824,9 +906,11 @@ def report_output_file_size(options, _log, input_file, output_file):
         """.format(ratio, explanation)))
 
 
-def run_pipeline():
-    options = parser.parse_args()
+def run_pipeline(args=None):
+    options = parser.parse_args(args=args)
     options.verbose_abbreviated_path = 1
+    if os.environ.get('_OCRMYPDF_THREADS'):
+        options.use_threads = True
 
     if not check_closed_streams(options):
         return ExitCode.bad_args
@@ -847,6 +931,14 @@ def run_pipeline():
             "security vulnerabilities with certain malformed PDFs. Consider "
             "upgrading to version 7.0.0 or newer.".format(qpdf.version()))
 
+    if ghostscript.version() == '9.24':
+        complain(
+            "Ghostscript 9.24 contains serious regressions and is not "
+            "supported. Please upgrade to Ghostscript 9.25 or use an older "
+            "version."
+        )
+        return ExitCode.missing_dependency
+
     # Any changes to options will not take effect for options that are already
     # bound to function parameters in the pipeline. (For example
     # options.input_file, options.pdf_renderer are already bound.)
@@ -855,9 +947,10 @@ def run_pipeline():
 
     # Performance is improved by setting Tesseract to single threaded. In tests
     # this gives better throughput than letting a smaller number of Tesseract
-    # jobs run multithreaded.
-    if tesseract.v4():
-        os.environ.setdefault('OMP_THREAD_LIMIT', '1')
+    # jobs run multithreaded. Same story for pngquant. Tess <4 ignores this
+    # variable, but harmless to set if ignored.
+    os.environ.setdefault('OMP_THREAD_LIMIT', '1')
+
     check_environ(options, _log)
     if os.environ.get('PYTEST_CURRENT_TEST'):
         os.environ['_OCRMYPDF_TEST_INFILE'] = options.input_file
@@ -886,7 +979,8 @@ def run_pipeline():
     except ruffus_exceptions.RethrownJobError as e:
         if options.verbose:
             _log.debug(str(e))  # stringify exception so logger doesn't have to
-        exitcode = traverse_ruffus_exception(e.args, options, _log)
+        exceptions = e.job_exceptions
+        exitcode = traverse_ruffus_exception(exceptions, options, _log)
         if exitcode is None:
             _log.error("Unexpected ruffus exception: " + str(e))
             _log.error(repr(e))
@@ -895,7 +989,7 @@ def run_pipeline():
     except ExitCodeException as e:
         return e.exit_code
     except Exception as e:
-        _log.error(e)
+        _log.error(str(e))
         return ExitCode.other_error
 
     if options.flowchart:
@@ -914,12 +1008,12 @@ def run_pipeline():
             else:
                 msg = 'Output file is okay but is not PDF/A (seems to be {})'
                 _log.warning(msg.format(pdfa_info['conformance']))
-                return ExitCode.invalid_output_pdf
+                return ExitCode.pdfa_conversion_failed
         if not qpdf.check(options.output_file, _log):
             _log.warning('Output file: The generated PDF is INVALID')
             return ExitCode.invalid_output_pdf
 
-        report_output_file_size(options, _log, start_input_file, 
+        report_output_file_size(options, _log, start_input_file,
                                 options.output_file)
 
     pdfinfo = context.get_pdfinfo()

src/ocrmypdf/_jobcontext.py

@@ -0,0 +1,84 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from contextlib import suppress
+from multiprocessing.managers import SyncManager
+import sys
+import shutil
+
+from .pdfinfo import PdfInfo
+
+
+class JobContext:
+    """Holds our context for a particular run of the pipeline
+
+    A multiprocessing manager effectively creates a separate process
+    that keeps the master job context object.  Other threads access
+    job context via multiprocessing proxy objects.
+
+    While this would naturally lend itself @property's it seems to make
+    a little more sense to use functions to make it explicitly that the
+    invocation requires marshalling data across a process boundary.
+
+    """
+
+    def __init__(self):
+        self.pdfinfo = None
+        self.options = None
+        self.work_folder = None
+        self.rotations = {}
+
+    def generate_pdfinfo(self, infile):
+        self.pdfinfo = PdfInfo(infile)
+
+    def get_pdfinfo(self):
+        "What we know about the input PDF"
+        return self.pdfinfo
+
+    def set_pdfinfo(self, pdfinfo):
+        self.pdfinfo = pdfinfo
+
+    def get_options(self):
+        return self.options
+
+    def set_options(self, options):
+        self.options = options
+
+    def get_work_folder(self):
+        return self.work_folder
+
+    def set_work_folder(self, work_folder):
+        self.work_folder = work_folder
+
+    def get_rotation(self, pageno):
+        return self.rotations.get(pageno, 0)
+
+    def set_rotation(self, pageno, value):
+        self.rotations[pageno] = value
+
+
+class JobContextManager(SyncManager):
+    pass
+
+
+def cleanup_working_files(work_folder, options):
+    if options.keep_temporary_files:
+        print("Temporary working files saved at:\n{0}".format(work_folder),
+              file=sys.stderr)
+    else:
+        with suppress(FileNotFoundError):
+            shutil.rmtree(work_folder)

src/ocrmypdf/_unicodefun.py

@@ -38,8 +38,13 @@ import sys
 import codecs
 
 
-def verify_python3_env():
+def verify_python3_env():  # pragma: no cover
     """Ensures that the environment is good for unicode on Python 3."""
+
+    # PEP 538 changes in Python 3.7 should make this wrangling unnecessary
+    if sys.version_info[0:3] >= (3, 7, 0):
+        return
+
     try:
         import locale
         fs_enc = codecs.lookup(locale.getpreferredencoding()).name

src/ocrmypdf/_weave.py

@@ -0,0 +1,418 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from pathlib import Path
+from itertools import groupby
+
+import pikepdf
+
+from .helpers import flatten_groups, page_number
+from .exec import tesseract
+
+
+def _update_page_resources(*, page, font, font_key, procset):
+    "Update this page's fonts with a reference to the Glyphless font"
+
+    if '/Resources' not in page:
+        page['/Resources'] = pikepdf.Dictionary({})
+    resources = page['/Resources']
+    try:
+        fonts = resources['/Font']
+    except KeyError:
+        fonts = pikepdf.Dictionary({})
+    if font_key not in fonts:
+        fonts[font_key] = font
+    resources['/Font'] = fonts
+
+    # Reassign /ProcSet to one that just lists everything - ProcSet is
+    # obsolete and doesn't matter but recommended for old viewer support
+    resources['/ProcSet'] = procset
+
+
+def strip_invisible_text(pdf, page, log):
+    stream = []
+    in_text_obj = False
+    render_mode = 0
+    text_objects = []
+
+    page.page_contents_coalesce()
+    for operands, operator in pikepdf.parse_content_stream(page, ''):
+        if not in_text_obj:
+            if operator == pikepdf.Operator('BT'):
+                in_text_obj = True
+                render_mode = 0
+                text_objects.append((operands, operator))
+            else:
+                stream.append((operands, operator))
+        else:
+            if operator == pikepdf.Operator('Tr'):
+                render_mode = operands[0]
+            text_objects.append((operands, operator))
+            if operator == pikepdf.Operator('ET'):
+                in_text_obj = False
+                if render_mode != 3:
+                    stream.extend(text_objects)
+                text_objects.clear()
+
+    def convert(op):
+        try:
+            return op.unparse()
+        except AttributeError:
+            return str(op).encode('ascii')
+
+    lines = []
+
+    for operands, operator in stream:
+        if operator == pikepdf.Operator('INLINE IMAGE'):
+            iim = operands[0]
+            line = iim.unparse()
+        else:
+            line = b' '.join(convert(op) for op in operands) + b' ' + operator.unparse()
+        lines.append(line)
+
+    content_stream = b'\n'.join(lines)
+    page.Contents = pikepdf.Stream(pdf, content_stream)
+
+
+def _weave_layers_graft(
+        *, pdf_base, page_num, text, font, font_key, procset, rotation,
+        strip_old_text, log):
+    """Insert the text layer from text page 0 on to pdf_base at page_num"""
+
+    log.debug("Grafting")
+    if Path(text).stat().st_size == 0:
+        return
+
+    # This is a pointer indicating a specific page in the base file
+    pdf_text = pikepdf.open(text)
+    pdf_text_contents = pdf_text.pages[0].Contents.read_bytes()
+
+    if not tesseract.has_textonly_pdf():
+        # If we don't have textonly_pdf, edit the stream to delete the
+        # instruction to draw the image Tesseract generated, which we do not
+        # use.
+        stream = bytearray(pdf_text_contents)
+        pattern = b'/Im1 Do'
+        idx = stream.find(pattern)
+        stream[idx:(idx + len(pattern))] = b' ' * len(pattern)
+        pdf_text_contents = bytes(stream)
+
+    base_page = pdf_base.pages.p(page_num)
+
+    # The text page always will be oriented up by this stage but the original
+    # content may have a rotation applied. Wrap the text stream with a rotation
+    # so it will be oriented the same way as the rest of the page content.
+    # (Previous versions OCRmyPDF rotated the content layer to match the text.)
+    mediabox = [float(pdf_text.pages[0].MediaBox[v])
+                for v in range(4)]
+    wt, ht = mediabox[2] - mediabox[0], mediabox[3] - mediabox[1]
+
+    mediabox = [float(base_page.MediaBox[v])
+                for v in range(4)]
+    wp, hp = mediabox[2] - mediabox[0], mediabox[3] - mediabox[1]
+
+    translate = pikepdf.PdfMatrix().translated(-wt / 2, -ht / 2)
+    untranslate = pikepdf.PdfMatrix().translated(wp / 2, hp / 2)
+    # -rotation because the input is a clockwise angle and this formula
+    # uses CCW
+    rotation = -rotation % 360
+    rotate = pikepdf.PdfMatrix().rotated(rotation)
+
+    # Because of rounding of DPI, we might get a text layer that is not
+    # identically sized to the target page. Scale to adjust. Normally this
+    # is within 0.998.
+    if rotation in (90, 270):
+        wt, ht = ht, wt
+    scale_x = wp / wt
+    scale_y = hp / ht
+
+    log.debug('%r', (scale_x, scale_y))
+    scale = pikepdf.PdfMatrix().scaled(scale_x, scale_y)
+
+    # Translate the text so it is centered at (0, 0), rotate it there, adjust
+    # for a size different between initial and text PDF, then untranslate
+    ctm = translate @ rotate @ scale @ untranslate
+
+    pdf_text_contents = (
+        b'q %s cm\n' % ctm.encode() +
+        pdf_text_contents +
+        b'\nQ\n'
+    )
+
+    new_text_layer = pikepdf.Stream(pdf_base, pdf_text_contents)
+
+    if strip_old_text:
+        strip_invisible_text(pdf_base, base_page, log)
+
+    base_page.page_contents_add(new_text_layer, prepend=True)
+
+    _update_page_resources(
+        page=base_page, font=font, font_key=font_key, procset=procset
+    )
+
+
+def _find_font(text, pdf_base):
+    "Copy a font from the filename text into pdf_base"
+
+    font, font_key = None, None
+    possible_font_names = ('/f-0-0', '/F1')
+    try:
+        pdf_text = pikepdf.open(text)
+        pdf_text_fonts = pdf_text.pages[0].Resources.get('/Font', {})
+    except Exception:
+        return None, None
+
+    for f in possible_font_names:
+        pdf_text_font = pdf_text_fonts.get(f, None)
+        if pdf_text_font is not None:
+            font_key = f
+            break
+    if pdf_text_font:
+        font = pdf_base.copy_foreign(pdf_text_font)
+    return font, font_key
+
+
+def _traverse_toc(pdf_base, visitor_fn, log):
+    """
+    Walk the table of contents, calling visitor_fn() at each node
+
+    The /Outlines data structure is a messy data structure, but rather than
+    navigating hierarchically we just track unique nodes.  Enqueue nodes when
+    we find them, and never visit them again.  set() is awesome.  We look for
+    the two types of object in the table of contents that can be page bookmarks
+    and update the page entry.
+
+    """
+
+    visited = set()
+    queue = set()
+    link_keys = ('/Parent', '/First', '/Last', '/Prev', '/Next')
+
+    if not '/Outlines' in pdf_base.root:
+        return
+
+    queue.add(pdf_base.root.Outlines.objgen)
+    while queue:
+        objgen = queue.pop()
+        visited.add(objgen)
+        node = pdf_base.get_object(objgen)
+        log.debug('fix toc: exploring outline entries at %r', objgen)
+
+        # Enumerate other nodes we could visit from here
+        for key in link_keys:
+            if key not in node:
+                continue
+            item = node[key]
+            if not item.is_indirect:
+                # Direct references are not allowed here, but it's not clear
+                # what we should do if we find any. Removing them is an option:
+                # node[key] = pdf_base.make_indirect(None)
+                continue
+            objgen = item.objgen
+            if objgen not in visited:
+                queue.add(objgen)
+
+        if visitor_fn:
+            visitor_fn(pdf_base, node, log)
+
+
+def _fix_toc(pdf_base, pageref_remap, log):
+    """Repair the table of contents
+
+    Whenever we replace a page wholesale, it gets assigned a new objgen number
+    and other references to it within the PDF become invalid, most notably in
+    the table of contents (/Outlines in PDF-speak).  In weave_layers we collect
+    pageref_remap, a mapping that describes the new objgen number given an old
+    one.  (objgen is a tuple, and the gen is almost always zero.)
+
+    It may ultimately be better to find a way to rebuild a page in place.
+
+    """
+
+    if not pageref_remap:
+        return
+
+    def remap_dest(dest_node):
+        """
+        Inner helper function: change the objgen for any page from the old we
+        invalidated to its new one.
+        """
+        if not isinstance(dest_node, pikepdf.Array):
+            return
+        pageref = dest_node[0]
+        if pageref['/Type'] == '/Page' and \
+                pageref.objgen in pageref_remap:
+            new_objgen = pageref_remap[pageref.objgen]
+            dest_node[0] = pdf_base.get_object(new_objgen)
+
+    def visit_remap_dest(pdf_base, node, log):
+        """
+        Visitor function to fix ToC entries
+
+        Test for the two types of references to pages that can occur in ToCs.
+        Both types have the same final format (an indirect reference to the
+        target page).
+        """
+        if '/Dest' in node:
+            # /Dest reference to another page (old method)
+            remap_dest(node['/Dest'])
+        elif '/A' in node:
+            # /A (action) command set to "GoTo" (newer method)
+            if '/S' in node['/A'] and node['/A']['/S'] == '/GoTo':
+                remap_dest(node['/A']['/D'])
+
+    _traverse_toc(pdf_base, visit_remap_dest, log)
+
+
+def weave_layers(
+        infiles,
+        output_file,
+        log,
+        context):
+    """Apply text layer and/or image layer changes to baseline file
+
+    This is where the magic happens. infiles will be the main PDF to modify,
+    and optional .text.pdf and .image-layer.pdf files, organized however ruffus
+    organizes them.
+
+    From .text.pdf, we copy the content stream (which contains the Tesseract
+    OCR results), and rotate it into place. The first time we do this, we also
+    copy the GlyphlessFont, and then reference that font again.
+
+    For .image-layer.pdf, we check if this is a "pointer" to the original file,
+    or a new file. If a new file, we replace the page and remember that we
+    replaced this page.
+
+    Every 100 open files, we save intermediate results, to avoid any resource
+    limits, since pikepdf/qpdf need to keep a lot of open file handles in the
+    background. When objects are copied from one file to another qpdf, qpdf
+    doesn't actually copy the data until asked to write, so all the resources
+    it may need to remain available.
+
+    For completeness, we set up a /ProcSet on every page, although it's
+    unlikely any PDF viewer cares about this anymore.
+
+    """
+
+    def input_sorter(key):
+        try:
+            return page_number(key)
+        except ValueError:
+            return -1
+    flat_inputs = sorted(flatten_groups(infiles), key=input_sorter)
+    groups = groupby(flat_inputs, key=input_sorter)
+
+    # Extract first item
+    _, basegroup = next(groups)
+    base = list(basegroup)[0]
+    path_base = Path(base).resolve()
+    pdf_base = pikepdf.open(path_base)
+    keep_open = []
+    font, font_key, procset = None, None, None
+    pdfinfo = context.get_pdfinfo()
+    pagerefs = {}
+
+    # Walk the table of contents first, to trigger pikepdf/qpdf to resolve all
+    # page references in the table of contents. Some PDF generators put invalid
+    # references in the ToC, so we want to resolve them to null before we
+    # create any references, or the ToC will be corrupted
+    _traverse_toc(pdf_base, None, log)
+
+    procset = pdf_base.make_indirect(
+        pikepdf.Object.parse(b'[ /PDF /Text /ImageB /ImageC /ImageI ]'))
+
+    # Iterate rest
+    for page_num, layers in groups:
+        layers = list(layers)
+        log.debug(page_num)
+        log.debug(layers)
+
+        text = next(
+            (ii for ii in layers if ii.endswith('.text.pdf')), None
+        )
+        image = next(
+            (ii for ii in layers if ii.endswith('.image-layer.pdf')), None
+        )
+
+        if text and not font:
+            font, font_key = _find_font(text, pdf_base)
+
+        replacing = False
+        content_rotation = pdfinfo[page_num - 1].rotation
+
+        path_image = Path(image).resolve() if image else None
+        if path_image is not None and path_image != path_base:
+            # We are replacing the old page with a rasterized PDF of the new
+            # page
+            log.debug("Replace")
+            old_objgen = pdf_base.pages[page_num - 1].objgen
+
+            pdf_image = pikepdf.open(image)
+            keep_open.append(pdf_image)
+            image_page = pdf_image.pages[0]
+            pdf_base.pages[page_num - 1] = image_page
+
+            # We're adding a new page, which will get a new objgen number pair,
+            # so we need to update any references to it.  qpdf did not like
+            # my attempt to update the old object in place, but that is an
+            # option to consider
+            pagerefs[old_objgen] = pdf_base.pages[page_num - 1].objgen
+            replacing = True
+
+        autorotate_correction = context.get_rotation(page_num - 1)
+        if replacing:
+            content_rotation = autorotate_correction
+        text_rotation = autorotate_correction
+        text_misaligned = (text_rotation - content_rotation) % 360
+        log.debug('%r', [
+            text_rotation, autorotate_correction, text_misaligned,
+            content_rotation]
+        )
+
+        if text and font:
+            # Graft the text layer onto this page, whether new or old
+            strip_old = context.get_options().redo_ocr
+            _weave_layers_graft(
+                pdf_base=pdf_base, page_num=page_num, text=text, font=font,
+                font_key=font_key, rotation=text_misaligned, procset=procset,
+                strip_old_text=strip_old, log=log
+            )
+
+        # Correct the rotation if applicable
+        pdf_base.pages[page_num - 1].Rotate = \
+            (content_rotation - autorotate_correction) % 360
+
+        if len(keep_open) > 100:
+            # qpdf limitations require us to keep files open when we intend
+            # to copy content from them before saving. However, we want to keep
+            # a lid on file handles and memory usage, so for big files we're
+            # going to stop and save periodically. Attach the font to page 1
+            # even if page 1 doesn't use it, so we have a way to get it back.
+            page0 = pdf_base.pages[0]
+            _update_page_resources(
+                page=page0, font=font, font_key=font_key, procset=procset)
+            interim = output_file + '_working{}.pdf'.format(page_num)
+            pdf_base.save(interim)
+            del pdf_base
+            keep_open = []
+
+            pdf_base = pikepdf.open(interim)
+            procset = pdf_base.pages[0].Resources.ProcSet
+            font = pdf_base.pages[0].Resources.Font.get(font_key)
+
+    _fix_toc(pdf_base, pagerefs, log)
+    pdf_base.save(output_file)

src/ocrmypdf/exceptions.py

@@ -17,6 +17,7 @@
 
 
 from enum import IntEnum
+from textwrap import dedent
 
 class ExitCode(IntEnum):
     ok = 0
@@ -29,12 +30,20 @@ class ExitCode(IntEnum):
     child_process_error = 7
     encrypted_pdf = 8
     invalid_config = 9
+    pdfa_conversion_failed = 10
     other_error = 15
     ctrl_c = 130
 
 
 class ExitCodeException(Exception):
     exit_code = ExitCode.other_error
+    message = ""
+
+    def __str__(self):
+        super_msg = super().__str__()  # Don't do str(super())
+        if self.message:
+            return self.message.format(super_msg)
+        return super_msg
 
 
 class BadArgsError(ExitCodeException):
@@ -43,7 +52,15 @@ class BadArgsError(ExitCodeException):
 
 class PdfMergeFailedError(ExitCodeException):
     exit_code = ExitCode.input_file
+    message = dedent('''\
+        Failed to merge PDF image layer with OCR layer
 
+        Usually this happens because the input PDF file is malformed and
+        ocrmypdf cannot automatically correct the problem on its own.
+
+        Try using
+            ocrmypdf --pdf-renderer sandwich  [..other args..]
+        ''')
 
 class MissingDependencyError(ExitCodeException):
     exit_code = ExitCode.missing_dependency
@@ -75,7 +92,18 @@ class SubprocessOutputError(ExitCodeException):
 
 class EncryptedPdfError(ExitCodeException):
     exit_code = ExitCode.encrypted_pdf
+    message = dedent('''\
+        Input PDF is encrypted. The encryption must be removed to
+        perform OCR.
+
+        For information about this PDF's security use
+            qpdf --show-encryption infilename
+
+        You can remove the encryption using
+            qpdf --decrypt [--password=[password]] infilename
+        ''')
 
 
 class TesseractConfigError(ExitCodeException):
     exit_code = ExitCode.invalid_config
+    message = "Error occurred while parsing a Tesseract configuration file"

src/ocrmypdf/exec/__init__.py

@@ -24,8 +24,7 @@ from subprocess import run, STDOUT, PIPE, CalledProcessError
 from ..exceptions import MissingDependencyError
 
 
-
-def get_version(program, *, 
+def get_version(program, *,
         version_arg='--version', regex=r'(\d+(\.\d+)*)'):
     "Get the version of the specified program"
     args_prog = [
@@ -37,6 +36,10 @@ def get_version(program, *,
             args_prog, close_fds=True, universal_newlines=True,
             stdout=PIPE, stderr=STDOUT, check=True)
         output = proc.stdout
+    except FileNotFoundError as e:
+        raise MissingDependencyError(
+            "Could not find program '{}' on the PATH".format(
+                program)) from e
     except CalledProcessError as e:
         if e.returncode < 0:
             raise MissingDependencyError(

src/ocrmypdf/exec/ghostscript.py

@@ -16,14 +16,13 @@
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
 from tempfile import NamedTemporaryFile
-from subprocess import run, PIPE, STDOUT, CalledProcessError
+from subprocess import run, PIPE, STDOUT
 from shutil import copy
 from functools import lru_cache
 import re
-import sys
 from PIL import Image
 from . import get_version
-from ..exceptions import SubprocessOutputError, MissingDependencyError
+from ..exceptions import SubprocessOutputError
 from ..helpers import fspath
 
 
@@ -33,44 +32,99 @@ def version():
 
 
 def jpeg_passthrough_available():
-    """
-    Ghostscript 9.23 introduced JPEG passthrough but it seems to corrupt the
-    last two bytes of certain images, for now we disable it for 9.23 and
-    do not mention it for < 9.23.
+    """Returns True if the installed version of Ghostscript supports JPEG passthru
 
+    Prior to 9.23, Ghostscript decode and re-encoded JPEGs internally. In 9.23
+    it gained the ability to keep JPEGs unmodified. However, the 9.23
+    implementation was buggy and would deletes the last two bytes of images in
+    some cases, as reported here.
     https://bugs.ghostscript.com/show_bug.cgi?id=699216
 
+    The issue was fixed for 9.24, hence that is the first version we consider
+    the feature available. (However, we don't use 9.24 at all, so the first
+    version that allows JPEG passthrough is 9.25.
+
     """
-    return False
+    return version() >= '9.24'
 
 
 def _gs_error_reported(stream):
     return re.search(r'error', stream, flags=re.IGNORECASE)
 
 
-def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
-                  pageno=1, page_dpi=None):
+def extract_text(input_file, pageno=1):
+    """Use the txtwrite device to get text layout information out
+
+    For details on options of -dTextFormat see
+    https://www.ghostscript.com/doc/current/VectorDevices.htm#TXT
+
+    Format is like
+    <page>
+    <line>
+    <span bbox="left top right bottom" font="..." size="...">
+    <char bbox="...." c="X"/>
+
+    :param pageno: number of page to extract, or all pages if None
+    :return: XML-ish text representation in bytes
     """
-    Rasterize one page of a PDF at resolution (xres, yres) in canvas units.
-    
-    The image is sized to match the integer pixels dimensions implied by 
+
+    if pageno is not None:
+        pages = [
+            '-dFirstPage=%i' % pageno,
+            '-dLastPage=%i' % pageno
+        ]
+    else:
+        pages = []
+
+    args_gs = [
+        'gs',
+        '-dQUIET',
+        '-dSAFER',
+        '-dBATCH',
+        '-dNOPAUSE',
+        '-sDEVICE=txtwrite',
+        '-dTextFormat=0',
+    ] + pages + [
+        '-o', '-',
+        fspath(input_file)
+    ]
+
+    p = run(args_gs, stdout=PIPE, stderr=PIPE)
+    if p.returncode != 0:
+        raise SubprocessOutputError(
+            'Ghostscript text extraction failed\n%s\n%s\n%s' % (
+                input_file, p.stdout.decode(), p.stderr.decode()
+            )
+        )
+
+    return p.stdout
+
+
+def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
+                  pageno=1, page_dpi=None, rotation=None, filter_vector=False):
+    """Rasterize one page of a PDF at resolution (xres, yres) in canvas units.
+
+    The image is sized to match the integer pixels dimensions implied by
     (xres, yres) even if those numbers are noninteger. The image's DPI will
      be overridden with the values in page_dpi.
-    
+
     :param input_file: pathlike
     :param output_file: pathlike
     :param xres: resolution at which to rasterize page
-    :param yres: 
-    :param raster_device: 
-    :param log: 
+    :param yres:
+    :param raster_device:
+    :param log:
     :param pageno: page number to rasterize (beginning at page 1)
-    :param page_dpi: resolution tuple (x, y) overriding output image DPI 
-    :return: 
+    :param page_dpi: resolution tuple (x, y) overriding output image DPI
+    :param rotation: 0, 90, 180, 270: clockwise angle to rotate page
+    :param filter_vector: if True, remove vector graphics objects
+    :return:
     """
     res = xres, yres
     int_res = round(xres), round(yres)
     if not page_dpi:
         page_dpi = res
+
     with NamedTemporaryFile(delete=True) as tmp:
         args_gs = [
             'gs',
@@ -82,10 +136,14 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
             '-dFirstPage=%i' % pageno,
             '-dLastPage=%i' % pageno,
             '-r{0}x{1}'.format(str(int_res[0]), str(int_res[1])),
+        ] + (['-dFILTERVECTOR'] if filter_vector else []) + [
             '-o', tmp.name,
+            '-dAutoRotatePages=/None',  # Probably has no effect on raster
+            '-f',
             fspath(input_file)
         ]
-        
+
+        log.debug(args_gs)
         p = run(args_gs, stdout=PIPE, stderr=STDOUT,
                 universal_newlines=True)
         if _gs_error_reported(p.stdout):
@@ -110,14 +168,42 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
                 log.debug(
                     "Ghostscript: resize output image {} -> {}".format(
                         im.size, expected_size))
-                im.resize(expected_size).save(
-                    fspath(output_file), dpi=page_dpi)
-            else:
-                copy(tmp.name, fspath(output_file))
+                im = im.resize(expected_size)
+
+            if rotation is not None:
+                log.debug("Rotating output by %i", rotation)
+                # rotation is a clockwise angle and Image.ROTATE_* is
+                # counterclockwise so this cancels out the rotation
+                if rotation == 90:
+                    im = im.transpose(Image.ROTATE_90)
+                elif rotation == 180:
+                    im = im.transpose(Image.ROTATE_180)
+                elif rotation == 270:
+                    im = im.transpose(Image.ROTATE_270)
+                if rotation % 180 == 90:
+                    page_dpi = page_dpi[1], page_dpi[0]
+            im.save(fspath(output_file), dpi=page_dpi)
 
 
 def generate_pdfa(pdf_pages, output_file, compression, log,
                   threads=1, pdf_version='1.5', pdfa_part='2'):
+    """Generate a PDF/A.
+
+    The pdf_pages, a list files, will be merged into output_file. One or more
+    PDF files may be merged. One of the files in this list must be a pdfmark
+    file that provides Ghostscript with details on how to perform the PDF/A
+    conversion. By default with we pick PDF/A-2b, but this works for 1 or 3.
+
+    compression can be 'jpeg', 'lossless', or an empty string. In 'jpeg',
+    Ghostscript is instructed to convert color and grayscale images to DCT
+    (JPEG encoding). In 'lossless' Ghostscript is told to convert images to
+    Flate (lossless/PNG). If the parameter is omitted Ghostscript is left to
+    make its own decisions about how to encode images; it appears to use a
+    heuristic to decide how to encode images. As of Ghostscript 9.25, we
+    support passthrough JPEG which allows Ghostscript to avoid transcoding
+    images entirely. (The feature was added in 9.23 but broken, and the 9.24
+    release of Ghostscript had regressions, so we don't support it until 9.25.)
+    """
     compression_args = []
     if compression == 'jpeg':
         compression_args = [
@@ -139,8 +225,8 @@ def generate_pdfa(pdf_pages, output_file, compression, log,
             "-dAutoFilterGrayImages=true",
         ]
 
-    # Older versions of Ghostscript expect a leading slash in 
-    # sColorConversionStrategy, newer ones should not have it. See Ghostscript 
+    # Older versions of Ghostscript expect a leading slash in
+    # sColorConversionStrategy, newer ones should not have it. See Ghostscript
     # git commit fe1c025d.
     strategy = 'RGB' if version() >= '9.19' else '/RGB'
 
@@ -151,6 +237,9 @@ def generate_pdfa(pdf_pages, output_file, compression, log,
         compression_args.append('-dPassThroughJPEGImages=false')
 
     with NamedTemporaryFile(delete=True) as gs_pdf:
+        # nb no need to specify ProcessColorModel when ColorConversionStrategy
+        # is set; see:
+        # https://bugs.ghostscript.com/show_bug.cgi?id=699392
         args_gs = [
             "gs",
             "-dQUIET",
@@ -160,15 +249,15 @@ def generate_pdfa(pdf_pages, output_file, compression, log,
             "-dNumRenderingThreads=" + str(threads),
             "-sDEVICE=pdfwrite",
             "-dAutoRotatePages=/None",
-            "-sColorConversionStrategy=" + strategy,
-            "-sProcessColorModel=DeviceRGB"
+            "-sColorConversionStrategy=" + strategy
         ] + compression_args + [
             "-dJPEGQ=95",
             "-dPDFA=" + pdfa_part,
             "-dPDFACompatibilityPolicy=1",
             "-sOutputFile=" + gs_pdf.name,
         ]
-        args_gs.extend(pdf_pages)
+        args_gs.extend(fspath(s) for s in pdf_pages)  # Stringify Path objs
+        log.debug(args_gs)
         p = run(args_gs, stdout=PIPE, stderr=STDOUT,
                 universal_newlines=True)
 
@@ -188,7 +277,7 @@ def generate_pdfa(pdf_pages, output_file, compression, log,
         if p.returncode == 0:
             # Ghostscript does not change return code when it fails to create
             # PDF/A - check PDF/A status elsewhere
-            copy(gs_pdf.name, output_file)
+            copy(gs_pdf.name, fspath(output_file))
         else:
             log.error('Ghostscript PDF/A rendering failed')
             raise SubprocessOutputError()

src/ocrmypdf/exec/jbig2enc.py

@@ -0,0 +1,63 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from subprocess import run, PIPE
+from functools import lru_cache
+
+from . import get_version
+from ..exceptions import MissingDependencyError
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('jbig2', regex=r'jbig2enc (\d+(\.\d+)*).*')
+
+
+def available():
+    try:
+        version()
+    except MissingDependencyError:
+        return False
+    return True
+
+
+def convert_group(*, cwd, infiles, out_prefix):
+    args = [
+        'jbig2',
+        '-b',
+        out_prefix,
+        '-s',  # symbol mode (lossy)
+        # '-r', # refinement mode (lossless symbol mode, currently disabled in
+                # jbig2)
+        '-p',
+    ]
+    args.extend(infiles)
+    proc = run(args, cwd=cwd, stdout=PIPE, stderr=PIPE)
+    proc.check_returncode()
+    return proc
+
+
+def convert_single(*, cwd, infile, outfile):
+    args = [
+        'jbig2',
+        '-p',
+        infile
+    ]
+    with open(outfile, 'wb') as fstdout:
+        proc = run(args, cwd=cwd, stdout=fstdout, stderr=PIPE)
+    proc.check_returncode()
+    return proc

src/ocrmypdf/exec/pngquant.py

@@ -0,0 +1,49 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from subprocess import run
+from functools import lru_cache
+
+from . import get_version
+from ..exceptions import MissingDependencyError
+
+
+@lru_cache(maxsize=1)
+def version():
+    return get_version('pngquant', regex=r'(\d+(\.\d+)*).*')
+
+
+def available():
+    try:
+        version()
+    except MissingDependencyError:
+        return False
+    return True
+
+
+def quantize(input_file, output_file, quality_min, quality_max):
+    args = [
+        'pngquant',
+        '--force',
+        '--skip-if-larger',
+        '--output', output_file,
+        '--quality', '{}-{}'.format(quality_min, quality_max),
+        '--',
+        input_file
+    ]
+    proc = run(args)
+    proc.check_returncode()

src/ocrmypdf/exec/qpdf.py

@@ -17,39 +17,33 @@
 
 from subprocess import CalledProcessError, STDOUT, PIPE, run
 from functools import lru_cache
-import sys
-import os
-import re
-import resource
 
-from ..exceptions import InputFileError, SubprocessOutputError, \
-    MissingDependencyError, EncryptedPdfError
 from . import  get_version
-from ..helpers import re_symlink
+from ..helpers import fspath
 
 
 @lru_cache(maxsize=1)
 def version():
-    return get_version('qpdf', regex=r'qpdf version (.+)')    
+    return get_version('qpdf', regex=r'qpdf version (.+)')
 
 
 def check(input_file, log=None):
     args_qpdf = [
         'qpdf',
         '--check',
-        input_file
+        fspath(input_file)
     ]
 
     if log is None:
         import logging as log
 
     try:
-        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
+        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True,
             check=True)
     except CalledProcessError as e:
         if e.returncode == 2:
-            log.error("{0}: not a valid PDF, and could not repair it.".format(
-                input_file))
+            log.error("%s: not a valid PDF, and could not repair it.",
+                input_file)
             log.error("Details:")
             log.error(e.output)
         elif e.returncode == 3:
@@ -59,169 +53,3 @@ def check(input_file, log=None):
             log.warning(e.output)
         return False
     return True
-
-
-def _probably_encrypted(e):
-    """qpdf can report a false positive "file is encrypted" message for damaged
-    files - suppress this"""
-    return e.returncode == 2 and \
-        'invalid password' in e.output and \
-        'file is damaged' not in e.output
-
-
-def repair(input_file, output_file, log):
-    args_qpdf = [
-        'qpdf', input_file, output_file
-    ]
-    try:
-        run(args_qpdf, stderr=STDOUT, stdout=PIPE, universal_newlines=True, 
-            check=True)
-    except CalledProcessError as e:
-        if e.returncode == 3 and e.output.find("operation succeeded"):
-            log.debug('qpdf found and fixed errors: ' + e.output)
-            return
-
-        if _probably_encrypted(e):
-            raise EncryptedPdfError() from e
-        elif e.returncode == 2:
-            log.error("{0}: not a valid PDF, and could not repair it.".format(
-                      input_file))
-            log.error("Details: " + e.output)
-            raise InputFileError() from e
-        else:
-            log.error("{0}: unknown error".format(
-                      input_file))
-            log.error(e.output)
-            raise SubprocessOutputError() from e
-
-
-def get_npages(input_file, log):
-    try:
-        pages = run(
-            ['qpdf', '--show-npages', input_file],
-            universal_newlines=True, check=True, stdout=PIPE, stderr=STDOUT)
-    except CalledProcessError as e:
-        if e.returncode == 2 and e.output.find('No such file'):
-            log.error(e.output)
-            raise InputFileError() from e
-    return int(pages)
-
-
-def split_pages(input_file, work_folder, npages):
-    """Split multipage PDF into individual pages.
-
-    Incredibly enough, this multiple process approach is about 70 times
-    faster than using Ghostscript.
-    """
-    for n in range(int(npages)):
-        args_qpdf = [
-            'qpdf', input_file,
-            '--pages', input_file, '{0}'.format(n + 1), '--',
-            os.path.join(work_folder, '{0:06d}.page.pdf'.format(n + 1))
-        ]
-        run(args_qpdf, check=True)
-
-
-def extract_page(input_file, output_file, pageno):
-    args_qpdf = [
-        'qpdf', input_file,
-        '--pages', input_file, '{0}'.format(pageno + 1), '--',
-        output_file
-    ]
-    run(args_qpdf, check=True)
-    
-
-def _merge_inner(input_files, output_file, min_version=None, log=None):
-    """Merge the list of input files (all filenames) into the output file.
-
-    The input files may contain one or more pages.
-    """
-
-    # Single page 'merges' should still be attempted to that the same error
-    # checking is applied to single page case
-
-    version_arg = ['--min-version={}'.format(min_version)] \
-                  if min_version else []
-
-    if log is None:
-        import logging as log
-
-    args_qpdf = [
-        'qpdf'
-    ] + version_arg + [
-        input_files[0], '--pages'
-    ] + input_files + ['--', output_file]
-
-    try:
-        run(args_qpdf, check=True, stderr=PIPE, universal_newlines=True)
-    except CalledProcessError as e:
-        if e.returncode == 3 and \
-                e.stderr.find("unknown token while reading object") and \
-                e.stderr.find("operation succeeded"):
-            # Only whitelist the 'unknown token' problem (decimal/string issue)
-            # qpdf issue #165
-            log.warning('qpdf found and fixed errors: ' + e.stderr)
-            return
-        raise e from e
-
-
-def merge(input_files, output_file, min_version=None, log=None, max_files=None):
-    """Merge the list of input files (all filenames) into the output file.
-
-    The input files may contain one or more pages.
-
-    """
-    # qpdf requires that every file that contributes to the output has a file 
-    # descriptor that remains open. That means, given our approach of one 
-    # intermediate PDF per, we can practically hit the number of file 
-    # descriptors. 
-
-    if max_files is None or max_files < 2:
-        # Find out how many open file descriptors we can get away with
-        ulimits = resource.getrlimit(resource.RLIMIT_NOFILE)
-        max_open_files = ulimits[0]
-        max_files = max_open_files // 2  # Conservative guess
-
-    # We'll write things alongside the output file
-    output_dir = os.path.dirname(output_file)
-
-    import random
-    import string    
-
-    def randstr():
-        return ''.join(random.sample(string.ascii_lowercase, 6))
-
-    # How many files to grab at once, merging all their contents
-    step_size = max_files
-
-    workqueue = input_files.copy()
-    counter = 1
-    next_workqueue = []
-    while len(workqueue) > 1 or len(next_workqueue) > 0:
-        # Take n files out of the queue
-        n = min(step_size, len(workqueue))
-        job = workqueue[0:n]
-        del workqueue[0:n]
-        log.debug('merging ' + repr(job))
-
-        # Merge them into 1 file, which will contain n^depth pages
-        merge_file = os.path.join(
-            output_dir, "merge-{:06d}-{}.pdf".format(counter, randstr()))
-        counter += 1
-        _merge_inner(job, merge_file, min_version=min_version, log=log)
-
-        # On the next 
-        next_workqueue.append(merge_file)
-        log.debug('next_workqueue ' + repr(next_workqueue))
-
-        # If we're out of things to do in this queue, move on to the next
-        # queue. On the counter-th pass of the workqueue we can chew through
-        # (step_size)**N pages, so on most systems the second pass finishes
-        # the job.
-        if len(workqueue) == 0:
-            workqueue = next_workqueue
-            next_workqueue = []
-
-    re_symlink(workqueue.pop(), output_file, log)
-
-

src/ocrmypdf/exec/tesseract.py

@@ -17,19 +17,15 @@
 
 import sys
 import os
-import re
 import shutil
 from functools import lru_cache
 from collections import namedtuple
 from textwrap import dedent
-from subprocess import PIPE, CalledProcessError, \
-    TimeoutExpired, check_output, STDOUT
+from subprocess import CalledProcessError, TimeoutExpired, check_output, STDOUT, run, PIPE
 from contextlib import suppress
 
-import PyPDF2 as pypdf
-
 from ..exceptions import MissingDependencyError, TesseractConfigError
-from ..helpers import page_number
+from ..helpers import page_number, fspath
 from . import get_version
 
 OrientationConfidence = namedtuple(
@@ -67,13 +63,14 @@ def v4():
 @lru_cache(maxsize=1)
 def has_textonly_pdf():
     """Does Tesseract have textonly_pdf capability?
-    
-    Available in 3.05.01, and v4.00.00alpha since January 2017. Best to 
+
+    Available in 3.05.01, and v4.00.00alpha since January 2017. Best to
     parse the parameter list
     """
     args_tess = [
         'tesseract',
-        '--print-parameters'
+        '--print-parameters',
+        'pdf'
     ]
     params = ''
     try:
@@ -95,22 +92,33 @@ def psm():
 
 @lru_cache(maxsize=1)
 def languages():
+    def lang_error(output):
+        msg = dedent("""Tesseract failed to report available languages.
+        Output from Tesseract:
+        -----------
+        """)
+        msg += output
+        print(msg, file=sys.stderr)
+
     args_tess = [
         'tesseract',
         '--list-langs'
     ]
     try:
-        langs = check_output(
-                args_tess, universal_newlines=True, stderr=STDOUT)
+        proc = run(
+            args_tess, universal_newlines=True, stdout=PIPE, stderr=STDOUT,
+            check=True
+        )
+        output = proc.stdout
     except CalledProcessError as e:
-        msg = dedent("""Tesseract failed to report available languages.
-        Output from Tesseract:
-        -----------
-        """)
-        msg += e.output
-        print(msg, file=sys.stderr)
+        lang_error(e.output)
         raise MissingDependencyError from e
-    return set(lang.strip() for lang in langs.splitlines()[1:])
+
+    header, *rest = output.splitlines()
+    if not header.startswith('List of available languages'):
+        lang_error(output)
+        raise MissingDependencyError
+    return set(lang.strip() for lang in rest)
 
 
 def tess_base_args(langs, engine_mode):
@@ -124,11 +132,10 @@ def tess_base_args(langs, engine_mode):
     return args
 
 
-def get_orientation(input_file, language: list, engine_mode, timeout: float,
-                    log):
+def get_orientation(input_file, engine_mode, timeout: float, log):
     args_tesseract = tess_base_args(['osd'], engine_mode) + [
         psm(), '0',
-        input_file,
+        fspath(input_file),
         'stdout'
     ]
 
@@ -159,7 +166,7 @@ def get_orientation(input_file, language: list, engine_mode, timeout: float,
             assert 'Rotate' not in osd
             angle = -angle % 360
         else:
-            # Tesseract == 3.04.01, hopefully also Tesseract > 3.04.01
+            # Tesseract >= 3.04.01
             # reports "Orientation in degrees" as a clockwise angle
             assert 'Rotate' in osd
 
@@ -190,6 +197,14 @@ def tesseract_log_output(log, stdout, input_file):
             log.warning(prefix + "lots of diacritics - possibly poor OCR")
         elif line.startswith('OSD: Weak margin'):
             log.warning(prefix + "unsure about page orientation")
+        elif 'Error in pixScanForForeground' in line:
+            pass  # Appears to be spurious/problem with nonwhite borders
+        elif 'Error in boxClipToRectangle' in line:
+            pass  # Always appears with pixScanForForeground message
+        elif 'parameter not found: ' in line.lower():
+            log.error(prefix + line.strip())
+            problem = line.split('found: ')[1]
+            raise TesseractConfigError(problem)
         elif 'error' in line.lower() or 'exception' in line.lower():
             log.error(prefix + line.strip())
         elif 'warning' in line.lower():
@@ -263,8 +278,6 @@ def generate_hocr(input_file, output_files, language: list, engine_mode,
         _generate_null_hocr(output_hocr, output_sidecar, input_file)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_file)
-        if b'read_params_file: parameter not found' in e.output:
-            raise TesseractConfigError() from e
         if b'Image too large' in e.output:
             _generate_null_hocr(output_hocr, output_sidecar, input_file)
             return
@@ -282,30 +295,19 @@ def use_skip_page(text_only, skip_pdf, output_pdf, output_text):
     with open(output_text, 'w') as f:
         f.write('[skipped page]')
 
-    if not text_only:
+    if skip_pdf and not text_only:
+        # Substitute a "skipped page"
         with suppress(FileNotFoundError):
             os.remove(output_pdf)  # In case it was partially created
         os.symlink(skip_pdf, output_pdf)
         return
 
-    # For text only we must create a blank page with dimensions identical
-    # to the skip page because this is equivalent to a page with no text
-
-    pdf_in = pypdf.PdfFileReader(skip_pdf)
-    page0 = pdf_in.pages[0]
-
+    # Or normally, just write a 0 byte file to the output to indicate a skip
     with open(output_pdf, 'wb') as out:
-        pdf_out = pypdf.PdfFileWriter()
-        w, h = page0.mediaBox.getWidth(), page0.mediaBox.getHeight()
-        # If skip page has a /Rotate key, replicate the rotation
-        rotation = int(page0.get('/Rotate', 0))
-        if rotation % 180 == 90:
-            w, h = h, w
-        pdf_out.addBlankPage(w, h)
-        pdf_out.write(out)
+        out.write(b'')
 
 
-def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
+def generate_pdf(*, input_image, skip_pdf=None, output_pdf, output_text,
                  language: list, engine_mode, text_only: bool,
                  tessconfig: list, timeout: float, pagesegmode: int,
                  user_words, user_patterns, log):
@@ -328,7 +330,7 @@ def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
     if pagesegmode is not None:
         args_tesseract.extend([psm(), str(pagesegmode)])
 
-    if text_only:
+    if text_only and has_textonly_pdf():
         args_tesseract.extend(['-c', 'textonly_pdf=1'])
 
     if user_words:
@@ -361,9 +363,6 @@ def generate_pdf(*, input_image, skip_pdf, output_pdf, output_text,
         use_skip_page(text_only, skip_pdf, output_pdf, output_text)
     except CalledProcessError as e:
         tesseract_log_output(log, e.output, input_image)
-        if b'read_params_file: parameter not found' in e.output:
-            raise TesseractConfigError() from e
-
         if b'Image too large' in e.output:
             use_skip_page(text_only, skip_pdf, output_pdf, output_text)
             return

src/ocrmypdf/helpers.py

@@ -17,7 +17,7 @@
 
 from functools import partial
 from collections.abc import Iterable
-from contextlib import suppress, contextmanager
+from contextlib import suppress
 from pathlib import Path
 import sys
 import os
@@ -29,7 +29,8 @@ def re_symlink(input_file, soft_link_name, log=None):
     """
     Helper function: relinks soft symbolic link if necessary
     """
-
+    input_file = fspath(input_file)  # For Py3.5
+    soft_link_name = fspath(soft_link_name)
     if log is None:
         prdebug = partial(print, file=sys.stderr)
     else:
@@ -70,7 +71,7 @@ def is_iterable_notstr(thing):
 
 
 def page_number(input_file):
-    "Get one-based page number implied by filename (000002.pdf -> 2)"
+    """Get one-based page number implied by filename (000002.pdf -> 2)"""
     return int(os.path.basename(fspath(input_file))[0:6])
 
 
@@ -103,18 +104,18 @@ def is_file_writable(test_file):
 
     if p.is_symlink():
         # Python 3.5 does not accept parameters for Path.resolve() and behaves
-        # as if strict=True (throws an exception on failure). Python 3.6 
+        # as if strict=True (throws an exception on failure). Python 3.6
         # defaults to strict=False. This implements strict=False like behavior
         # for Python 3.5.
         if sys.version_info[0:2] <= (3, 5):
-            p = Path(os.path.realpath(str(p)))
+            p = Path(os.path.realpath(fspath(p)))
         else:
             p = p.resolve(strict=False)
 
     # p.is_file() throws an exception in some cases
     if p.exists() and p.is_file():
         return os.access(
-            str(p), os.W_OK,
+            fspath(p), os.W_OK,
             effective_ids=(os.access in os.supports_effective_ids))
     else:
         try:
@@ -129,17 +130,9 @@ def is_file_writable(test_file):
 
 
 if sys.version_info[0:2] <= (3, 5):
-    def universal_open(p, *args, **kwargs):
-        "Work around Python 3.5's inability to open(pathlib.Path())"
-        try:
-            return p.open(*args, **kwargs)
-        except AttributeError:
-            return open(p, *args, **kwargs)
-
-
     def fspath(path):
+        """https://www.python.org/dev/peps/pep-0519/#os"""
         import pathlib
-        '''https://www.python.org/dev/peps/pep-0519/#os'''
         if isinstance(path, (str, bytes)):
             return path
 
@@ -166,5 +159,12 @@ if sys.version_info[0:2] <= (3, 5):
             + path_type.__name__)
 
 else:
-    universal_open = open
     fspath = os.fspath
+
+
+def flatten_groups(groups):
+    for obj in groups:
+        if is_iterable_notstr(obj):
+            yield from obj
+        else:
+            yield obj

src/ocrmypdf/hocrtransform.py

@@ -16,10 +16,10 @@
 # distribute, sublicense, and/or sell copies of the Software, and to
 # permit persons to whom the Software is furnished to do so, subject to
 # the following conditions:
-# 
+#
 # The above copyright notice and this permission notice shall be included
 # in all copies or substantial portions of the Software.
-# 
+#
 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 # OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 # MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
@@ -31,7 +31,6 @@
 from reportlab.pdfgen.canvas import Canvas
 from reportlab.lib.units import inch
 from xml.etree import ElementTree
-from PIL import Image
 from collections import namedtuple
 from math import atan, sin, cos
 import re
@@ -55,7 +54,7 @@ class HocrTransform():
 
     box_pattern = re.compile(r'bbox((\s+\d+){4})')
     baseline_pattern = re.compile(r'''
-        baseline \s+ 
+        baseline \s+
         ([\-\+]?\d*\.?\d*) \s+  # +/- decimal float
         ([\-\+]?\d+)            # +/- int''', re.VERBOSE)
     ligatures = str.maketrans({
@@ -139,7 +138,7 @@ class HocrTransform():
             matches = cls.baseline_pattern.search(element.attrib['title'])
             if matches:
                 return float(matches.group(1)), int(matches.group(2))
-        return (0, 0) 
+        return (0, 0)
 
     def pt_from_pixel(self, pxl):
         """
@@ -193,12 +192,12 @@ class HocrTransform():
                 pdf.rect(
                     pt.x1, self.height - pt.y2, pt.x2 - pt.x1, pt.y2 - pt.y1,
                     fill=1)
-                    
+
         found_lines = False
         for line in self.hocr.findall(
                 ".//%sspan[@class='%s']" % (self.xmlns, "ocr_line")):
             found_lines = True
-            self._do_line(pdf, line, "ocrx_word", fontname, invisibleText, 
+            self._do_line(pdf, line, "ocrx_word", fontname, invisibleText,
                           interwordSpaces, showBoundingboxes)
 
         if not found_lines:
@@ -221,7 +220,7 @@ class HocrTransform():
         return x * poly[0] + poly[1]
 
 
-    def _do_line(self, pdf, line, elemclass, fontname, invisibleText, 
+    def _do_line(self, pdf, line, elemclass, fontname, invisibleText,
                  interwordSpaces, showBoundingboxes):
         pxl_line_coords = self.element_coordinates(line)
         line_box = self.pt_from_pixel(pxl_line_coords)
@@ -255,10 +254,10 @@ class HocrTransform():
             pdf.setLineWidth(0.5)
             # negate slope because it is defined as a rise/run in pixel
             # coordinates and page coordinates have the y axis flipped
-            pdf.line(line_box.x1, 
+            pdf.line(line_box.x1,
                      baseline_y2,
                      line_box.x2,
-                     self.polyval((-slope, baseline_y2), 
+                     self.polyval((-slope, baseline_y2),
                                   line_box.x2 - line_box.x1))
             # light green for bounding box of word/line
             pdf.setDash(6, 3)
@@ -282,16 +281,16 @@ class HocrTransform():
             box = self.pt_from_pixel(pxl_coords)
             if interwordSpaces:
                 # if  `--interword-spaces` is true, append a space
-                # to the end of each text element to allow simpler PDF viewers 
-                # such as PDF.js to better recognize words in search and copy 
+                # to the end of each text element to allow simpler PDF viewers
+                # such as PDF.js to better recognize words in search and copy
                 # and paste. Do not remove space from last word in line, even
                 # though it would look better, because it will interfere with
                 # naive text extraction. \n does not work either.
                 elemtxt += ' '
                 box = Rect._make((
-                    box.x1, 
+                    box.x1,
                     line_box.y1,
-                    box.x2 + pdf.stringWidth(' ', fontname, line_height), 
+                    box.x2 + pdf.stringWidth(' ', fontname, line_height),
                     line_box.y2))
             box_width = box.x2 - box.x1
             font_width = pdf.stringWidth(elemtxt, fontname, fontsize)
@@ -312,14 +311,14 @@ class HocrTransform():
             # content stream while this issues a "offset" (Td) command.
             # .moveCursor() is relative to start of the text line, where the
             # "text line" means whatever reportlab defines it as. Do not use
-            # use .getCursor(), since moveCursor() rather unintuitively plans 
+            # use .getCursor(), since moveCursor() rather unintuitively plans
             # its moves relative to .getStartOfLine().
             # For skewed lines, in the text transform we set up a rotated
-            # coordinate system, so we don't have to account for the 
+            # coordinate system, so we don't have to account for the
             # incremental offset. Surprisingly most PDF viewers can handle this.
             cursor = text.getStartOfLine()
             dx = box.x1 - cursor[0]
-            dy = baseline_y2 - cursor[1]            
+            dy = baseline_y2 - cursor[1]
             text.moveCursor(dx, dy)
 
             # If reportlab tells us this word is 0 units wide, our best seems

src/ocrmypdf/leptonica.py

@@ -20,41 +20,53 @@
 #
 # Python FFI wrapper for Leptonica library
 
-import argparse
-import sys
-import os
-import logging
-from tempfile import TemporaryFile
+from collections.abc import Sequence
+from contextlib import suppress
 from ctypes.util import find_library
-from .lib._leptonica import ffi
 from functools import lru_cache
-from enum import Enum
+from io import BytesIO
+from tempfile import TemporaryFile
+import argparse
+import logging
+import os
+import sys
+import warnings
+
+from .lib._leptonica import ffi
 from .helpers import fspath
 
-lept = ffi.dlopen(find_library('lept'))
+# pylint: disable=protected-access
 
 logger = logging.getLogger(__name__)
 
+lept = ffi.dlopen(find_library('lept'))
+lept.setMsgSeverity(lept.L_SEVERITY_WARNING)
+
 
 def stderr(*objs):
-    """Python 2/3 compatible print to stderr.
-    """
+    """Shorthand print to stderr."""
     print("leptonica.py:", *objs, file=sys.stderr)
 
 
-class LeptonicaErrorTrap(object):
-    """Context manager to trap errors reported by Leptonica.
+class _LeptonicaErrorTrap:
+    """
+    Context manager to trap errors reported by Leptonica.
 
-    Leptonica's error return codes are unreliable to the point of being
-    almost useless.  It does, however, write errors to stderr provided that is
-    not disabled at its compile time.  Fortunately this is done using error
-    macros so it is very self-consistent.
+    Leptonica's error return codes don't provide much informatino about what
+    went wrong. Leptonica does, however, write more detailed errors to stderr
+    (provided this is not disabled at compile time). The Leptonica source
+    code is very consistent in its use of macros to generate errors.
 
     This context manager redirects stderr to a temporary file which is then
     read and parsed for error messages.  As a side benefit, debug messages
     from Leptonica are also suppressed.
 
     """
+
+    def __init__(self):
+        self.tmpfile = None
+        self.copy_of_stderr = -1
+
     def __enter__(self):
         from io import UnsupportedOperation
         self.tmpfile = TemporaryFile()
@@ -95,6 +107,8 @@ class LeptonicaErrorTrap(object):
                 raise FileNotFoundError()
             if 'pixWrite: stream not opened' in leptonica_output:
                 raise LeptonicaIOError()
+            if 'index not valid' in leptonica_output:
+                raise IndexError()
             raise LeptonicaError(leptonica_output)
 
         return False
@@ -108,8 +122,45 @@ class LeptonicaIOError(LeptonicaError):
     pass
 
 
-class Pix:
-    """Wrapper around leptonica's PIX object.
+class LeptonicaObject:
+    """General wrapper for Leptonica objects
+
+    When Leptonica returns an object, we bundled it in a wrapper class, which
+    manages its memory. The wrapper class assumes that it will be calling some
+    sort of lept.thingDestroy() function when the instance is deleted. Most
+    Leptonica objects are reference counted, and destroy decrements the
+    refcount.
+
+    Most of the time, when Leptonica returns something, we wrap and it the job
+    is done. When wrapping objects that came from a Leptonica container, like
+    a PIXA returning PIX, the subclass must clone the object before passing it
+    here, to maintain the reference count.
+
+    CFFI ensures that the destroy function is called at garbage collection time
+    so we do not need to mess with __del__.
+    """
+
+    cdata_destroy = lambda cdata: None
+    LEPTONICA_TYPENAME = ''
+
+    def __init__(self, cdata):
+        if not cdata:
+            raise ValueError('Tried to wrap a NULL ' + self.LEPTONICA_TYPENAME)
+        self._cdata = ffi.gc(cdata, self._destroy)
+
+    @classmethod
+    def _destroy(cls, cdata):
+        """Destroy some cdata"""
+        # Leptonica API uses double-pointers for its destroy APIs to prevent
+        # dangling pointers. This means we need to put our single pointer,
+        # cdata, in a temporary CDATA**.
+        pp = ffi.new('{} **'.format(cls.LEPTONICA_TYPENAME), cdata)
+        cls.cdata_destroy(pp)
+
+
+class Pix(LeptonicaObject):
+    """
+    Wrapper around leptonica's PIX object.
 
     Leptonica uses referencing counting on PIX objects. Also, many Leptonica
     functions return the original object with an increased reference count
@@ -125,19 +176,18 @@ class Pix:
     This class treats Pix objects as immutable.  All methods return new
     modified objects.  This allows convenient chaining:
 
-    >>>   Pix.read('filename.jpg').scale((0.5, 0.5)).deskew().show()
-
+    >>>   Pix.open('filename.jpg').scale((0.5, 0.5)).deskew().show()
     """
 
-    def __init__(self, pix):
-        self._pix = ffi.gc(pix, Pix._pix_destroy)
+    LEPTONICA_TYPENAME = "PIX"
+    cdata_destroy = lept.pixDestroy
 
     def __repr__(self):
-        if self._pix:
+        if self._cdata:
             s = "<leptonica.Pix image size={0}x{1} depth={2}{4} at 0x{3:x}>"
-            return s.format(self._pix.w, self._pix.h, self._pix.d,
-                            int(ffi.cast('intptr_t', self._pix)),
-                            '(colormapped)' if self._pix.colormap else '')
+            return s.format(self._cdata.w, self._cdata.h, self._cdata.d,
+                            int(ffi.cast('intptr_t', self._cdata)),
+                            '(colormapped)' if self._cdata.colormap else '')
         else:
             return "<leptonica.Pix image NULL>"
 
@@ -150,7 +200,7 @@ class Pix:
         data = ffi.new('l_uint8 **')
         size = ffi.new('size_t *')
 
-        err = lept.pixWriteMemPng(data, size, self._pix, 0)
+        err = lept.pixWriteMemPng(data, size, self._cdata, 0)
         if err != 0:
             raise LeptonicaIOError("pixWriteMemPng")
 
@@ -161,7 +211,7 @@ class Pix:
         data = ffi.new('l_uint32 **')
         size = ffi.new('size_t *')
 
-        err = lept.pixSerializeToMemory(self._pix, data, size)
+        err = lept.pixSerializeToMemory(self._cdata, data, size)
         if err != 0:
             raise LeptonicaIOError("pixSerializeToMemory")
 
@@ -178,32 +228,38 @@ class Pix:
         cdata_bytes = ffi.new('char[]', state['data'])
         cdata_uint32 = ffi.cast('l_uint32 *', cdata_bytes)
 
-        pix = lept.pixDeserializeFromMemory(
-            cdata_uint32, len(state['data']))
+        pix = lept.pixDeserializeFromMemory(cdata_uint32, len(state['data']))
         Pix.__init__(self, pix)
 
     def __eq__(self, other):
-        return self.__getstate__() == other.__getstate__()
+        if not isinstance(other, Pix):
+            return NotImplemented
+        same = ffi.new('l_int32 *', 0)
+        with _LeptonicaErrorTrap():
+            err = lept.pixEqual(self._cdata, other._cdata, same)
+            if err:
+                raise TypeError()
+        return bool(same[0])
 
     @property
     def width(self):
-        return self._pix.w
+        return self._cdata.w
 
     @property
     def height(self):
-        return self._pix.h
+        return self._cdata.h
 
     @property
     def depth(self):
-        return self._pix.d
+        return self._cdata.d
 
     @property
     def size(self):
-        return (self._pix.w, self._pix.h)
+        return (self._cdata.w, self._cdata.h)
 
     @property
     def info(self):
-        return {'dpi': (self._pix.xres, self._pix.yres)}
+        return {'dpi': (self._cdata.xres, self._cdata.yres)}
 
     @property
     def mode(self):
@@ -212,20 +268,25 @@ class Pix:
             return '1'
         elif self.depth >= 16:
             return 'RGB'
-        elif not self._pix.colormap:
+        elif not self._cdata.colormap:
             return 'L'
         else:
             return 'P'
 
     @classmethod
     def read(cls, path):
+        warnings.warn('Use Pix.open() instead', DeprecationWarning)
+        return cls.open(path)
+
+    @classmethod
+    def open(cls, path):
         """Load an image file into a PIX object.
 
         Leptonica can load TIFF, PNM (PBM, PGM, PPM), PNG, and JPEG.  If
         loading fails then the object will wrap a C null pointer.
         """
         filename = fspath(path)
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             return cls(lept.pixRead(os.fsencode(filename)))
 
     def write_implied_format(
@@ -236,13 +297,24 @@ class Pix:
         jpeg_progressive -- (iff JPEG; 0 for baseline seq., 1 for progressive)
         """
         filename = fspath(path)
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             lept.pixWriteImpliedFormat(
                 os.fsencode(filename),
-                self._pix, jpeg_quality, jpeg_progressive)
+                self._cdata, jpeg_quality, jpeg_progressive)
+
+    @classmethod
+    def frompil(self, pillow_image):
+        """Create a copy of a PIL.Image from this Pix"""
+        bio = BytesIO()
+        pillow_image.save(bio, format='png', compress_level=1)
+        py_buffer = bio.getbuffer()
+        c_buffer = ffi.from_buffer(py_buffer)
+        with _LeptonicaErrorTrap():
+            pix = Pix(lept.pixReadMem(c_buffer, len(c_buffer)))
+            return pix
 
     def topil(self):
-        "Returns a PIL.Image version of this Pix"
+        """Returns a PIL.Image version of this Pix"""
         from PIL import Image
 
         # Leptonica manages data in words, so it implicitly does an endian
@@ -255,17 +327,17 @@ class Pix:
                 raw_mode = 'ABGR'
             elif self.mode == '1':
                 raw_mode = '1;I'
-                pix = Pix(lept.pixEndianByteSwapNew(pix._pix))
+                pix = Pix(lept.pixEndianByteSwapNew(pix._cdata))
             else:
                 raw_mode = self.mode
-                pix = Pix(lept.pixEndianByteSwapNew(pix._pix))
+                pix = Pix(lept.pixEndianByteSwapNew(pix._cdata))
         else:
             raw_mode = self.mode  # no endian swap needed
 
-        size = (pix._pix.w, pix._pix.h)
-        bytecount = pix._pix.wpl * 4 * pix._pix.h
-        buf = ffi.buffer(pix._pix.data, bytecount)
-        stride = pix._pix.wpl * 4
+        size = (pix._cdata.w, pix._cdata.h)
+        bytecount = pix._cdata.wpl * 4 * pix._cdata.h
+        buf = ffi.buffer(pix._cdata.data, bytecount)
+        stride = pix._cdata.wpl * 4
 
         im = Image.frombytes(self.mode, size, buf, 'raw', raw_mode, stride)
 
@@ -283,65 +355,65 @@ class Pix:
         reduction_factor -- amount to downsample (0 for default) when searching
             for skew angle
         """
-        with LeptonicaErrorTrap():
-            return Pix(lept.pixDeskew(self._pix, reduction_factor))
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixDeskew(self._cdata, reduction_factor))
 
     def scale(self, scale_xy):
         "Returns the pix object rescaled according to the proportions given."
-        with LeptonicaErrorTrap():
-            return Pix(lept.pixScale(self._pix, scale_xy[0], scale_xy[1]))
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixScale(self._cdata, scale_xy[0], scale_xy[1]))
 
     def rotate180(self):
-        with LeptonicaErrorTrap():
-            return Pix(lept.pixRotate180(ffi.NULL, self._pix))
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixRotate180(ffi.NULL, self._cdata))
 
     def rotate_orth(self, quads):
         "Orthographic rotation, quads: 0-3, number of clockwise rotations"
-        with LeptonicaErrorTrap():
-            return Pix(lept.pixRotateOrth(self._pix, quads))
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixRotateOrth(self._cdata, quads))
 
     def find_skew(self):
         """Returns a tuple (deskew angle in degrees, confidence value).
 
         Returns (None, None) if no angle is available.
         """
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             angle = ffi.new('float *', 0.0)
             confidence = ffi.new('float *', 0.0)
-            result = lept.pixFindSkew(self._pix, angle, confidence)
+            result = lept.pixFindSkew(self._cdata, angle, confidence)
             if result == 0:
                 return (angle[0], confidence[0])
             else:
                 return (None, None)
 
     def convert_rgb_to_luminance(self):
-        with LeptonicaErrorTrap():
-            gray_pix = lept.pixConvertRGBToLuminance(self._pix)
+        with _LeptonicaErrorTrap():
+            gray_pix = lept.pixConvertRGBToLuminance(self._cdata)
             if gray_pix:
                 return Pix(gray_pix)
             return None
 
     def remove_colormap(self, removal_type):
-        """Remove a palette (colormap); if no colormap, returns a copy of this 
+        """Remove a palette (colormap); if no colormap, returns a copy of this
         image
 
             removal_type - any of lept.REMOVE_CMAP_*
 
         """
-        return self
-        # with LeptonicaErrorTrap():
-        #     return Pix(lept.pixRemoveColormapGeneral(
-        #             self._pix, removal_type, lept.L_COPY))
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixRemoveColormapGeneral(
+                    self._cdata, removal_type, lept.L_COPY))
 
     def otsu_adaptive_threshold(
             self, tile_size=(300, 300), kernel_size=(4, 4), scorefract=0.1):
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             sx, sy = tile_size
             smoothx, smoothy = kernel_size
             p_pix = ffi.new('PIX **')
 
+            pix = Pix(lept.pixConvertTo8(self._cdata, 0))
             result = lept.pixOtsuAdaptiveThreshold(
-                self._pix,
+                pix._cdata,
                 sx, sy,
                 smoothx, smoothy,
                 scorefract,
@@ -355,33 +427,53 @@ class Pix:
     def otsu_threshold_on_background_norm(
             self, mask=None, tile_size=(10, 15), thresh=100, mincount=50,
             bgval=255, kernel_size=(2, 2), scorefract=0.1):
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             sx, sy = tile_size
             smoothx, smoothy = kernel_size
-            if mask is None:
-                mask = ffi.NULL
+            mask = ffi.NULL
             if isinstance(mask, Pix):
-                mask = mask._pix
+                mask = mask._cdata
 
+            pix = Pix(lept.pixConvertTo8(self._cdata, 0))
             thresh_pix = lept.pixOtsuThreshOnBackgroundNorm(
-                self._pix,
+                pix._cdata,
                 mask,
                 sx, sy,
                 thresh, mincount, bgval,
                 smoothx, smoothy,
                 scorefract,
                 ffi.NULL
-                )
-            if thresh_pix == ffi.NULL:
-                return None
+            )
+            return Pix(thresh_pix)
+
+    def masked_threshold_on_background_norm(
+            self, mask=None, tile_size=(10, 15), thresh=100, mincount=50,
+            kernel_size=(2, 2), scorefract=0.1):
+        with _LeptonicaErrorTrap():
+            sx, sy = tile_size
+            smoothx, smoothy = kernel_size
+            mask = ffi.NULL
+            if isinstance(mask, Pix):
+                mask = mask._cdata
+
+            pix = Pix(lept.pixConvertTo8(self._cdata, 0))
+            thresh_pix = lept.pixMaskedThreshOnBackgroundNorm(
+                pix._cdata,
+                mask,
+                sx, sy,
+                thresh, mincount,
+                smoothx, smoothy,
+                scorefract,
+                ffi.NULL
+            )
             return Pix(thresh_pix)
 
     def crop_to_foreground(
             self, threshold=128, mindist=70, erasedist=30, pagenum=0,
             showmorph=0, display=0, pdfdir=ffi.NULL):
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             cropbox = Box(lept.pixFindPageForeground(
-                self._pix,
+                self._cdata,
                 threshold,
                 mindist,
                 erasedist,
@@ -390,20 +482,18 @@ class Pix:
                 display,
                 pdfdir))
 
-            print(repr(cropbox))
-
             cropped_pix = lept.pixClipRectangle(
-                self._pix,
-                cropbox._box,
+                self._cdata,
+                cropbox._cdata,
                 ffi.NULL)
 
             return Pix(cropped_pix)
 
     def clean_background_to_white(
             self, mask=None, grayscale=None, gamma=1.0, black=0, white=255):
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             return Pix(lept.pixCleanBackgroundToWhite(
-                self._pix,
+                self._cdata,
                 mask or ffi.NULL,
                 grayscale or ffi.NULL,
                 gamma,
@@ -411,10 +501,10 @@ class Pix:
                 white))
 
     def gamma_trc(self, gamma=1.0, minval=0, maxval=255):
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             return Pix(lept.pixGammaTRC(
                 ffi.NULL,
-                self._pix,
+                self._cdata,
                 gamma,
                 minval,
                 maxval
@@ -425,9 +515,9 @@ class Pix:
             min_count=40, bg_val=200, smooth_kernel=(2, 1)):
         # Background norm doesn't work on color mapped Pix, so remove colormap
         target_pix = self.remove_colormap(lept.REMOVE_CMAP_BASED_ON_SRC)
-        with LeptonicaErrorTrap():
+        with _LeptonicaErrorTrap():
             return Pix(lept.pixBackgroundNorm(
-                target_pix._pix,
+                target_pix._cdata,
                 mask or ffi.NULL,
                 grayscale or ffi.NULL,
                 tile_size[0],
@@ -437,7 +527,7 @@ class Pix:
                 bg_val,
                 smooth_kernel[0],
                 smooth_kernel[1]
-                ))
+            ))
 
     @staticmethod
     @lru_cache(maxsize=1)
@@ -451,73 +541,236 @@ class Pix:
             # implementation of pixCorrelationBinary that overflows on larger
             # images.  Ubuntu 14.04/trusty has 1.70. Ubuntu PPA
             # ppa:alex-p/tesseract-ocr has leptonlib 1.75.
-            pix1_count = ffi.new('l_int32 *')
-            pix2_count = ffi.new('l_int32 *')
-            pixn_count = ffi.new('l_int32 *')
-            tab8 = Pix.make_pixel_sum_tab8()
+            raise LeptonicaError("Leptonica version is too old")
 
-            lept.pixCountPixels(pix1._pix, pix1_count, tab8)
-            lept.pixCountPixels(pix2._pix, pix2_count, tab8)
-            pixn = Pix(lept.pixAnd(ffi.NULL, pix1._pix, pix2._pix))
-            lept.pixCountPixels(pixn._pix, pixn_count, tab8)
+        correlation = ffi.new('float *', 0.0)
+        result = lept.pixCorrelationBinary(pix1._cdata, pix2._cdata,
+                                           correlation)
+        if result != 0:
+            raise LeptonicaError("Correlation failed")
+        return correlation[0]
 
-            # Python converts these int32s to larger units as needed
-            # to avoid overflow. Overflow happens easily here.
-            correlation = (
-                    (pixn_count[0] * pixn_count[0]) /
-                    (pix1_count[0] * pix2_count[0])
-                    )
-            return correlation
+    def generate_pdf_ci_data(self, type_, quality):
+        "Convert to PDF data, with transcoding"
+        p_compdata = ffi.new('L_COMP_DATA **')
+        result = lept.pixGenerateCIData(self._cdata, type_, quality, 0,
+                                        p_compdata)
+        if result != 0:
+            raise LeptonicaError("Generate PDF data failed")
+        return CompressedData(p_compdata[0])
+
+    def invert(self):
+        return Pix(lept.pixInvert(ffi.NULL, self._cdata))
+
+    def locate_barcodes(self):
+        try:
+            with _LeptonicaErrorTrap():
+                pix = Pix(lept.pixConvertTo8(self._cdata, 0))
+                pixa_candidates = PixArray(lept.pixExtractBarcodes(pix._cdata, 0))
+                sarray = StringArray(lept.pixReadBarcodes(
+                    pixa_candidates._cdata,
+                    lept.L_BF_ANY,
+                    lept.L_USE_WIDTHS,
+                    ffi.NULL,
+                    0
+                ))
+        except (LeptonicaError, ValueError) as e:
+            return
+
+        for n, s in enumerate(sarray):
+            decoded = s.decode()
+            if decoded.strip() == '':
+                continue
+            box = pixa_candidates.get_box(n)
+            left, top = box.x, box.y
+            right, bottom = box.x + box.w, box.y + box.h
+            yield (decoded, (left, top, right, bottom))
+
+    def despeckle(self, size):
+        if size == 2:
+            speckle2 = """
+                oooo
+                oC o
+                o  o
+                oooo
+                """
+            sel1 = Sel.from_selstr(speckle2, 'speckle2')
+            sel2 = Sel.create_brick(2, 2, 0, 0, lept.SEL_HIT)
+        elif size == 3:
+            speckle3 = """
+                ooooo
+                oC  o
+                o   o
+                o   o
+                ooooo
+                """
+            sel1 = Sel.from_selstr(speckle3, 'speckle3')
+            sel2 = Sel.create_brick(3, 3, 0, 0, lept.SEL_HIT)
         else:
-            correlation = ffi.new('float *', 0.0)
-            result = lept.pixCorrelationBinary(pix1._pix, pix2._pix,
-                                               correlation)
-            if result != 0:
-                raise LeptonicaError("Correlation failed")
-            return correlation[0]
+            raise ValueError(size)
 
-    @staticmethod
-    def _pix_destroy(pix):
-        p_pix = ffi.new('PIX **', pix)
-        lept.pixDestroy(p_pix)
-        # print('pix destroy ' + repr(pix))
+        pixhmt = Pix(lept.pixHMT(ffi.NULL, self._cdata, sel1._cdata))
+        pixdilated = Pix(lept.pixDilate(ffi.NULL, pixhmt._cdata, sel2._cdata))
+
+        pixsub = Pix(lept.pixSubtract(ffi.NULL, self._cdata, pixdilated._cdata))
+        return pixsub
 
 
-class Box:
-    """Wrapper around Leptonica's BOX objects.
+class CompressedData(LeptonicaObject):
+    """Wrapper for L_COMP_DATA - abstract compressed image data"""
 
-    See class Pix for notes about reference counting.
+    LEPTONICA_TYPENAME = 'L_COMP_DATA'
+    cdata_destroy = lept.l_CIDataDestroy
+
+    @classmethod
+    def open(cls, path, jpeg_quality=75):
+        "Open compressed data, without transcoding"
+        filename = fspath(path)
+
+        p_compdata = ffi.new('L_COMP_DATA **')
+        result = lept.l_generateCIDataForPdf(
+            os.fsencode(filename), ffi.NULL, jpeg_quality, p_compdata)
+        if result != 0:
+            raise LeptonicaError("CompressedData.open")
+        return CompressedData(p_compdata[0])
+
+    def __len__(self):
+        return self._cdata.nbytescomp
+
+    def read(self):
+        buf = ffi.buffer(self._cdata.datacomp, self._cdata.nbytescomp)
+        return bytes(buf)
+
+    def __getattr__(self, name):
+        if hasattr(self._cdata, name):
+            return getattr(self._cdata, name)
+        raise AttributeError(name)
+
+    def get_palette_pdf_string(self):
+        "Returns palette pre-formatted for use in PDF"
+        buflen = len('< ') + len(' rrggbb') * self._cdata.ncolors + len('>')
+        buf = ffi.buffer(self._cdata.cmapdatahex, buflen)
+        return bytes(buf)
+
+
+class PixArray(LeptonicaObject, Sequence):
+    """Wrapper around PIXA (array of PIX)"""
+
+    LEPTONICA_TYPENAME = 'PIXA'
+    cdata_destroy = lept.pixaDestroy
+
+    def __len__(self):
+        return self._cdata[0].n
+
+    def __getitem__(self, n):
+        with _LeptonicaErrorTrap():
+            return Pix(lept.pixaGetPix(self._cdata, n, lept.L_CLONE))
+
+    def get_box(self, n):
+        with _LeptonicaErrorTrap():
+            return Box(lept.pixaGetBox(self._cdata, n, lept.L_CLONE))
+
+
+class Box(LeptonicaObject):
+    """Wrapper around Leptonica's BOX objects (a pixel rectangle)
+
+    Uses x, y, w, h coordinates.
     """
 
-    def __init__(self, box):
-        self._box = ffi.gc(box, Box._box_destroy)
+    LEPTONICA_TYPENAME = 'BOX'
+    cdata_destroy = lept.boxDestroy
 
     def __repr__(self):
-        if self._box:
+        if self._cdata:
             return '<leptonica.Box x={0} y={1} w={2} h={3}>'.format(
                 self.x, self.y, self.w, self.h)
         return '<leptonica.Box NULL>'
 
     @property
     def x(self):
-        return self._box.x
+        return self._cdata.x
 
     @property
     def y(self):
-        return self._box.y
+        return self._cdata.y
 
     @property
     def w(self):
-        return self._box.w
+        return self._cdata.w
 
     @property
     def h(self):
-        return self._box.h
+        return self._cdata.h
 
-    @staticmethod
-    def _box_destroy(box):
-        p_box = ffi.new('BOX **', box)
-        lept.boxDestroy(p_box)
+
+class BoxArray(LeptonicaObject, Sequence):
+    """Wrapper around Leptonica's BOXA (Array of BOX) objects."""
+
+    LEPTONICA_TYPENAME = 'BOXA'
+    cdata_destroy = lept.boxaDestroy
+
+    def __repr__(self):
+        if not self._cdata:
+            return '<BoxArray>'
+        boxes = (repr(box) for box in self)
+        return '<BoxArray [' + ', '.join(boxes) + ']>'
+
+    def __len__(self):
+        return self._cdata.n
+
+    def __getitem__(self, n):
+        if not isinstance(n, int):
+            raise TypeError('list indices must be integers')
+        if 0 <= n < len(self):
+            return Box(lept.boxaGetBox(self._cdata, n, lept.L_CLONE))
+        raise IndexError(n)
+
+
+class StringArray(LeptonicaObject, Sequence):
+    """Leptonica SARRAY/string array"""
+
+    LEPTONICA_TYPENAME = 'SARRAY'
+    cdata_destroy = lept.sarrayDestroy
+
+    def __len__(self):
+        return self._cdata.n
+
+    def __getitem__(self, n):
+        if 0 <= n < len(self):
+            return ffi.string(self._cdata.array[n])
+        raise IndexError(n)
+
+
+class Sel(LeptonicaObject):
+    """Leptonica 'sel'/selection element for hit-miss transform"""
+
+    LEPTONICA_TYPENAME = 'SEL'
+    cdata_destroy = lept.selDestroy
+
+    @classmethod
+    def from_selstr(cls, selstr, name):
+        # TODO this will strip a horizontal line of don't care's
+        lines = [line.strip() for line in selstr.split('\n') if line.strip()]
+        h = len(lines)
+        w = len(lines[0])
+        lengths = set(len(line) for line in lines)
+        if len(lengths) != 1:
+            raise ValueError("All lines in selstr must be same length")
+
+        repacked = ''.join(line.strip() for line in lines)
+        buf_selstr = ffi.from_buffer(repacked.encode('ascii'))
+        buf_name = ffi.from_buffer(name.encode('ascii'))
+        sel = lept.selCreateFromString(buf_selstr, h, w, buf_name)
+        return cls(sel)
+
+    @classmethod
+    def create_brick(cls, h, w, cy, cx, type_):
+        sel = lept.selCreateBrick(h, w, cy, cx, type_)
+        return cls(sel)
+
+    def __repr__(self):
+        selstr = ffi.gc(lept.selPrintToString(self._cdata), lept.lept_free)
+        return '<Sel \n' + ffi.string(selstr).decode('ascii') + '\n>'
 
 
 @lru_cache(maxsize=1)
@@ -533,7 +786,7 @@ def get_leptonica_version():
 
 def deskew(infile, outfile, dpi):
     try:
-        pix_source = Pix.read(infile)
+        pix_source = Pix.open(infile)
     except LeptonicaIOError:
         raise LeptonicaIOError("Failed to open file: %s" % infile)
 
@@ -552,7 +805,7 @@ def deskew(infile, outfile, dpi):
 def remove_background(infile, outfile, tile_size=(40, 60), gamma=1.0,
                       black_threshold=70, white_threshold=190):
     try:
-        pix = Pix.read(infile)
+        pix = Pix.open(infile)
     except LeptonicaIOError:
         raise LeptonicaIOError("Failed to open file: %s" % infile)
 
@@ -580,4 +833,4 @@ if __name__ == '__main__':
     parser_deskew.set_defaults(func=deskew)
 
     args = parser.parse_args()
-    args.func(args)
+    args.func(args)

src/ocrmypdf/lib/__init__.py

@@ -16,13 +16,3 @@
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
 """Bindings to external libraries"""
-
-import os as _os
-
-try:
-    import fitz
-except ImportError:
-    fitz = None
-
-if _os.environ.get('_OCRMYPDF_NO_FITZ'):
-    fitz = None

src/ocrmypdf/lib/compile_leptonica.py

@@ -19,7 +19,6 @@
 from cffi import FFI
 
 ffibuilder = FFI()
-ffibuilder.set_source("ocrmypdf.lib._leptonica", None)
 ffibuilder.cdef("""
 typedef signed char             l_int8;
 typedef unsigned char           l_uint8;
@@ -32,6 +31,8 @@ typedef double                  l_float64;
 typedef long long               l_int64;
 typedef unsigned long long      l_uint64;
 
+typedef int l_ok; /*!< return type 0 if OK, 1 on error */
+
 struct Pix
 {
     l_uint32             w;           /* width in pixels                   */
@@ -61,6 +62,17 @@ struct PixColormap
 };
 typedef struct PixColormap  PIXCMAP;
 
+/*! Array of pix */
+struct Pixa
+{
+    l_int32             n;          /*!< number of Pix in ptr array        */
+    l_int32             nalloc;     /*!< number of Pix ptrs allocated      */
+    l_uint32            refcount;   /*!< reference count (1 if no clones)  */
+    struct Pix        **pix;        /*!< the array of ptrs to pix          */
+    struct Boxa        *boxa;       /*!< array of boxes                    */
+};
+typedef struct Pixa PIXA;
+
 struct Box
 {
     l_int32            x;
@@ -72,6 +84,69 @@ struct Box
 };
 typedef struct Box    BOX;
 
+/*! Array of Box */
+struct Boxa
+{
+    l_int32            n;           /*!< number of box in ptr array        */
+    l_int32            nalloc;      /*!< number of box ptrs allocated      */
+    l_uint32           refcount;    /*!< reference count (1 if no clones)  */
+    struct Box       **box;         /*!< box ptr array                     */
+};
+typedef struct Boxa BOXA;
+
+/*! String array: an array of C strings */
+struct Sarray
+{
+    l_int32          nalloc;    /*!< size of allocated ptr array         */
+    l_int32          n;         /*!< number of strings allocated         */
+    l_int32          refcount;  /*!< reference count (1 if no clones)    */
+    char           **array;     /*!< string array                        */
+};
+typedef struct Sarray SARRAY;
+
+/*! Pdf formatted encoding types */
+enum {
+    L_DEFAULT_ENCODE  = 0,  /*!< use default encoding based on image        */
+    L_JPEG_ENCODE     = 1,  /*!< use dct encoding: 8 and 32 bpp, no cmap    */
+    L_G4_ENCODE       = 2,  /*!< use ccitt g4 fax encoding: 1 bpp           */
+    L_FLATE_ENCODE    = 3,  /*!< use flate encoding: any depth, cmap ok     */
+    L_JP2K_ENCODE     = 4   /*!< use jp2k encoding: 8 and 32 bpp, no cmap   */
+};
+
+/*! Compressed image data */
+struct L_Compressed_Data
+{
+    l_int32            type;         /*!< encoding type: L_JPEG_ENCODE, etc   */
+    l_uint8           *datacomp;     /*!< gzipped raster data                 */
+    size_t             nbytescomp;   /*!< number of compressed bytes          */
+    char              *data85;       /*!< ascii85-encoded gzipped raster data */
+    size_t             nbytes85;     /*!< number of ascii85 encoded bytes     */
+    char              *cmapdata85;   /*!< ascii85-encoded uncompressed cmap   */
+    char              *cmapdatahex;  /*!< hex pdf array for the cmap          */
+    l_int32            ncolors;      /*!< number of colors in cmap            */
+    l_int32            w;            /*!< image width                         */
+    l_int32            h;            /*!< image height                        */
+    l_int32            bps;          /*!< bits/sample; typ. 1, 2, 4 or 8      */
+    l_int32            spp;          /*!< samples/pixel; typ. 1 or 3          */
+    l_int32            minisblack;   /*!< tiff g4 photometry                  */
+    l_int32            predictor;    /*!< flate data has PNG predictors       */
+    size_t             nbytes;       /*!< number of uncompressed raster bytes */
+    l_int32            res;          /*!< resolution (ppi)                    */
+};
+typedef struct L_Compressed_Data L_COMP_DATA;
+
+/*! Selection */
+struct Sel
+{
+    l_int32       sy;        /*!< sel height                               */
+    l_int32       sx;        /*!< sel width                                */
+    l_int32       cy;        /*!< y location of sel origin                 */
+    l_int32       cx;        /*!< x location of sel origin                 */
+    l_int32     **data;      /*!< {0,1,2}; data[i][j] in [row][col] order  */
+    char         *name;      /*!< used to find sel by name                 */
+};
+typedef struct Sel SEL;
+
 enum {
     REMOVE_CMAP_TO_BINARY = 0,     /*!< remove colormap for conv to 1 bpp  */
     REMOVE_CMAP_TO_GRAYSCALE = 1,  /*!< remove colormap for conv to 8 bpp  */
@@ -90,10 +165,48 @@ enum {
                       /*!< the array with clones (e.g., pix)              */
 };
 
+/*! Flags for method of extracting barcode widths */
+enum {
+    L_USE_WIDTHS = 1,     /*!< use histogram of barcode widths           */
+    L_USE_WINDOWS = 2     /*!< find best window for decoding transitions */
+};
+
+/*! Flags for barcode formats */
+enum {
+    L_BF_UNKNOWN = 0,     /*!< unknown format                            */
+    L_BF_ANY = 1,         /*!< try decoding with all known formats       */
+    L_BF_CODE128 = 2,     /*!< decode with Code128 format                */
+    L_BF_EAN8 = 3,        /*!< decode with EAN8 format                   */
+    L_BF_EAN13 = 4,       /*!< decode with EAN13 format                  */
+    L_BF_CODE2OF5 = 5,    /*!< decode with Code 2 of 5 format            */
+    L_BF_CODEI2OF5 = 6,   /*!< decode with Interleaved 2 of 5 format     */
+    L_BF_CODE39 = 7,      /*!< decode with Code39 format                 */
+    L_BF_CODE93 = 8,      /*!< decode with Code93 format                 */
+    L_BF_CODABAR = 9,     /*!< decode with Code93 format                 */
+    L_BF_UPCA = 10        /*!< decode with UPC A format                  */
+};
+
+enum {
+    L_SEVERITY_EXTERNAL = 0,   /* Get the severity from the environment   */
+    L_SEVERITY_ALL      = 1,   /* Lowest severity: print all messages     */
+    L_SEVERITY_DEBUG    = 2,   /* Print debugging and higher messages     */
+    L_SEVERITY_INFO     = 3,   /* Print informational and higher messages */
+    L_SEVERITY_WARNING  = 4,   /* Print warning and higher messages       */
+    L_SEVERITY_ERROR    = 5,   /* Print error and higher messages         */
+    L_SEVERITY_NONE     = 6    /* Highest severity: print no messages     */
+};
+
+enum {
+    SEL_DONT_CARE  = 0,
+    SEL_HIT        = 1,
+    SEL_MISS       = 2
+};
+
 """)
 
 ffibuilder.cdef("""
 PIX * pixRead ( const char *filename );
+PIX * pixReadMem ( const l_uint8 *data, size_t size );
 PIX * pixScale ( PIX *pixs, l_float32 scalex, l_float32 scaley );
 l_int32 pixFindSkew ( PIX *pixs, l_float32 *pangle, l_float32 *pconf );
 l_int32 pixWriteImpliedFormat ( const char *filename, PIX *pix, l_int32 quality, l_int32 progressive );
@@ -105,6 +218,11 @@ pixWriteMemPng(l_uint8  **pdata,
 
 void pixDestroy ( PIX **ppix );
 
+l_ok
+pixEqual(PIX      *pix1,
+         PIX      *pix2,
+         l_int32 *psame);
+
 PIX *
 pixEndianByteSwapNew(PIX  *pixs);
 
@@ -125,6 +243,8 @@ l_int32 pixSerializeToMemory ( PIX *pixs, l_uint32 **pdata, size_t *pnbytes );
 
 PIX * pixConvertRGBToLuminance(PIX *pixs);
 
+PIX * pixConvertTo8(PIX     *pixs, l_int32  cmapflag);
+
 PIX * pixRemoveColormap(PIX *pixs, l_int32  type);
 
 l_int32
@@ -150,6 +270,18 @@ pixOtsuThreshOnBackgroundNorm(PIX       *pixs,
                               l_float32  scorefract,
                               l_int32   *pthresh);
 
+PIX *
+pixMaskedThreshOnBackgroundNorm(PIX       *pixs,
+                                PIX       *pixim,
+                                l_int32    sx,
+                                l_int32    sy,
+                                l_int32    thresh,
+                                l_int32    mincount,
+                                l_int32    smoothx,
+                                l_int32    smoothy,
+                                l_float32  scorefract,
+                                l_int32   *pthresh);
+
 PIX *
 pixCleanBackgroundToWhite(PIX       *pixs,
                           PIX       *pixim,
@@ -233,13 +365,13 @@ pixGetAverageMaskedRGB(PIX        *pixs,
                        l_float32  *pgval,
                        l_float32  *pbval);
 
-PIX * 
+PIX *
 pixGlobalNormRGB(PIX * 	pixd,
                  PIX * 	pixs,
                  l_int32 	rval,
                  l_int32 	gval,
                  l_int32 	bval,
-                 l_int32 	mapval); 
+                 l_int32 	mapval);
 
 PIX *
 pixInvert(PIX * pixd,
@@ -250,13 +382,108 @@ pixRemoveColormapGeneral(PIX     *pixs,
                          l_int32  type,
                          l_int32  ifnocmap);
 
-void                 
+l_int32
+pixGenerateCIData(PIX           *pixs,
+                  l_int32        type,
+                  l_int32        quality,
+                  l_int32        ascii85,
+                  L_COMP_DATA **pcid);
+
+SARRAY *
+pixProcessBarcodes(PIX      *pixs,
+                   l_int32   format,
+                   l_int32   method,
+                   SARRAY  **psaw,
+                   l_int32 debugflag);
+
+PIX *
+pixaGetPix(PIXA    *pixa,
+           l_int32  index,
+           l_int32 accesstype);
+
+BOX*
+pixaGetBox 	(PIXA *  	pixa,
+		    l_int32  	index,
+		    l_int32  	accesstype );
+
+PIXA *
+pixExtractBarcodes(PIX     *pixs,
+                   l_int32 debugflag);
+
+BOXA *
+pixLocateBarcodes ( PIX *pixs, l_int32 thresh, PIX **ppixb, PIX **ppixm );
+
+SARRAY *
+pixReadBarcodes(PIXA     *pixa,
+                l_int32   format,
+                l_int32   method,
+                SARRAY  **psaw,
+                l_int32 debugflag);
+
+l_int32
+l_generateCIDataForPdf(const char *fname,
+                       PIX *pix,
+                       l_int32 quality,
+                       L_COMP_DATA **pcid);
+
+
+BOX *
+boxClone ( BOX *box );
+
+BOX *
+boxaGetBox ( BOXA *boxa, l_int32 index, l_int32 accessflag );
+
+SEL *
+selCreateFromString ( const char *text, l_int32 h, l_int32 w, const char *name );
+
+SEL *
+selCreateBrick ( l_int32 h, l_int32 w, l_int32 cy, l_int32 cx, l_int32 type );
+
+char *
+selPrintToString(SEL  *sel);
+
+PIX *
+pixDilate ( PIX *pixd, PIX *pixs, SEL *sel );
+
+PIX *
+pixErode ( PIX *pixd, PIX *pixs, SEL *sel );
+
+PIX *
+pixHMT ( PIX *pixd, PIX *pixs, SEL *sel );
+
+PIX *
+pixSubtract ( PIX *pixd, PIX *pixs1, PIX *pixs2 );
+
+void
 boxDestroy(BOX  **pbox);
 
 void
-lept_free(void *ptr);  
+boxaDestroy ( BOXA **pboxa );
+
+void
+pixaDestroy(PIXA **ppixa);
+
+l_ok
+pixRenderBoxa ( PIX *pix, BOXA *boxa, l_int32 width, l_int32 op );
+
+void
+l_CIDataDestroy(L_COMP_DATA **pcid);
+
+void
+sarrayDestroy(SARRAY  **psa);
+
+void
+lept_free(void *ptr);
+
+void selDestroy ( SEL **psel );
+
+l_int32
+setMsgSeverity(l_int32 newsev);
+
 """)
 
 
+ffibuilder.set_source("ocrmypdf.lib._leptonica", None)
+
 if __name__ == '__main__':
     ffibuilder.compile(verbose=True)

src/ocrmypdf/optimize.py

@@ -0,0 +1,485 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+from pathlib import Path
+import concurrent.futures
+from collections import defaultdict
+import logging
+import sys
+
+from PIL import Image
+
+import pikepdf
+
+from ._jobcontext import JobContext
+from . import leptonica
+from .helpers import re_symlink, fspath
+from .exec import pngquant, jbig2enc
+
+DEFAULT_JPEG_QUALITY = 75
+DEFAULT_PNG_QUALITY = 70
+
+
+def img_name(root, xref, ext):
+    return fspath(root / '{:08d}{}'.format(xref, ext))
+
+
+def png_name(root, xref):
+    return img_name(root, xref, '.png')
+
+
+def jpg_name(root, xref):
+    return img_name(root, xref, '.jpg')
+
+
+def tif_name(root, xref):
+    return img_name(root, xref, '.tif')
+
+
+def extract_image_filter(pike, root, log, image, xref):
+    if image.Subtype != '/Image':
+        return None
+    if image.Length < 100:
+        log.debug("Skipping small image, xref %s", xref)
+        return None
+
+    pim = pikepdf.PdfImage(image)
+
+    if len(pim.filter_decodeparms) > 1:
+        log.debug("Skipping multiply filtered, xref %s", xref)
+        return None
+    filtdp = pim.filter_decodeparms[0]
+
+    if pim.bits_per_component > 8:
+        return None  # Don't mess with wide gamut images
+
+    if filtdp[0] == '/JPXDecode':
+        return None  # Don't do JPEG2000
+
+    return pim, filtdp
+
+
+def extract_image_jbig2(*, pike, root, log, image, xref, options):
+    result = extract_image_filter(pike, root, log, image, xref)
+    if result is None:
+        return None
+    pim, filtdp = result
+
+    if pim.bits_per_component == 1 \
+            and filtdp != '/JBIG2Decode' \
+            and jbig2enc.available():
+        try:
+            imgname = Path(root / '{:08d}'.format(xref))
+            with imgname.open('wb') as f:
+                ext = pim.extract_to(stream=f)
+            imgname.rename(imgname.with_suffix(ext))
+        except pikepdf.UnsupportedImageTypeError:
+            return None
+        return xref, ext
+    return None
+
+
+def extract_image_generic(*, pike, root, log, image, xref, options):
+    result = extract_image_filter(pike, root, log, image, xref)
+    if result is None:
+        return None
+    pim, filtdp = result
+
+    if filtdp[0] == '/DCTDecode' \
+            and options.optimize >= 2:
+        # This is a simple heuristic derived from some training data, that has
+        # about a 70% chance of guessing whether the JPEG is high quality,
+        # and possibly recompressible, or not. The number itself doesn't mean
+        # anything.
+        # bytes_per_pixel = int(raw_jpeg.Length) / (w * h)
+        # jpeg_quality_estimate = 117.0 * (bytes_per_pixel ** 0.213)
+        # if jpeg_quality_estimate < 65:
+        #     return None
+
+        # We could get the ICC profile here, but there's no need to look at it
+        # for quality transcoding
+        # if icc:
+        #     stream = BytesIO(raw_jpeg.read_raw_bytes())
+        #     iccbytes = icc.read_bytes()
+        #     with Image.open(stream) as im:
+        #         im.save(jpg_name(root, xref), icc_profile=iccbytes)
+        try:
+            imgname = Path(root / '{:08d}'.format(xref))
+            with imgname.open('wb') as f:
+                ext = pim.extract_to(stream=f)
+            imgname.rename(imgname.with_suffix(ext))
+        except pikepdf.UnsupportedImageTypeError:
+            return None
+        return xref, ext
+    elif pim.indexed \
+            and pim.colorspace in pim.SIMPLE_COLORSPACES \
+            and options.optimize >= 3:
+        # Try to improve on indexed images - these are far from low hanging
+        # fruit in most cases
+        pim.as_pil_image().save(png_name(root, xref))
+        return xref, '.png'
+    elif not pim.indexed and pim.colorspace in pim.SIMPLE_COLORSPACES:
+        # An optimization opportunity here, not currently taken, is directly
+        # generating a PNG from compressed data
+        pim.as_pil_image().save(png_name(root, xref))
+        return xref, '.png'
+
+    return None
+
+
+
+def extract_images(pike, root, log, options, extract_fn):
+    """Extract image using extract_fn
+
+    extract_fn decides where the image is interesting in this case
+    """
+
+    include_xrefs = set()
+    exclude_xrefs = set()
+    errors = 0
+    for pageno, page in enumerate(pike.pages):
+        try:
+            xobjs = page.Resources.XObject
+        except AttributeError:
+            continue
+        for _imname, image in dict(xobjs).items():
+            if image.objgen[1] != 0:
+                continue  # Ignore images in an incremental PDF
+            xref = image.objgen[0]
+            if hasattr(image, 'SMask'):
+                # Ignore soft masks
+                smask_xref = image.SMask.objgen[0]
+                exclude_xrefs.add(smask_xref)
+            include_xrefs.add(xref)
+
+    working_xrefs = include_xrefs - exclude_xrefs
+    for xref in working_xrefs:
+        image = pike.get_object((xref, 0))
+        try:
+            result = extract_fn(
+                pike=pike, root=root, log=log, image=image,
+                xref=xref, options=options
+            )
+        except Exception as e:
+            log.debug("Image xref %s", xref)
+            log.debug(repr(e))
+            errors += 1
+        else:
+            if result:
+                _, ext = result
+                yield pageno, xref, ext
+
+
+def extract_images_generic(pike, root, log, options):
+    """Extract any >=2bpp image we think we can improve"""
+
+    jpegs = []
+    pngs = []
+    for _, xref, ext in extract_images(
+            pike, root, log, options, extract_image_generic):
+        log.debug('xref = %s ext = %s', xref, ext)
+        if ext == '.png':
+            pngs.append(xref)
+        elif ext == '.jpg':
+            jpegs.append(xref)
+    log.debug(
+        "Optimizable images: "
+        "JPEGs: %s PNGs: %s", len(jpegs), len(pngs)
+    )
+    return jpegs, pngs
+
+
+def extract_images_jbig2(pike, root, log, options):
+    """Extract any bitonal image that we think we can improve as JBIG2"""
+
+    jbig2_groups = defaultdict(list)
+    for pageno, xref, ext in extract_images(
+            pike, root, log, options, extract_image_jbig2):
+        group = pageno // options.jbig2_page_group_size
+        jbig2_groups[group].append((xref, ext))
+
+    # Elide empty groups
+    jbig2_groups = {group: xrefs for group, xrefs in jbig2_groups.items()
+                    if len(xrefs) > 0}
+    log.debug(
+        "Optimizable images: "
+        "JBIG2 groups: %s", (len(jbig2_groups),)
+    )
+    return jbig2_groups
+
+
+def _produce_jbig2_images(jbig2_groups, root, log, options):
+    """Produce JBIG2 images from their groups"""
+
+    def jbig2_group_futures(executor, root, groups):
+        for group, xref_exts in groups.items():
+            prefix = 'group{:08d}'.format(group)
+            future = executor.submit(
+                jbig2enc.convert_group,
+                cwd=fspath(root),
+                infiles=(img_name(root, xref, ext) for xref, ext in xref_exts),
+                out_prefix=prefix
+            )
+            yield future
+
+    def jbig2_single_futures(executor, root, groups):
+        for group, xref_exts in groups.items():
+            prefix = 'group{:08d}'.format(group)
+            # Second loop is to ensure multiple images per page are unpacked
+            for n, xref_ext in enumerate(xref_exts):
+                xref, ext = xref_ext
+                future = executor.submit(
+                    jbig2enc.convert_single,
+                    cwd=fspath(root),
+                    infile=img_name(root, xref, ext),
+                    outfile=root / ('{}.{:04d}'.format(prefix, n))
+                )
+                yield future
+
+    if options.jbig2_page_group_size > 1:
+        jbig2_futures = jbig2_group_futures
+    else:
+        jbig2_futures = jbig2_single_futures
+
+    with concurrent.futures.ThreadPoolExecutor(
+            max_workers=options.jobs) as executor:
+        futures = jbig2_futures(executor, root, jbig2_groups)
+        for future in concurrent.futures.as_completed(futures):
+            proc = future.result()
+            log.debug(proc.stderr.decode())
+
+
+def convert_to_jbig2(pike, jbig2_groups, root, log, options):
+    """Convert images to JBIG2 and insert into PDF.
+
+    When the JBIG2 page group size is > 1 we do several JBIG2 images at once
+    and build a symbol dictionary that will span several pages. Each JBIG2
+    image must reference to its symbol dictionary. If too many pages shared the
+    same dictionary JBIG2 encoding becomes more expensive and less efficient.
+    The default value of 10 was determined through testing. Currently this
+    must be lossy encoding since jbig2enc does not support refinement coding.
+
+    When the JBIG2 symbolic coder is not used, each JBIG2 stands on its own
+    and needs no dictionary. Currently this is must be lossless JBIG2.
+    """
+
+    _produce_jbig2_images(jbig2_groups, root, log, options)
+
+    for group, xref_exts in jbig2_groups.items():
+        prefix = 'group{:08d}'.format(group)
+        jbig2_symfile = root / (prefix + '.sym')
+        if jbig2_symfile.exists():
+            jbig2_globals_data = jbig2_symfile.read_bytes()
+            jbig2_globals = pikepdf.Stream(pike, jbig2_globals_data)
+            jbig2_globals_dict = pikepdf.Dictionary({
+                    '/JBIG2Globals': jbig2_globals
+            })
+        elif options.jbig2_page_group_size == 1:
+            jbig2_globals_dict = None
+        else:
+            raise FileNotFoundError(jbig2_symfile)
+
+        for n, xref_ext in enumerate(xref_exts):
+            xref, _ = xref_ext
+            jbig2_im_file = root / (prefix + '.{:04d}'.format(n))
+            jbig2_im_data = jbig2_im_file.read_bytes()
+            im_obj = pike.get_object(xref, 0)
+            im_obj.write(
+                jbig2_im_data, pikepdf.Name('/JBIG2Decode'),
+                jbig2_globals_dict
+            )
+
+
+def transcode_jpegs(pike, jpegs, root, log, options):
+    for xref in jpegs:
+        in_jpg = Path(jpg_name(root, xref))
+        opt_jpg = in_jpg.with_suffix('.opt.jpg')
+
+        # This produces a debug warning from PIL
+        # DEBUG:PIL.Image:Error closing: 'NoneType' object has no attribute
+        # 'close'.  Seems to be mostly harmless
+        # https://github.com/python-pillow/Pillow/issues/1144
+        with Image.open(fspath(in_jpg)) as im:
+            im.save(fspath(opt_jpg),
+                    optimize=True,
+                    quality=options.jpeg_quality)
+        # pylint: disable=no-member
+        if opt_jpg.stat().st_size > in_jpg.stat().st_size:
+            log.debug("xref %s, jpeg, made larger - skip", xref)
+            continue
+
+        compdata = leptonica.CompressedData.open(opt_jpg)
+        im_obj = pike.get_object(xref, 0)
+        im_obj.write(
+            compdata.read(), filter=pikepdf.Name('/DCTDecode')
+        )
+
+
+def transcode_pngs(pike, pngs, root, log, options):
+    if options.optimize >= 2:
+        png_quality = (
+            max(10, options.png_quality - 10),
+            min(100, options.png_quality + 10)
+        )
+        with concurrent.futures.ThreadPoolExecutor(
+                max_workers=options.jobs) as executor:
+            for xref in pngs:
+                executor.submit(
+                    pngquant.quantize,
+                    png_name(root, xref), png_name(root, xref),
+                    png_quality[0], png_quality[1])
+
+    for xref in pngs:
+        im_obj = pike.get_object(xref, 0)
+
+        # Open, transcode (!), package for PDF
+        try:
+            pix = leptonica.Pix.open(png_name(root, xref))
+            if pix.depth == 1:
+                pix = pix.invert()  # PDF assumes 1 is black for monochrome
+            compdata = pix.generate_pdf_ci_data(
+                leptonica.lept.L_FLATE_ENCODE, 0
+            )
+        except leptonica.LeptonicaError as e:
+            log.error(e)
+            continue
+
+        # This is what we should be doing: open the compressed data without
+        # transcoding. However this shifts each pixel row by one for some
+        # reason.
+        #compdata = leptonica.CompressedData.open(png_name(root, xref))
+        if len(compdata) > int(im_obj.stream_dict.Length):
+            continue  # If we produced a larger image, don't use
+
+        predictor = None
+        if compdata.predictor > 0:
+            predictor = pikepdf.Dictionary({'/Predictor': compdata.predictor})
+
+        im_obj.BitsPerComponent = compdata.bps
+        im_obj.Width = compdata.w
+        im_obj.Height = compdata.h
+
+        if compdata.ncolors > 0:
+            palette_pdf_string = compdata.get_palette_pdf_string()
+            palette_data = pikepdf.Object.parse(palette_pdf_string)
+            palette_stream = pikepdf.Stream(pike, bytes(palette_data))
+            palette = [pikepdf.Name('/Indexed'), pikepdf.Name('/DeviceRGB'),
+                       compdata.ncolors - 1, palette_stream]
+            cs = palette
+        else:
+            if compdata.spp == 1:
+                cs = pikepdf.Name('/DeviceGray')
+            elif compdata.spp == 3:
+                cs = pikepdf.Name('/DeviceRGB')
+            elif compdata.spp == 4:
+                cs = pikepdf.Name('/DeviceCMYK')
+        im_obj.ColorSpace = cs
+        im_obj.write(
+            compdata.read(),
+            filter=pikepdf.Name('/FlateDecode'), decode_parms=predictor
+        )
+
+
+def optimize(
+    input_file,
+    output_file,
+    log,
+    context):
+
+    options = context.get_options()
+    if options.optimize == 0:
+        re_symlink(input_file, output_file, log)
+        return
+
+    if options.jpeg_quality == 0:
+        options.jpeg_quality = \
+                DEFAULT_JPEG_QUALITY if options.optimize < 3 else 40
+    if options.png_quality == 0:
+        options.png_quality = \
+                DEFAULT_PNG_QUALITY if options.optimize < 3 else 30
+    if options.jbig2_page_group_size == 0:
+        options.jbig2_page_group_size = \
+                10 if options.jbig2_lossy else 1
+
+    pike = pikepdf.Pdf.open(input_file)
+
+    root = Path(output_file).parent / 'images'
+    root.mkdir(exist_ok=True)  # pylint: disable=no-member
+
+    jpegs, pngs = extract_images_generic(pike, root, log, options)
+    transcode_jpegs(pike, jpegs, root, log, options)
+    transcode_pngs(pike, pngs, root, log, options)
+
+    jbig2_groups = extract_images_jbig2(pike, root, log, options)
+    convert_to_jbig2(pike, jbig2_groups, root, log, options)
+
+    target_file = Path(output_file).with_suffix('.opt.pdf')
+    pike.remove_unreferenced_resources()
+    pike.save(target_file, preserve_pdfa=True,
+              object_stream_mode=pikepdf.ObjectStreamMode.generate)
+
+    input_size = Path(input_file).stat().st_size
+    output_size = Path(target_file).stat().st_size
+    ratio = input_size / output_size
+    savings = 1 - output_size / input_size
+    log.info("Optimize ratio: {:.2f} savings: {:.1f}%".format(
+        ratio, 100 * savings))
+
+    if savings < 0:
+        log.info("Optimize did not improve the file - discarded")
+        re_symlink(input_file, output_file, log)
+    else:
+        re_symlink(target_file, output_file, log)
+
+
+def main(infile, outfile, level, jobs=1):
+    from tempfile import TemporaryDirectory
+    from shutil import copy
+
+    class OptimizeOptions:
+        """Emulate ocrmypdf's options"""
+
+        def __init__(
+                self, jobs, optimize, jpeg_quality, png_quality, jb2lossy):
+            self.jobs = jobs
+            self.optimize = optimize
+            self.jpeg_quality = jpeg_quality
+            self.png_quality = png_quality
+            self.jbig2_page_group_size = 0
+            self.jbig2_lossy = jb2lossy
+
+    logging.basicConfig(level=logging.DEBUG)
+    log = logging.getLogger()
+
+    ctx = JobContext()
+    options = OptimizeOptions(
+        jobs=jobs,
+        optimize=int(level),
+        jpeg_quality=0,  # Use default
+        png_quality=0,
+        jb2lossy=False
+    )
+    ctx.set_options(options)
+
+    with TemporaryDirectory() as td:
+        tmpout = Path(td) / 'out.pdf'
+        optimize(infile, tmpout, log, ctx)
+        copy(fspath(tmpout), fspath(outfile))
+
+
+if __name__ == '__main__':
+    main(sys.argv[1], sys.argv[2], sys.argv[3])

src/ocrmypdf/pdfa.py

@@ -15,16 +15,32 @@
 # You should have received a copy of the GNU General Public License
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
-# Generate a PDFA_def.ps file for Ghostscript >= 9.14
+"""
+Generate a PDFMARK file for Ghostscript >= 9.14, for PDF/A conversion
+
+pdfmark is an extension to the Postscript language that describes some PDF
+features like bookmarks and annotations. It was originally specified Adobe
+Distiller, for Postscript to PDF conversion:
+https://www.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/pdfmark_reference.pdf
+
+Ghostscript uses pdfmark for PDF to PDF/A conversion as well. To use Ghostscript
+to create a PDF/A, we need to create a pdfmark file with the necessary metadata.
+
+This takes care of the many version-specific bugs and pecularities in
+Ghostscript's handling of pdfmark.
+
+"""
 
-from string import Template
 from binascii import hexlify
 from datetime import datetime
-from xml.parsers.expat import ExpatError
+from pathlib import Path
+from string import Template
 import pkg_resources
-import PyPDF2 as pypdf
-from defusedxml.minidom import parseString as defused_parseString
-from unittest.mock import patch
+import os
+
+from libxmp.utils import file_to_dict
+from libxmp import consts
+
 
 ICC_PROFILE_RELPATH = 'data/sRGB.icc'
 
@@ -36,12 +52,8 @@ SRGB_ICC_PROFILE = pkg_resources.resource_filename(
 # files, from the Ghostscript documentation. Lines beginning with % are
 # comments. Python substitution variables have a '$' prefix.
 pdfa_def_template = u"""%!
-% This is derived from Ghostscript's template for creating a PDF/A document.
-% This is a small PostScript program that includes some necessary information
-% to create a PDF/A compliant file.
-
 % Define entries in the document Info dictionary :
-/ICCProfile ($icc_profile)
+/ICCProfile $icc_profile
 def
 
 [$docinfo
@@ -78,7 +90,8 @@ def
 
 
 def encode_text_string(s: str) -> str:
-    '''Encode text string to hex string for use in a PDF
+    """
+    Encode text string to hex string for use in a PDF
 
     From PDF 32000-1:2008 a string object may be included in hexademical form
     if it is enclosed in angle brackets.  For general Unicode the string should
@@ -86,7 +99,7 @@ def encode_text_string(s: str) -> str:
     ASCII strings could be encoded as PdfDocEncoding literals provided
     that certain Postscript sequences are escaped.  But it's far simpler to
     encode everything as UTF-16.
-    '''
+    """
 
     # Sometimes lazy C programmers leave their NULs at the end of strings they
     # insert into PDFs
@@ -102,10 +115,30 @@ def encode_text_string(s: str) -> str:
     return ascii_hex_str
 
 
-def encode_pdf_date(d: datetime) -> str:
-    """Encode Python datetime object as PDF date string
+def _encode_ascii(s: str) -> str:
+    """
+    Aggressively strip non-ASCII and PDF escape sequences
 
-    From Adobe pdfmark manual:    
+    Ghostscript 9.24+ lost support for UTF-16BE in pdfmark files for reasons
+    given in GhostPDL commit e997c683. Our temporary workaround is use ASCII
+    and drop all non-ASCII characters. A slightly improved alternative would
+    be to implement PdfDocEncoding in pikepdf and encode to that, or handle
+    metadata there.
+    """
+    trans = str.maketrans({
+        '(': '',
+        ')': '',
+        '\\': '',
+        '\0': ''
+    })
+    return s.translate(trans).encode('ascii', errors='replace').decode()
+
+
+def encode_pdf_date(d: datetime) -> str:
+    """
+    Encode Python datetime object as PDF date string
+
+    From Adobe pdfmark manual:
     (D:YYYYMMDDHHmmSSOHH'mm')
     D: is an optional prefix. YYYY is the year. All fields after the year are
     optional. MM is the month (01-12), DD is the day (01-31), HH is the
@@ -137,27 +170,36 @@ def encode_pdf_date(d: datetime) -> str:
 
 
 def decode_pdf_date(s: str) -> datetime:
+    """
+    Decode a pdfmark date to a Python datetime object
+
+    A pdfmark date is a string in a paritcular format. See the pdfmark
+    Reference for the specification.
+
+    """
     if s.startswith('D:'):
         s = s[2:]
 
-    # Literal Z00'00', is incorrect but found in the wild, 
+    # Literal Z00'00', is incorrect but found in the wild,
     # probably made by OS X Quartz -- standardize
     if s.endswith("Z00'00'"):
         s = s.replace("Z00'00'", '+0000')
     elif s.endswith('Z'):
         s = s.replace('Z', '+0000')
-    
+
     s = s.replace("'", "")  # Remove apos from PDF time strings
 
     return datetime.strptime(s, r'%Y%m%d%H%M%S%z')
 
 
 def _get_pdfmark_dates(pdfmark):
-    """Encode dates for pdfmark Postscript.  The best way to deal with a
-    missing date entry is set it to null, because if the key is omitted 
-    Ghostscript will set it to now - we do not want to erase the fact that
-    the value was unknown.  Setting to an empty string breaks Ghostscript
-    9.22 as reported here:
+    """
+    Encode dates in the expected format for pdfmark Postscript
+
+    The best way to deal with amissing date entry is set it to null, because if
+    the key is omitted Ghostscript will set it to now - we do not want to erase
+    the fact that the value was unknown.  Setting to an empty string breaks
+    Ghostscript 9.22 as reported here:
     https://bugs.ghostscript.com/show_bug.cgi?id=699182
     """
 
@@ -172,30 +214,48 @@ def _get_pdfmark_dates(pdfmark):
             date_str = date_str[2:]
         try:
             yield '  {} (D:{})'.format(
-                    key, 
+                    key,
                     encode_pdf_date(decode_pdf_date(date_str)))
         except ValueError:
             yield '  {} null'.format(key)
 
 
-def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
-    """Create a Postscript file for Ghostscript.  pdfmark contains the various
-    objects as strings; these must be encoded in ASCII, and dates have a 
-    special format."""
+def _get_pdfa_def(icc_profile, icc_identifier, pdfmark, ascii_docinfo=False):
+    """
+    Create a Postscript pdfmark file for Ghostscript.
+
+    pdfmark contains the various objects as strings; these must be encoded in
+    ASCII, and dates have a special format.
+
+    :param icc_profile: filename of the ICC profile to include in pdfmark
+    :param icc_identifier: ICC identifier such as 'sRGB'
+    :param pdfmark: a dictionary containing keys to include the pdfmark
+    :param ascii_docinfo: if True, the docinfo block must be encoded in pure
+        ASCII and may not contain UTF-16BE-BOM-hex encoded strings, as
+        required for Ghostscript 9.24+
+
+    :returns: a string containing the entire pdfmark
+
+    """
 
     # Ghostscript <= 9.21 has a bug where null entries in DOCINFO might produce
     # ERROR: VMerror (-25) on closing pdfwrite device.
     # https://bugs.ghostscript.com/show_bug.cgi?id=697684
     # Work around this by only adding keys that have a nontrivial value
     docinfo_keys = ('/Title', '/Author', '/Subject', '/Creator', '/Keywords')
-    docinfo_line_template = '  {key} <{value}>'
 
     def docinfo_gen():
+        if not ascii_docinfo:
+            docinfo_line_template = '  {key} <{value}>'
+            encode = encode_text_string
+        else:
+            docinfo_line_template = '  {key} ({value})'
+            encode = _encode_ascii
         yield from _get_pdfmark_dates(pdfmark)
         for key in docinfo_keys:
             if key in pdfmark and pdfmark[key].strip() != '':
                 line = docinfo_line_template.format(
-                    key=key, value=encode_text_string(pdfmark[key]))
+                    key=key, value=encode(pdfmark[key]))
                 yield line
     docinfo = '\n'.join(docinfo_gen())
 
@@ -206,18 +266,28 @@ def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
     return result
 
 
-def generate_pdfa_ps(target_filename, pdfmark, icc='sRGB'):
+def generate_pdfa_ps(target_filename, pdfmark, icc='sRGB', ascii_docinfo=False):
     if icc == 'sRGB':
         icc_profile = SRGB_ICC_PROFILE
     else:
         raise NotImplementedError("Only supporting sRGB")
 
-    ps = _get_pdfa_def(icc_profile, icc, pdfmark)
+    # pdfmark must contain the full path to the ICC profile, and pdfmark must
+    # also encoded in ASCII. ocrmypdf can be installed anywhere, including to
+    # paths that have a non-ASCII character in the filename. Ghostscript
+    # accepts hex-encoded strings and converts them to byte strings, so
+    # we encode the path with fsencode() and use the hex representation.
+    # UTF-16 not accepted here. (Even though ASCII encodable is the usual case,
+    # do this always to avoid making it a rare conditional.)
+    bytes_icc_profile = os.fsencode(icc_profile)
+    hex_icc_profile = hexlify(bytes_icc_profile)
+    icc_profile = '<' + hex_icc_profile.decode('ascii') + '>'
+
+    ps = _get_pdfa_def(icc_profile, icc, pdfmark, ascii_docinfo=ascii_docinfo)
 
     # We should have encoded everything to pure ASCII by this point, and
     # to be safe, only allow ASCII in PostScript
-    with open(target_filename, 'w', encoding='ascii') as f:
-        f.write(ps)
+    Path(target_filename).write_text(ps, encoding='ascii')
 
 
 def file_claims_pdfa(filename):
@@ -229,38 +299,35 @@ def file_claims_pdfa(filename):
 
     This checks if the XMP metadata contains a PDF/A marker.
     """
-    pdf = pypdf.PdfFileReader(filename)
-    try:
-        # Monkeypatch PyPDF2 to use defusedxml as its XML parser, for safety
-        with patch('xml.dom.minidom.parseString', new=defused_parseString):
-            xmp = pdf.getXmpMetadata()
-    except ExpatError:
-        return {'pass': False, 'output': 'pdf',
-                'conformance': 'Invalid XML metadata'}
 
-    try:
-        pdfa_nodes = xmp.getNodesInNamespace(
-            aboutUri='',
-            namespace='http://www.aiim.org/pdfa/ns/id/')
-    except AttributeError:
+    xmp = file_to_dict(filename)
+    if not xmp:
         return {'pass': False, 'output': 'pdf',
                 'conformance': 'No XMP metadata'}
 
-    pdfa_dict = {attr.localName: attr.value for attr in pdfa_nodes}
-    if not pdfa_dict:
+    if not consts.XMP_NS_PDFA_ID in xmp:
         return {'pass': False, 'output': 'pdf',
-                'conformance': 'No XMP metadata'}
+                'conformance': 'No PDF/A metadata in XMP'}
 
-    part_conformance = pdfa_dict['part'] + pdfa_dict['conformance']
+    pdfa_node = xmp[consts.XMP_NS_PDFA_ID]
+    def read_node(node, key):
+        return next(
+            (v for k, v, meta in node if k == key), ''
+        )
+
+    part = read_node(pdfa_node, 'pdfaid:part')
+    conformance = read_node(pdfa_node, 'pdfaid:conformance')
+
+    part_conformance = part + conformance
     valid_part_conforms = {'1A', '1B', '2A', '2B', '2U', '3A', '3B', '3U'}
 
     conformance = 'PDF/A-{}'.format(
         part_conformance)
 
+    pdfa_dict = {}
     if part_conformance in valid_part_conforms:
         pdfa_dict['pass'] = True
         pdfa_dict['output'] = 'pdfa'
     pdfa_dict['conformance'] = conformance
 
     return pdfa_dict
-

src/ocrmypdf/pdfinfo/__init__.py

@@ -16,25 +16,22 @@
 # You should have received a copy of the GNU General Public License
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
-from subprocess import Popen, PIPE
-from decimal import Decimal
-from math import hypot, isclose
-import re
-import sys
 from collections import namedtuple
-from collections.abc import MutableMapping, Mapping
-import warnings
-from pathlib import Path
+from decimal import Decimal
 from enum import Enum
-from contextlib import contextmanager
+from math import hypot, isclose
+from pathlib import Path
+from unittest.mock import Mock
+import re
 
-import PyPDF2 as pypdf
-from .lib import fitz
-from .helpers import universal_open, fspath
+from pikepdf import PdfMatrix
+import pikepdf
 
+from . import ghosttext
+from .layout import get_page_analysis, get_text_boxes
 
+from ..helpers import fspath
 
-matrix_mult = pypdf.pdf.utils.matrixMultiply
 
 Colorspace = Enum('Colorspace',
                   'gray rgb cmyk lab icc index sep devn pattern jpeg2000')
@@ -43,7 +40,6 @@ Encoding = Enum('Encoding',
                 'ccitt jpeg jpeg2000 jbig2 asciihex ascii85 lzw flate ' + \
                 'runlength')
 
-
 FRIENDLY_COLORSPACE = {
     '/DeviceGray': Colorspace.gray,
     '/CalGray': Colorspace.gray,
@@ -87,28 +83,6 @@ FRIENDLY_COMP = {
 
 UNIT_SQUARE = (1.0, 0.0, 0.0, 1.0, 0.0, 0.0)
 
-
-def _matrix_from_shorthand(shorthand):
-    """Convert from PDF matrix shorthand to full matrix
-
-    PDF 1.7 spec defines a shorthand for describing the entries of a matrix
-    since the last column is always (0, 0, 1).
-    """
-
-    a, b, c, d, e, f = map(float, shorthand)
-    return ((a, b, 0),
-            (c, d, 0),
-            (e, f, 1))
-
-
-def _shorthand_from_matrix(matrix):
-    """Convert from transformation matrix to PDF shorthand."""
-    a, b = matrix[0][0], matrix[0][1]
-    c, d = matrix[1][0], matrix[1][1]
-    e, f = matrix[2][0], matrix[2][1]
-    return tuple(map(float, (a, b, c, d, e, f)))
-
-
 def _is_unit_square(shorthand):
     values = map(float, shorthand)
     pairwise = zip(values, UNIT_SQUARE)
@@ -118,30 +92,33 @@ XobjectSettings = namedtuple('XobjectSettings',
     ['name', 'shorthand', 'stack_depth'])
 
 InlineSettings = namedtuple('InlineSettings',
-    ['settings', 'shorthand', 'stack_depth'])
+    ['iimage', 'shorthand', 'stack_depth'])
 
-ContentsInfo = namedtuple('ContentsInfo', 
-    ['xobject_settings', 'inline_images', 'found_text'])
+ContentsInfo = namedtuple('ContentsInfo',
+    ['xobject_settings', 'inline_images', 'found_vector'])
+
+TextboxInfo = namedtuple('TextboxInfo',
+    ['bbox', 'is_visible', 'is_corrupt'])
 
 
-def _normalize_stack(operations):
-    """Fix runs of qQ's in the stack
+class VectorInfo:
+    def __init__(self):
+        pass
 
-    For some reason PyPDF2 converts runs of qqq, QQ, QQQq, etc. into single
-    operations.  Break this silliness up and issue each stack operation
-    individually so we don't lose count.
 
-    """
-    for operands, command in operations:
-        if re.match(br'Q*q+$', command):   # Zero or more Q, one or more q
-            for char in command:           # Split into individual bytes
-                yield ([], bytes([char]))  # Yield individual bytes
+def _normalize_stack(graphobjs):
+    """Convert runs of qQ's in the stack into single graphobjs"""
+    for operands, operator in graphobjs:
+        operator = str(operator)
+        if re.match(r'Q*q+$', operator):   # Zero or more Q, one or more q
+            for char in operator:          # Split into individual
+                yield ([], char)           # Yield individual
         else:
-            yield (operands, command)
+            yield (operands, operator)
 
 
 def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE):
-    """Interpret the PDF content stream
+    """Interpret the PDF content stream.
 
     The stack represents the state of the PDF graphics stack.  We are only
     interested in the current transformation matrix (CTM) so we only track
@@ -158,54 +135,53 @@ def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE):
     page.
 
     PDF units suit our needs so we initialize ctm to the identity matrix.
-
-    PyPDF2 replaces inline images with a fake "INLINE IMAGE" operator.
-
     """
 
-    operations = contentstream.operations
     stack = []
-    ctm = _matrix_from_shorthand(initial_shorthand)
+    ctm = PdfMatrix(initial_shorthand)
     xobject_settings = []
     inline_images = []
-    found_text = False
+    found_vector = False
+    vector_ops = set('S s f F f* B B* b b*'.split())
+    image_ops = set('BI ID EI q Q Do cm'.split())
+    operator_whitelist = ' '.join(vector_ops | image_ops)
 
-    for n, op in enumerate(_normalize_stack(operations)):
-        operands, command = op
-        if command == b'q':
+    for n, graphobj in enumerate(_normalize_stack(
+            pikepdf.parse_content_stream(contentstream, operator_whitelist))):
+        operands, operator = graphobj
+
+        if operator == 'q':
             stack.append(ctm)
             if len(stack) > 32:
                 raise RuntimeError(
-                    "PDF graphics stack overflow, command %i" % n)
-        elif command == b'Q':
+                    "PDF graphics stack overflow, operator %i" % n)
+        elif operator == 'Q':
             try:
                 ctm = stack.pop()
             except IndexError:
                 raise RuntimeError(
-                    "PDF graphics stack underflow, command %i" % n)
-        elif command == b'cm':
-            ctm = matrix_mult(
-                _matrix_from_shorthand(operands), ctm)
-        elif command == b'Do':
+                    "PDF graphics stack underflow, operator %i" % n)
+        elif operator == 'cm':
+            ctm = PdfMatrix(operands) @ ctm
+        elif operator == 'Do':
             image_name = operands[0]
             settings = XobjectSettings(
-                name=image_name, shorthand=_shorthand_from_matrix(ctm),
+                name=image_name, shorthand=ctm.shorthand,
                 stack_depth=len(stack))
             xobject_settings.append(settings)
-        elif command == b'INLINE IMAGE':
-            settings = operands['settings']
+        elif operator == 'INLINE IMAGE':   # BI/ID/EI are grouped into this
+            iimage = operands[0]
             inline = InlineSettings(
-                settings=settings, shorthand=_shorthand_from_matrix(ctm),
+                iimage=iimage, shorthand=ctm.shorthand,
                 stack_depth=len(stack))
             inline_images.append(inline)
-        elif command in (b'Tj', b'TJ', b'"', b"'"):
-            found_text = True
-
+        elif operator in vector_ops:
+            found_vector = True
 
     return ContentsInfo(
         xobject_settings=xobject_settings,
         inline_images=inline_images,
-        found_text=found_text)
+        found_vector=found_vector)
 
 
 def _get_dpi(ctm_shorthand, image_size):
@@ -278,70 +254,47 @@ class ImageInfo:
     def __init__(self, *, name='', pdfimage=None, inline=None,
                  shorthand=None):
 
-        self._name = name
+        self._name = str(name)
         self._shorthand = shorthand
-        if inline:
-            # Fixme does not work for inline images with non abbreviated
-            # fields
+
+        if inline is not None:
             self._origin = 'inline'
-            self._width = inline.settings['/W']
-            self._height = inline.settings['/H']
-            self._bpc = inline.settings.get('/BPC', 8)
-            try:
-                self._color = FRIENDLY_COLORSPACE[inline.settings['/CS']]
-            except Exception:
-                self._color = '-'
-            self._comp = FRIENDLY_COMP.get(self._color, '?')
-            if '/F' in inline.settings:
-                filter_ = inline.settings['/F']
-                if isinstance(filter_, pypdf.generic.ArrayObject):
-                    filter_ = filter_[0]
-                self._enc = FRIENDLY_ENCODING.get(filter_, 'image')
-            else:
-                self._enc = 'image'
-        elif pdfimage:
+            pim = inline.iimage
+        elif pdfimage is not None:
             self._origin = 'xobject'
-            self._width = pdfimage['/Width']
-            self._height = pdfimage['/Height']
-            if '/BitsPerComponent' in pdfimage:
-                self._bpc = pdfimage['/BitsPerComponent']
-            else:
-                self._bpc = 8
+            pim = pikepdf.PdfImage(pdfimage)
+        self._width = pim.width
+        self._height = pim.height
 
-            # Fixme: this is incorrectly treats explicit masks as stencil masks,
-            # but good enough for now. Explicit masks have /ImageMask true but are
-            # never called for in content stream, instead are drawn as a /Mask on
-            # other images. For our purposes finding out the details of /Mask
-            # will seldom matter.
-            if '/ImageMask' in pdfimage:
-                self._type = 'stencil' if pdfimage['/ImageMask'].value \
-                    else 'image'
-            else:
-                self._type = 'image'
-            if '/Filter' in pdfimage:
-                filter_ = pdfimage['/Filter']
-                if isinstance(filter_, pypdf.generic.ArrayObject):
-                    filter_ = filter_[0]
-                self._enc = FRIENDLY_ENCODING.get(filter_, 'image')
-            else:
-                self._enc = 'image'
-            if '/ColorSpace' in pdfimage:
-                cs = pdfimage['/ColorSpace']
-                if isinstance(cs, pypdf.generic.ArrayObject):
-                    cs = cs[0]
-                self._color = FRIENDLY_COLORSPACE.get(cs, '-')
-            else:
-                self._color = FRIENDLY_COLORSPACE[Colorspace.jpeg2000] \
-                    if self._enc == Encoding.jpeg2000 else '?'
+        # If /ImageMask is true, then this image is a stencil mask
+        # (Images that draw with this stencil mask will have a reference to
+        # it in their /Mask, but we don't actually need that information)
+        if pim.image_mask:
+            self._type = 'stencil'
+        else:
+            self._type = 'image'
 
-            self._comp = FRIENDLY_COMP.get(self._color, '?')
+        self._bpc = int(pim.bits_per_component)
+        try:
+            self._enc = FRIENDLY_ENCODING.get(pim.filters[0], 'image')
+        except IndexError:
+            self._enc = '?'
 
-            # Bit of a hack... infer grayscale if component count is uncertain
-            # but encoding must be monochrome. This happens if a monochrome image
-            # has an ICC profile attached. Better solution would be to examine
-            # the ICC profile.
-            if self._comp == '?' and self._enc in (Encoding.ccitt, 'jbig2'):
-                self._comp = FRIENDLY_COMP[Colorspace.gray]
+        try:
+            self._color = FRIENDLY_COLORSPACE.get(pim.colorspace, '?')
+        except NotImplementedError:
+            self._color = '?'
+        if self._enc == Encoding.jpeg2000:
+            self._color = Colorspace.jpeg2000
+
+        self._comp = FRIENDLY_COMP.get(self._color, '?')
+
+        # Bit of a hack... infer grayscale if component count is uncertain
+        # but encoding must be monochrome. This happens if a monochrome image
+        # has an ICC profile attached. Better solution would be to examine
+        # the ICC profile.
+        if self._comp == '?' and self._enc in (Encoding.ccitt, 'jbig2'):
+            self._comp = FRIENDLY_COMP[Colorspace.gray]
 
     @property
     def name(self):
@@ -383,20 +336,6 @@ class ImageInfo:
     def yres(self):
         return _get_dpi(self._shorthand, (self._width, self._height))[1]
 
-    def __getitem__(self, item):
-        warnings.warn("ImageInfo.__getitem__", DeprecationWarning)
-        if item in ('name', 'width', 'height', 'bpc', 'color', 'comp', 'enc'):
-            return getattr(self, item)
-        elif item == 'dpi_w':
-            return Decimal(self.xres).quantize(self.DPI_PREC)
-        elif item == 'dpi_h':
-            return Decimal(self.yres).quantize(self.DPI_PREC)
-        elif item == 'dpi':
-            return Decimal(self.xres * self.yres).sqrt().quantize(
-                self.DPI_PREC)
-        else:
-            raise KeyError(item)
-
     def __repr__(self):
         class_locals = {attr: getattr(self, attr, None) for attr in dir(self)
                         if not attr.startswith('_')}
@@ -431,8 +370,9 @@ def _image_xobjects(container):
     resources = container['/Resources']
     if '/XObject' not in resources:
         return
-    for xobj in resources['/XObject']:
-        candidate = resources['/XObject'][xobj]
+    xobjs = resources['/XObject'].as_dict()
+    for xobj in xobjs:
+        candidate = xobjs[xobj]
         if candidate['/Subtype'] == '/Image':
             pdfimage = candidate
             yield (pdfimage, xobj)
@@ -445,7 +385,6 @@ def _find_regular_images(container, contentsinfo):
     that contains images.
 
     Generates images with their DPI at time of drawing.
-
     """
 
     for pdfimage, xobj in _image_xobjects(container):
@@ -479,8 +418,9 @@ def _find_form_xobject_images(pdf, container, contentsinfo):
     resources = container['/Resources']
     if '/XObject' not in resources:
         return
-    for xobj in resources['/XObject']:
-        candidate = resources['/XObject'][xobj]
+    xobjs = resources['/XObject'].as_dict()
+    for xobj in xobjs:
+        candidate = xobjs[xobj]
         if candidate['/Subtype'] != '/Form':
             continue
 
@@ -494,11 +434,11 @@ def _find_form_xobject_images(pdf, container, contentsinfo):
             # but in practice both Form XObjects and multiple drawing of the
             # same object are both very rare.
             ctm_shorthand = settings.shorthand
-            yield from _find_images(
+            yield from _process_content_streams(
                 pdf=pdf, container=form_xobject, shorthand=ctm_shorthand)
 
 
-def _find_images(*, pdf, container, shorthand=None):
+def _process_content_streams(*, pdf, container, shorthand=None):
     """Find all individual instances of images drawn in the container
 
     Usually the container is a page, but it may also be a Form XObject.
@@ -507,7 +447,7 @@ def _find_images(*, pdf, container, shorthand=None):
     in an XObject.
 
     Form XObjects may include inline images, XObject images,
-    and recursively, other Form XObjects; and also vector drawing commands.
+    and recursively, other Form XObjects; and also vector graphic objects.
 
     Every instance of an image being drawn somewhere is flattened and
     treated as a unique image, since if the same image is drawn multiple times
@@ -518,99 +458,104 @@ def _find_images(*, pdf, container, shorthand=None):
     """
 
     if container.get('/Type') == '/Page' and '/Contents' in container:
-        # For a /Page the content stream is attached to the page's /Contents
-        page = container
-        contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
         initial_shorthand = shorthand or UNIT_SQUARE
     elif container.get('/Type') == '/XObject' and \
             container['/Subtype'] == '/Form':
-        # For a Form XObject that content stream is attached to the XObject
-        contentstream = pypdf.pdf.ContentStream(container, pdf)
-
         # Set the CTM to the state it was when the "Do" operator was
         # encountered that is drawing this instance of the Form XObject
-        ctm = _matrix_from_shorthand(shorthand or UNIT_SQUARE)
+        ctm = PdfMatrix(shorthand) if shorthand else PdfMatrix.identity()
 
         # A Form XObject may provide its own matrix to map form space into
         # user space. Get this if one exists
-        form_matrix = _matrix_from_shorthand(
-                container.get('/Matrix', UNIT_SQUARE))
+        form_shorthand = container.get('/Matrix', PdfMatrix.identity())
+        form_matrix = PdfMatrix(form_shorthand)
 
         # Concatenate form matrix with CTM to ensure CTM is correct for
         # drawing this instance of the XObject
-        ctm = matrix_mult(form_matrix, ctm)
-        initial_shorthand = _shorthand_from_matrix(ctm)
+        ctm = form_matrix @ ctm
+        initial_shorthand = ctm.shorthand
     else:
         return
 
-    contentsinfo = _interpret_contents(contentstream, initial_shorthand)
+    contentsinfo = _interpret_contents(container, initial_shorthand)
 
+    if contentsinfo.found_vector:
+        yield VectorInfo()
     yield from _find_inline_images(contentsinfo)
     yield from _find_regular_images(container, contentsinfo)
     yield from _find_form_xobject_images(pdf, container, contentsinfo)
 
 
-def _naive_find_text(*, pdf, page):
-    if not(page.get('/Type') == '/Page' and '/Contents' in page):
-        # Not a page, or has no /Contents => no text
-        return False
+def _page_has_text(text_blocks, page_width, page_height):
+    """Smarter text detection that ignores text in margins"""
 
-    # First we check the main content stream    
-    contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
-    contentsinfo = _interpret_contents(contentstream, UNIT_SQUARE)
-    if contentsinfo.found_text:
-        return True
+    pw, ph = float(page_width), float(page_height)
 
-    # Then see if there is a Form XObject with with a content stream
-    # that might have text.  For full completeness we should recursively
-    # search nested Form XObjects, as we do with images.  But that is
-    # rare.
-    if '/Resources' in page:
-        resources = page['/Resources']
-        if '/XObject' in resources:    
-            for xobj in resources['/XObject']:
-                candidate = resources['/XObject'][xobj]
-                if candidate['/Subtype'] != '/Form':
-                    continue
-                form_xobject = candidate                
-                # Content stream is attached to Form XObject dictionary
-                contentstream = pypdf.pdf.ContentStream(form_xobject, pdf)
-                sub_contentsinfo = _interpret_contents(
-                    contentstream, UNIT_SQUARE)
-                if sub_contentsinfo.found_text:
-                    return True
-    return False
+    margin_ratio = 0.125
+    interior_bbox = (
+        margin_ratio * pw,  # left
+        (1 - margin_ratio) * ph, # top
+        (1 - margin_ratio) * pw, # right
+        margin_ratio * ph  # bottom  (first quadrant: bottom < top)
+    )
+
+    def rects_intersect(a, b):
+        """
+        Where (a,b) are 4-tuple rects (left-0, top-1, right-2, bottom-3)
+        https://stackoverflow.com/questions/306316/determine-if-two-rectangles-overlap-each-other
+        Formula assumes all boxes are in first quadrant
+        """
+        return a[0] < b[2] and a[2] > b[0] and a[1] > b[3] and a[3] < b[1]
+
+    has_text = False
+    for bbox in text_blocks:
+        if rects_intersect(bbox, interior_bbox):
+            has_text = True
+            break
+    return has_text
 
 
-def _page_has_text(infile, pageno):
-    doc = fitz.Document(infile)
-    text = doc.getPageText(pageno)
-    if text.strip() != '':
-        return True
-    return False
+def simplify_textboxes(miner):
+    """Extract only limited content from text boxes
+
+    We do this to save memory and ensure that our objects are pickleable.
+    """
+    for box in get_text_boxes(miner):
+        first_line = box._objs[0]
+        first_char = first_line._objs[0]
+
+        visible = (first_char.rendermode != 3)
+        corrupt = (first_char.get_text() == '\ufffd')
+        yield TextboxInfo(box.bbox, visible, corrupt)
 
 
-def _pdf_get_pageinfo(pdf, pageno: int, infile):
+def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext):
     pageinfo = {}
     pageinfo['pageno'] = pageno
     pageinfo['images'] = []
 
-    if isinstance(pdf, Path):
-        pdf = pypdf.PdfFileReader(str(pdf))
-    elif isinstance(pdf, str):
-        pdf = pypdf.PdfFileReader(pdf)
-
     page = pdf.pages[pageno]
+    mediabox = [Decimal(d) for d in page.MediaBox.as_list()]
+    width_pt = mediabox[2] - mediabox[0]
+    height_pt = mediabox[3] - mediabox[1]
 
-    if fitz:
-        pageinfo['has_text'] = _page_has_text(str(infile), pageno)
+    if xmltext:
+        bboxes = ghosttext.page_get_textblocks(
+            fspath(infile), pageno, xmltext=xmltext, height=height_pt)
+        pageinfo['bboxes'] = bboxes
     else:
-        pageinfo['has_text'] = _naive_find_text(pdf=pdf, page=page)
+        pscript5_mode = str(pdf.metadata.get('/Creator')).startswith('PScript5')
+        miner = get_page_analysis(infile, pageno, pscript5_mode)
+        pageinfo['textboxes'] = list(simplify_textboxes(miner))
+        bboxes = (box.bbox for box in pageinfo['textboxes'])
 
-    width_pt = page.mediaBox.getWidth()
-    height_pt = page.mediaBox.getHeight()
+    pageinfo['has_text'] = _page_has_text(
+        bboxes, width_pt, height_pt
+    )
 
     userunit = page.get('/UserUnit', Decimal(1.0))
+    if not isinstance(userunit, Decimal):
+        userunit = Decimal(userunit)
     pageinfo['userunit'] = userunit
     pageinfo['width_inches'] = width_pt * userunit / Decimal(72.0)
     pageinfo['height_inches'] = height_pt * userunit / Decimal(72.0)
@@ -621,12 +566,19 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile):
         pageinfo['rotate'] = 0
 
     userunit_shorthand = (userunit, 0, 0, userunit, 0, 0)
-    pageinfo['images'] = [im for im in
-                          _find_images(pdf=pdf, container=page,
-                                       shorthand=userunit_shorthand)]
+    contentsinfo = [ci for ci in
+                    _process_content_streams(pdf=pdf, container=page,
+                                             shorthand=userunit_shorthand)]
+
+    pageinfo['has_vector'] = False
+    if any(isinstance(ci, VectorInfo) for ci in contentsinfo):
+        pageinfo['has_vector'] = True
+
+    pageinfo['images'] = [im for im in contentsinfo
+                          if isinstance(im, ImageInfo)]
     if pageinfo['images']:
-        xres = max(image['dpi_w'] for image in pageinfo['images'])
-        yres = max(image['dpi_h'] for image in pageinfo['images'])
+        xres = Decimal(max(image.xres for image in pageinfo['images']))
+        yres = Decimal(max(image.yres for image in pageinfo['images']))
         pageinfo['xres'], pageinfo['yres'] = xres, yres
         pageinfo['width_pixels'] = \
             int(round(xres * pageinfo['width_inches']))
@@ -636,17 +588,31 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile):
     return pageinfo
 
 
-def _pdf_get_all_pageinfo(infile):
-    with universal_open(infile, 'rb') as f:
-        pdf = pypdf.PdfFileReader(f)
-        return [PageInfo(pdf, n, infile) for n in range(pdf.numPages)]
+def _pdf_get_all_pageinfo(infile, detailed_analysis=False, log=None):
+    if not log:
+        log = Mock()
+
+    pdf = pikepdf.open(infile)
+    if detailed_analysis:
+        pages_xml = None
+    else:
+        pages_xml = ghosttext.extract_text_xml(infile, pdf, pageno=None, log=log)
+
+    pages = []
+    for n in range(len(pdf.pages)):
+        page_xml = pages_xml[n] if pages_xml else None
+        page = PageInfo(pdf, n, infile, page_xml, detailed_analysis)
+        pages.append(page)
+
+    return pages, pdf
 
 
 class PageInfo:
-    def __init__(self, pdf, pageno, infile):
+    def __init__(self, pdf, pageno, infile, xmltext, detailed_analysis=False):
         self._pageno = pageno
         self._infile = infile
-        self._pageinfo = _pdf_get_pageinfo(pdf, pageno, infile)
+        self._pageinfo = _pdf_get_pageinfo(pdf, pageno, infile, xmltext)
+        self._detailed_analysis = detailed_analysis
 
     @property
     def pageno(self):
@@ -656,6 +622,16 @@ class PageInfo:
     def has_text(self):
         return self._pageinfo['has_text']
 
+    @property
+    def has_corrupt_text(self):
+        if not self._detailed_analysis:
+            raise NotImplementedError('Did not do detailed analysis')
+        return any(tbox.is_corrupt for tbox in self._pageinfo['textboxes'])
+
+    @property
+    def has_vector(self):
+        return self._pageinfo['has_vector']
+
     @property
     def width_inches(self):
         return self._pageinfo['width_inches']
@@ -687,6 +663,27 @@ class PageInfo:
     def images(self):
         return self._pageinfo['images']
 
+    def get_textareas(self, visible=None, corrupt=None):
+        def predicate(obj, want_visible, want_corrupt):
+            result = True
+            if want_visible is not None:
+                if obj.is_visible != want_visible:
+                    result = False
+            if want_corrupt is not None:
+                if obj.is_corrupt != want_corrupt:
+                    result = False
+            return result
+
+        if 'textboxes' not in self._pageinfo:
+            if visible is not None and corrupt is not None:
+                raise NotImplementedError(
+                    'Ghostscript textboxes cannot be classified')
+            return self._pageinfo['bboxes']
+
+        return (obj.bbox for obj in self._pageinfo['textboxes']
+                if predicate(obj, visible, corrupt))
+
+
     @property
     def xres(self):
         return self._pageinfo.get('xres', None)
@@ -717,16 +714,14 @@ class PageInfo:
 
 
 class PdfInfo:
-    """Get summary information about a PDF
-    
-    """
-    def __init__(self, infile):
+    """Get summary information about a PDF"""
+
+    def __init__(self, infile, detailed_page_analysis=False, log=None):
         self._infile = infile
-        self._pages = _pdf_get_all_pageinfo(infile)
-        if fitz:
-            self._toc = fitz.Document(fspath(infile)).getToC()
-        else:
-            self._toc = []
+        self._pages, pdf = _pdf_get_all_pageinfo(
+            infile, detailed_page_analysis, log=log)
+        self._needs_rendering = pdf.root.get('/NeedsRendering', False)
+        self._has_acroform = '/AcroForm' in pdf.root
 
     @property
     def pages(self):
@@ -741,6 +736,10 @@ class PdfInfo:
     def has_userunit(self):
         return any(page.userunit != 1.0 for page in self.pages)
 
+    @property
+    def has_acroform(self):
+        return self._has_acroform
+
     @property
     def filename(self):
         if not isinstance(self._infile, (str, Path)):
@@ -748,8 +747,8 @@ class PdfInfo:
         return self._infile
 
     @property
-    def table_of_contents(self):
-        return self._toc
+    def needs_rendering(self):
+        return self._needs_rendering
 
     def __getitem__(self, item):
         return self._pages[item]
@@ -760,15 +759,6 @@ class PdfInfo:
     def __repr__(self):
         return "<PdfInfo('...'), page count={}>".format(len(self))
 
-    # def __getstate__(self):
-    #     state = {'_infile': self._infile}
-    #     return state
-    #
-    # def __setstate__(self, state):
-    #     self._infile = state['_infile']
-    #     self._pages = _pdf_get_all_pageinfo(self._infile)
-
-
 def main():
     import argparse
     parser = argparse.ArgumentParser()

src/ocrmypdf/pdfinfo/ghosttext.py

@@ -0,0 +1,96 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+import re
+import xml.etree.ElementTree as ET
+
+from ..exec import ghostscript
+
+# Forgive me for I have sinned
+# I am using regular expressions to parse XML. However the XML in this case,
+# generated by Ghostscript, is self-consistent enough to be parseable.
+regex_remove_char_tags = re.compile(br"""
+    <char\b
+    (?:   [^>]   # anything single character but >
+        | \">\"  # special case: trap ">"
+    )*
+    />           # terminate with '/>'
+""", re.VERBOSE)
+
+
+def page_get_textblocks(infile, pageno, xmltext, height):
+    """Get text boxes out of Ghostscript txtwrite xml"""
+
+    root = xmltext
+    if not hasattr(xmltext, 'findall'):
+        return []
+
+    def blocks():
+        for span in root.findall('.//span'):
+            bbox_str = span.attrib['bbox']
+            font_size = span.attrib['size']
+            pts = [int(pt) for pt in bbox_str.split()]
+            pts[1] = pts[1] - int(float(font_size) + 0.5)
+            bbox_topdown = tuple(pts)
+            bb = bbox_topdown
+            bbox_bottomup = (bb[0], height - bb[3], bb[2], height - bb[1])
+            yield bbox_bottomup
+
+    def joined_blocks():
+        prev = None
+        for bbox in blocks():
+            if prev is None:
+                prev = bbox
+            if bbox[1] == prev[1] and bbox[3] == prev[3]:
+                gap = prev[2] - bbox[0]
+                height = abs(bbox[3] - bbox[1])
+                if gap < height:
+                    # Join boxes
+                    prev = (prev[0], prev[1], bbox[2], bbox[3])
+                    continue
+            # yield previously joined bboxes and start anew
+            yield prev
+            prev = bbox
+        if prev is not None:
+            yield prev
+
+    return [block for block in joined_blocks()]
+
+
+def extract_text_xml(infile, pdf, pageno=None, log=None):
+    existing_text = ghostscript.extract_text(infile, pageno=None)
+    existing_text = regex_remove_char_tags.sub(b' ', existing_text)
+
+    try:
+        root = ET.fromstringlist([
+            b'<document>\n', existing_text, b'</document>\n'
+        ])
+        page_xml = root.findall('page')
+    except ET.ParseError as e:
+        log.error(
+            "An error occurred while attempting to retrieve existing text in "
+            "the input file. Will attempt to continue assuming that there is "
+            "no existing text in the file. The error was:")
+        log.error(e)
+        page_xml = [None] * len(pdf.pages)
+
+    page_count_difference = len(pdf.pages) - len(page_xml)
+    if page_count_difference != 0:
+        log.error("The number of pages in the input file is inconsistent.")
+        if page_count_difference > 0:
+            page_xml.extend([None] * page_count_difference)
+    return page_xml

src/ocrmypdf/pdfinfo/layout.py

@@ -0,0 +1,237 @@
+# © 2018 James R. Barlow: github.com/jbarlow83
+#
+# This file is part of OCRmyPDF.
+#
+# OCRmyPDF is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# OCRmyPDF is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
+
+import re
+from math import copysign
+from pathlib import Path
+from unittest.mock import patch
+
+import pdfminer.encodingdb
+import pdfminer.pdfdevice
+import pdfminer.pdfinterp
+from pdfminer.converter import PDFLayoutAnalyzer
+from pdfminer.glyphlist import glyphname2unicode
+from pdfminer.layout import (LAParams, LTChar, LTContainer, LTLayoutContainer,
+                             LTPage, LTTextBox, LTTextLine)
+from pdfminer.pdfdocument import PDFTextExtractionNotAllowed
+from pdfminer.pdffont import (PDFCIDFont, PDFFont, PDFType3Font,
+                              PDFUnicodeNotDefined, PDFSimpleFont)
+from pdfminer.pdfpage import PDFPage
+from pdfminer.utils import bbox2str, fsplit, matrix2str
+
+from ..exceptions import EncryptedPdfError
+
+STRIP_NAME = re.compile(r'[0-9]+')
+
+#
+# Unconditional pdfminer patches
+#
+
+def name2unicode(name):
+    """Fix pdfminer's regex in name2unicode function
+
+    Font cids that are mapped to names of the form /g123 seem to be, by convention
+    characters with no corresponding Unicode entry. These can be subsetted fonts
+    or symbolic fonts. There seems to be no way to map /g123 fonts to Unicode,
+    barring a ToUnicode data structure.
+    """
+    if name in glyphname2unicode:
+        return glyphname2unicode[name]
+    if name.startswith('g') or name.startswith('a'):
+        raise KeyError(name)
+    if name.startswith('uni'):
+        try:
+            return chr(int(name[3:], 16))
+        except ValueError:  # Not hexadecimal
+            raise KeyError(name)
+    m = STRIP_NAME.search(name)
+    if not m:
+        raise KeyError(name)
+    return chr(int(m.group(0)))
+pdfminer.encodingdb.name2unicode = name2unicode
+
+original_PDFFont_init = PDFFont.__init__
+def PDFFont__init__(self, descriptor, widths, default_width=None):
+    original_PDFFont_init(self, descriptor, widths, default_width)
+    # PDF spec says descent should be negative
+    # A font with a positive descent implies it floats entirely above the
+    # baseline, i.e. it's not really a baseline anymore. I have fonts that
+    # claim a positive descent, but treating descent as positive always seems
+    # to misposition text.
+    if self.descent > 0:
+        self.descent = -self.descent
+PDFFont.__init__ = PDFFont__init__
+
+original_PDFSimpleFont_init = PDFSimpleFont.__init__
+def PDFSimpleFont__init__(self, descriptor, widths, spec):
+    # Font encoding is specified either by a name of
+    # built-in encoding or a dictionary that describes
+    # the differences.
+    original_PDFSimpleFont_init(self, descriptor, widths, spec)
+    # pdfminer is incorrect. If there is no ToUnicode and no Encoding, do not
+    # assume Unicode conversion is possible. RM 9.10.2
+    if not self.unicode_map and 'Encoding' not in spec:
+        self.cid2unicode = {}
+    return
+PDFSimpleFont.__init__ = PDFSimpleFont__init__
+#
+# pdfminer patches when creator is PScript5.dll
+#
+
+def PDFType3Font__PScript5_get_height(self):
+    h = self.bbox[3]-self.bbox[1]
+    if h == 0:
+        h = self.ascent - self.descent
+    return h * copysign(1.0, self.vscale)
+
+def PDFType3Font__PScript5_get_descent(self):
+    return self.descent * copysign(1.0, self.vscale)
+
+def PDFType3Font__PScript5_get_ascent(self):
+    return self.ascent * copysign(1.0, self.vscale)
+
+
+class LTStateAwareChar(LTChar):
+    """A subclass of LTChar that tracks text render mode at time of drawing"""
+
+    __slots__ = (
+        'rendermode', '_text', 'matrix', 'fontname', 'adv', 'upright', 'size',
+        'width', 'height', 'bbox', 'x0', 'x1', 'y0', 'y1'
+    )
+
+    def __init__(self, matrix, font, fontsize, scaling, rise, text, textwidth,
+                 textdisp, textstate, *args):
+        super().__init__(matrix, font, fontsize, scaling, rise, text, textwidth,
+                         textdisp, *args)
+        self.rendermode = textstate.render
+
+    def is_compatible(self, obj):
+        """Check if characters can be combined into a textline
+
+        We consider characters compatible if:
+            - the Unicode mapping is known, and both have the same render mode
+            - the Unicode mapping is unknown but both are part of the same font
+        """
+        both_unicode_mapped = (isinstance(self._text, str) and
+                               isinstance(obj._text, str))
+        try:
+            if both_unicode_mapped:
+                return self.rendermode == obj.rendermode
+            font0, _ = self._text
+            font1, _ = obj._text
+            return font0 == font1 and self.rendermode == obj.rendermode
+        except (ValueError, AttributeError):
+            return False
+
+    def get_text(self):
+        if isinstance(self._text, tuple):
+            return '�'
+        return self._text
+
+    def __repr__(self):
+        return ('<%s %s matrix=%s rendermode=%r font=%r adv=%s text=%r>' %
+                (self.__class__.__name__, bbox2str(self.bbox),
+                 matrix2str(self.matrix), self.rendermode, self.fontname, self.adv,
+                 self.get_text()))
+
+
+class TextPositionTracker(PDFLayoutAnalyzer):
+    """A page layout analyzer that pays attention to text visibility"""
+
+    def __init__(self, rsrcmgr, pageno=1, laparams=None):
+        super().__init__(rsrcmgr, pageno, laparams)
+        self.textstate = None
+        self.result = None
+
+    def begin_page(self, page, ctm):
+        super().begin_page(page, ctm)
+        self.cur_item = LTPage(self.pageno, page.mediabox)
+
+    def end_page(self, page):
+        assert not self._stack, str(len(self._stack))
+        assert isinstance(self.cur_item, LTPage), str(type(self.cur_item))
+        if self.laparams is not None:
+            self.cur_item.analyze(self.laparams)
+        self.pageno += 1
+        self.receive_layout(self.cur_item)
+
+    def render_string(self, textstate, seq, *args):
+        self.textstate = textstate.copy()
+        super().render_string(self.textstate, seq, *args)
+
+    def render_char(self, matrix, font, fontsize, scaling, rise, cid, *args):
+        try:
+            text = font.to_unichr(cid)
+            assert isinstance(text, str), str(type(text))
+        except PDFUnicodeNotDefined:
+            text = self.handle_undefined_char(font, cid)
+        textwidth = font.char_width(cid)
+        textdisp = font.char_disp(cid)
+        item = LTStateAwareChar(
+                matrix, font, fontsize, scaling, rise, text,
+                textwidth, textdisp, self.textstate, *args)
+        self.cur_item.add(item)
+        return item.adv
+
+    def handle_undefined_char(self, font, cid):
+        #log.info('undefined: %r, %r', font, cid)
+        return (font.fontname, cid)
+
+    def receive_layout(self, ltpage):
+        self.result = ltpage
+
+    def get_result(self):
+        return self.result
+
+
+def get_page_analysis(infile, pageno, pscript5_mode):
+    rman = pdfminer.pdfinterp.PDFResourceManager(caching=True)
+    dev = TextPositionTracker(rman, laparams=LAParams())
+    interp = pdfminer.pdfinterp.PDFPageInterpreter(rman, dev)
+
+    if pscript5_mode:
+        patcher = patch.multiple(
+            'pdfminer.pdffont.PDFType3Font',
+            spec=True,
+            get_ascent=PDFType3Font__PScript5_get_ascent,
+            get_descent=PDFType3Font__PScript5_get_descent,
+            get_height=PDFType3Font__PScript5_get_height
+        )
+        patcher.start()
+
+    try:
+        with Path(infile).open('rb') as f:
+            page = PDFPage.get_pages(f, pagenos=[pageno], maxpages=0)
+            interp.process_page(next(page))
+    except PDFTextExtractionNotAllowed:
+        raise EncryptedPdfError()
+    finally:
+        if pscript5_mode:
+            patcher.stop()
+
+    return dev.get_result()
+
+
+def get_text_boxes(obj):
+    for child in obj:
+        if isinstance(child, (LTTextBox)):
+            yield child
+        else:
+            try:
+                yield from get_text_boxes(child)
+            except TypeError:
+                continue

test_requirements.txt

@@ -1,5 +0,0 @@
-pytest >= 3.2
-pytest-helpers-namespace
-pytest-xdist
-pytest-cov
-pytest-timeout

tests/cache/2400dpi/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -1 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/2400dpi/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -1,25 +0,0 @@
-a la Waterman
-
-4 ons linzen
-
-3 liter water
-
-3 uien
-
-bloem, boter
-
-2 kopjes melk
-
-laurier, kruidnagel, kerrie, zout
-
-De linzgen wassen en in -l liter kokend wa-
-ter 1 dag laten weken, 2 liter water bij
-de linzen voegen, zonder het water waarin
-ze geweekt zijn af te gieten, De helft van
-de uien bakken met laurier en Kruicdnagel.
-Alle uien, kerrie en gout bij de linzen
-voegen, Alles aan de kook brengen,. Van de
-bloem met boter en melk een papje maken en
-verder afmaken met de soep, Als de linzen
-gaar Zijn is de soep klaar.
-

tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/2400dpi/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -0,0 +1,25 @@
+é la Waterman
+
+h ens linzen
+
+3 liter water
+
+3 mien
+
+bloem} hater
+
+2 kopjes melk
+
+laurier, kruidnagel, kerrie, gout
+
+Be linzen wassen en in¢l liter kokend wa-
+ter 1 dag laten weken. 2 liter water bi'
+de linzen voegen, gender hat water wagrin
+ze geweekt zijn af te gieten. De helft Van
+de mien bakken met laurier en kruiénagal.
+Alla mien, kerrie en zout bij de linzen
+voegen. A1135 aan de Rook brengem. Van de
+bloem mat boter-en melk een papje maken en
+verder afmaken met de soap. A13 de linzen
+gaar zijn is de seep klaar.
+

tests/cache/aspect/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -1,4 +0,0 @@
- 
-
-This should be a perfect circle:
-

tests/cache/aspect/__-l__eng__000001.image__000001.rendered__pdf__txt/stderr.bin

@@ -1,3 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
-Too few characters. Skipping this page
-OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__000001.image__000001.rendered__pdf__txt/txt.bin

@@ -1,4 +0,0 @@
- 
-
-This should be a perfect circle:
-

tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -1,3 +1,3 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica
 Too few characters. Skipping this page
 OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -0,0 +1,4 @@
+ 
+
+This shnuld be a perfect circle-
+

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/hocr.bin

@@ -5,11 +5,11 @@
  <head>
   <title></title>
 <meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
-  <meta name='ocr-system' content='tesseract 4.0.0-beta.1' />
+  <meta name='ocr-system' content='tesseract 3.05.01' />
   <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
 </head>
 <body>
-  <div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.4nwu81i8/000001.ocr.png"; bbox 0 0 1000 800; ppageno 0'>
+  <div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.7vjmeixk/000001.ocr.png"; bbox 0 0 1000 800; ppageno 0'>
    <div class='ocr_carea' id='block_1_1' title="bbox 296 96 704 504">
     <p class='ocr_par' id='par_1_1' lang='eng' title="bbox 296 96 704 504">
      <span class='ocr_line' id='line_1_1' title="bbox 296 96 704 504; baseline 0 296; x_size 169.33333; x_descenders 42.333332; x_ascenders 42.333336"><span class='ocrx_word' id='word_1_1' title='bbox 296 96 704 504; x_wconf 95'><strong> </strong></span> 
@@ -18,7 +18,7 @@
    </div>
    <div class='ocr_carea' id='block_1_2' title="bbox 150 592 841 622">
     <p class='ocr_par' id='par_1_2' lang='eng' title="bbox 150 592 841 622">
-     <span class='ocr_line' id='line_1_2' title="bbox 150 592 841 622; baseline 0 -6; x_size 30; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_2' title='bbox 150 592 230 616; x_wconf 96'><strong>This</strong></span> <span class='ocrx_word' id='word_1_3' title='bbox 260 592 384 616; x_wconf 95'><strong>should</strong></span> <span class='ocrx_word' id='word_1_4' title='bbox 413 592 449 616; x_wconf 95'><strong>be</strong></span> <span class='ocrx_word' id='word_1_5' title='bbox 479 600 493 616; x_wconf 95'><strong>a</strong></span> <span class='ocrx_word' id='word_1_6' title='bbox 523 592 668 622; x_wconf 95'><strong>perfect</strong></span> <span class='ocrx_word' id='word_1_7' title='bbox 698 592 841 616; x_wconf 55'><strong>circle:</strong></span> 
+     <span class='ocr_line' id='line_1_2' title="bbox 150 592 841 622; baseline 0 -6; x_size 30; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_2' title='bbox 150 592 230 616; x_wconf 85'><strong>This</strong></span> <span class='ocrx_word' id='word_1_3' title='bbox 260 592 384 616; x_wconf 81'>shnuld</span> <span class='ocrx_word' id='word_1_4' title='bbox 413 592 449 616; x_wconf 81'><strong>be</strong></span> <span class='ocrx_word' id='word_1_5' title='bbox 479 600 493 616; x_wconf 85'><em>a</em></span> <span class='ocrx_word' id='word_1_6' title='bbox 523 592 668 622; x_wconf 76'>perfect</span> <span class='ocrx_word' id='word_1_7' title='bbox 698 592 841 616; x_wconf 76'><strong>circle-</strong></span> 
      </span>
     </p>
    </div>

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/stderr.bin

@@ -1,3 +1,3 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica
 Too few characters. Skipping this page
 OSD: Weak margin (0.00) for 24 blob text block, but using orientation anyway: 0

tests/cache/aspect/__-l__eng__000001.ocr.png__000001__hocr__txt/txt.bin

@@ -1,4 +1,4 @@
  
 
-This should be a perfect circle:
-
+This shnuld be a perfect circle-
+

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -1 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000002.ocr.png__000002.text__pdf__txt/stderr.bin

@@ -1 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000003.ocr.png__000003.text__pdf__txt/stderr.bin

@@ -1 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000003.ocr.png__000003.text__pdf__txt/txt.bin

@@ -1,123 +0,0 @@
-The LinnSequencer
-32 Track MIDI Sequence Recorder
-
-The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
-
-extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
-
-¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
-FORWARD, REWIND, and LOCATE controls.
-
-e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
-be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
-
-synthesizers!
-
-¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
-
-per disk!
-
-¢ One or all tracks may be TRANSPOSED at the touch of a key.
-e Exclusive real-time ERASE function makes editing FAST.
-* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
-
-rhythmic value.
-
-¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
-
-¢ Optional SMPTE time code synchronization.
-
-© Optional remote control.
-
-Recording a Sequence
-
-To record a sequence, simply press RECORD and PLAY,
-then play your MIDI keyboard in time to the Sequencer’s
-click track. When the sequence loops back around to bar 1,
-you’ ll hear what you played—only all timing errors will be
-
-corrected! (Timing correction may be adjusted or defeated).
-
-Any additional notes played will be added into the track
-— existing notes are not erased while recording!
-
-FAST FORWARD, REWIND, and LOCATE controls
-may be used at any time to quickly access any location in
-your sequence for spot-recording. To overdub a new part,
-select a different track and start recording—while you
-record, the first track will play in perfect sync (unless you
-MUTE it, or SOLO another track). In this way, up to 32
-tracks may be overdubbed! All MIDI effects are recorded
-including pitch bend, modulation, velocity, aftertouch,
-sustain pedal, and program changes!
-
-Editing
-
-To erase a wrong note, simply hold ERASE and press
-the note to be erased just before it plays in the sequence—
-when played back, it will be gone. Notes may also be
-
-added, erased, or changed using the SINGLE STEP func-
-tion. To overdub notes at specific points within a sequence,
-
-Additional Features
-
-simply use LOCATE, FAST FORWARD, or REWIND to
-find the desired bar number, then start recording.
-
-The INSERT/COPY function allows you to move bars
-from one location to another—in the same sequence or a
-different one. For example, you might insert a copy of the
-first verse between the second chorus and the bridge.
-DELETE BARS operates the same way to remove
-unwanted sections,
-
-Creating a Song
-
-One way to create a song is to record each track all the
-way through (up to 999 bars). Another way is to record
-each basic section (verse, chorus, etc.) in individual
-sequences, then use the CREATE SONG function to “chain”
-them together. CREATE SONG will then automatically
-copy all the parts into a new sequence. If desired, you can
-even set the last few bars to repeat infinitely, for a fadeout.
-
-Composition Without Compromise
-
-The technology you use should never be so complex that
-it interferes with the creative process. That’s precisely why
-the LinnSequencer is designed to let you compose, record
-and edit while devoting your undivided attention to your
-music. See your Linn dealer today for a demonstration!
-
-* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
-
-HELP button displays additional explanations.
-
-* Non-destructive recording—existing notes are not erased while recording.
-¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
-
-ERASE, REPEAT, PLAY/STOP, or LOCATE.
-
-¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
-
-© Will sync to standard LinnDrum or Linn 9000 sync tone.
-
-© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
-* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
-
-(even drop frame!)
-
-¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
-
-on the TAP TEMPO button.
-
-¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
-¢ Any TIME SIGNATURE may be used, and may be changed within a song.
-
-linn
-Linn Electronics, Inc.
-
-18720 Oxnard Street, Tarzana, CA 91356
-(818) 708-8131 TELEX #298949 LINN UR
-

tests/cache/cardinal/__-l__eng__-c__textonly_pdf=1__000004.ocr.png__000004.text__pdf__txt/stderr.bin

@@ -1 +0,0 @@
-Tesseract Open Source OCR Engine v4.0.0-beta.1 with Leptonica

tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/cardinal/__-l__eng__000001.ocr.png__000001.text__pdf__txt/txt.bin

@@ -1,51 +1,51 @@
 The LinnSequencer
 32 Track MIDI Sequence Recorder
 
-The LinnSequencer is a state-of-the-art composition and performance tool for the professional musician. It is
+The LinnSequencer is a state—of—the-art composition and performance tool for the professional musician. It is
 
 extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
 
-¢ Operation is similar to multi-track tape recorder with PLAY, STOP, RECORD, FAST
+0 Operation is similar to multi—track tape recorder with PLAY, STOP, RECORD, FAST
 FORWARD, REWIND, and LOCATE controls.
 
-e Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+0 Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
 be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
 
-synthesizers!
+synthesizers !
 
-¢ Ultra-fast 3%” disk drive stores complex songs in seconds and holds over 110,000 notes
+0 Ultra—fast 3 1/2 ” disk drive stores complex songs in seconds and holds over 110,000 notes
 
 per disk!
 
-¢ One or all tracks may be TRANSPOSED at the touch of a key.
-e Exclusive real-time ERASE function makes editing FAST.
-* Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+0 One or all tracks may be TRANSPOSED at the touch of a key.
+0 Exclusive real—time ERASE function makes editing FAST.
+0 Exclusive REPEAT function automatically repeats any held notes at a pre-selected
 
 rhythmic value.
 
-¢ TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+0 TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
 
-¢ Optional SMPTE time code synchronization.
+0 Optional SMPTE time code synchronization.
 
-© Optional remote control.
+0 Optional remote control.
 
-Recording a Sequence
+Recording 3 Sequence
 
 To record a sequence, simply press RECORD and PLAY,
 then play your MIDI keyboard in time to the Sequencer’s
 click track. When the sequence loops back around to bar 1,
-you’ ll hear what you played—only all timing errors will be
+you’ll hear what you played—only all timing errors will be
 
 corrected! (Timing correction may be adjusted or defeated).
 
 Any additional notes played will be added into the track
-— existing notes are not erased while recording!
+—existing notes are not erased while recording!
 
 FAST FORWARD, REWIND, and LOCATE controls
 may be used at any time to quickly access any location in
-your sequence for spot-recording. To overdub a new part,
-select a different track and start recording—while you
-record, the first track will play in perfect sync (unless you
+your sequence for spot—recording. To overdub a new part,
+select a different track and start recording~while you
+record, the first‘track will play in perfect sync (unless you
 MUTE it, or SOLO another track). In this way, up to 32
 tracks may be overdubbed! All MIDI effects are recorded
 including pitch bend, modulation, velocity, aftertouch,
@@ -58,7 +58,7 @@ the note to be erased just before it plays in the sequence—
 when played back, it will be gone. Notes may also be
 
 added, erased, or changed using the SINGLE STEP func-
-tion. To overdub notes at specific points within a sequence,
+tion. To overdub notes at specific points within a sequence,
 
 Additional Features
 
@@ -70,7 +70,7 @@ from one location to another—in the same sequence or a
 different one. For example, you might insert a copy of the
 first verse between the second chorus and the bridge.
 DELETE BARS operates the same way to remove
-unwanted sections,
+unwanted sections.
 
 Creating a Song
 
@@ -90,34 +90,34 @@ the LinnSequencer is designed to let you compose, record
 and edit while devoting your undivided attention to your
 music. See your Linn dealer today for a demonstration!
 
-* Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+0 Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
 
 HELP button displays additional explanations.
 
-* Non-destructive recording—existing notes are not erased while recording.
-¢ Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+0 Non-destructive recording—existing notes are not erased while recording.
+0 Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
 
 ERASE, REPEAT, PLAY/STOP, or LOCATE.
 
-¢ Iwo TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+0 Two TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
 
-© Will sync to standard LinnDrum or Linn 9000 sync tone.
+0 Will sync to standard LinnDrum or Linn 9000 sync tone.
 
-© Utilizes ultra high-speed, 8 MHz 80186 16 bit computer internally for FAST operation.
-* TEMPO may be specified in BEATS-PER-MINUTE or FRAMES-PER-BEAT at 24, 25, or 30 frames per second,
+' Utilizes ultra high—speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+0 TEMPO may be specified in BEATS-PER—MINUTE or FRAMES-PER—BEAT at 24, 25, or 30 frames per second,
 
 (even drop frame!)
 
-¢ TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+0 TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
 
 on the TAP TEMPO button.
 
-¢ TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
-¢ Any TIME SIGNATURE may be used, and may be changed within a song.
+0 TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+° Any TIME SIGNATURE may be used, and may be changed within a song.
 
-linn
+E1111
 Linn Electronics, Inc.
 
 18720 Oxnard Street, Tarzana, CA 91356
-(818) 708-8131 TELEX #298949 LINN UR
-
+(818) 7088131 T ELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/hocr.bin

@@ -0,0 +1,335 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ <head>
+  <title></title>
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
+  <meta name='ocr-system' content='tesseract 3.05.01' />
+  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
+</head>
+<body>
+  <div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.wahjk6fa/000001.ocr.png"; bbox 0 0 2550 3300; ppageno 0'>
+   <div class='ocr_carea' id='block_1_1' title="bbox 582 131 1968 303">
+    <p class='ocr_par' id='par_1_1' lang='eng' title="bbox 582 131 1968 303">
+     <span class='ocr_line' id='line_1_1' title="bbox 882 131 1657 217; baseline 0.001 -17; x_size 85; x_descenders 16; x_ascenders 19"><span class='ocrx_word' id='word_1_1' title='bbox 882 132 1036 202; x_wconf 86'>The</span> <span class='ocrx_word' id='word_1_2' title='bbox 1061 131 1657 217; x_wconf 84'>LinnSequencer</span> 
+     </span>
+     <span class='ocr_line' id='line_1_2' title="bbox 582 215 1968 303; baseline 0 -17; x_size 87; x_descenders 16; x_ascenders 21"><span class='ocrx_word' id='word_1_3' title='bbox 582 215 674 286; x_wconf 90'>32</span> <span class='ocrx_word' id='word_1_4' title='bbox 697 218 923 288; x_wconf 84'>Track</span> <span class='ocrx_word' id='word_1_5' title='bbox 948 218 1181 287; x_wconf 94'><strong>MIDI</strong></span> <span class='ocrx_word' id='word_1_6' title='bbox 1208 217 1575 303; x_wconf 89'>Sequence</span> <span class='ocrx_word' id='word_1_7' title='bbox 1600 218 1968 288; x_wconf 83'>Recorder</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_2' title="bbox 347 380 2188 574">
+    <p class='ocr_par' id='par_1_2' lang='eng' title="bbox 347 380 2188 423">
+     <span class='ocr_line' id='line_1_3' title="bbox 347 380 2188 423; baseline -0.001 -12; x_size 38; x_descenders 8; x_ascenders 10"><span class='ocrx_word' id='word_1_8' title='bbox 347 380 412 410; x_wconf 86'>The</span> <span class='ocrx_word' id='word_1_9' title='bbox 424 380 676 417; x_wconf 85'>LinnSequencer</span> <span class='ocrx_word' id='word_1_10' title='bbox 688 380 712 411; x_wconf 87'>is</span> <span class='ocrx_word' id='word_1_11' title='bbox 724 390 743 411; x_wconf 90'>a</span> <span class='ocrx_word' id='word_1_12' title='bbox 754 381 1005 423; x_wconf 79'>state—of—the-art</span> <span class='ocrx_word' id='word_1_13' title='bbox 1017 380 1226 418; x_wconf 86'>composition</span> <span class='ocrx_word' id='word_1_14' title='bbox 1238 381 1299 411; x_wconf 88'>and</span> <span class='ocrx_word' id='word_1_15' title='bbox 1311 380 1525 418; x_wconf 85'>performance</span> <span class='ocrx_word' id='word_1_16' title='bbox 1536 380 1602 411; x_wconf 85'>tool</span> <span class='ocrx_word' id='word_1_17' title='bbox 1615 380 1663 411; x_wconf 82'>for</span> <span class='ocrx_word' id='word_1_18' title='bbox 1674 381 1725 410; x_wconf 94'>the</span> <span class='ocrx_word' id='word_1_19' title='bbox 1737 380 1940 417; x_wconf 85'>professional</span> <span class='ocrx_word' id='word_1_20' title='bbox 1952 380 2112 411; x_wconf 87'>musician.</span> <span class='ocrx_word' id='word_1_21' title='bbox 2127 381 2152 410; x_wconf 90'>It</span> <span class='ocrx_word' id='word_1_22' title='bbox 2164 380 2188 410; x_wconf 90'>is</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_3' lang='eng' title="bbox 347 430 1988 468">
+     <span class='ocr_line' id='line_1_4' title="bbox 347 430 1988 468; baseline 0 -8; x_size 37; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_23' title='bbox 347 430 507 467; x_wconf 86'>extremely</span> <span class='ocrx_word' id='word_1_24' title='bbox 518 430 677 467; x_wconf 84'>powerful,</span> <span class='ocrx_word' id='word_1_25' title='bbox 691 435 739 467; x_wconf 91'>yet</span> <span class='ocrx_word' id='word_1_26' title='bbox 752 430 924 468; x_wconf 87'>amazingly</span> <span class='ocrx_word' id='word_1_27' title='bbox 936 430 1044 468; x_wconf 85'>simple</span> <span class='ocrx_word' id='word_1_28' title='bbox 1055 435 1087 461; x_wconf 87'>to</span> <span class='ocrx_word' id='word_1_29' title='bbox 1099 431 1183 461; x_wconf 88'>learn</span> <span class='ocrx_word' id='word_1_30' title='bbox 1195 431 1257 461; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_31' title='bbox 1269 440 1329 461; x_wconf 89'>use.</span> <span class='ocrx_word' id='word_1_32' title='bbox 1344 431 1393 461; x_wconf 87'>It’s</span> <span class='ocrx_word' id='word_1_33' title='bbox 1406 440 1499 467; x_wconf 91'>many</span> <span class='ocrx_word' id='word_1_34' title='bbox 1510 430 1701 461; x_wconf 88'>remarkable</span> <span class='ocrx_word' id='word_1_35' title='bbox 1714 430 1843 460; x_wconf 85'>features</span> <span class='ocrx_word' id='word_1_36' title='bbox 1856 430 1988 461; x_wconf 87'>include:</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_4' lang='eng' title="bbox 350 482 2093 574">
+     <span class='ocr_line' id='line_1_5' title="bbox 350 482 2093 527; baseline 0 -9; x_size 43; x_descenders 7; x_ascenders 11"><span class='ocrx_word' id='word_1_37' title='bbox 350 490 368 508; x_wconf 68'><strong>0</strong></span> <span class='ocrx_word' id='word_1_38' title='bbox 383 482 585 526; x_wconf 88'>Operation</span> <span class='ocrx_word' id='word_1_39' title='bbox 598 482 627 518; x_wconf 89'>is</span> <span class='ocrx_word' id='word_1_40' title='bbox 641 482 776 518; x_wconf 88'>similar</span> <span class='ocrx_word' id='word_1_41' title='bbox 789 488 829 518; x_wconf 88'>to</span> <span class='ocrx_word' id='word_1_42' title='bbox 843 482 1062 519; x_wconf 86'>multi—track</span> <span class='ocrx_word' id='word_1_43' title='bbox 1076 488 1160 527; x_wconf 92'>tape</span> <span class='ocrx_word' id='word_1_44' title='bbox 1173 483 1336 519; x_wconf 88'>recorder</span> <span class='ocrx_word' id='word_1_45' title='bbox 1350 482 1436 518; x_wconf 87'>with</span> <span class='ocrx_word' id='word_1_46' title='bbox 1451 483 1580 525; x_wconf 90'>PLAY,</span> <span class='ocrx_word' id='word_1_47' title='bbox 1598 483 1724 525; x_wconf 90'>STOP,</span> <span class='ocrx_word' id='word_1_48' title='bbox 1741 483 1957 525; x_wconf 89'>RECORD,</span> <span class='ocrx_word' id='word_1_49' title='bbox 1974 483 2093 518; x_wconf 88'>FAST</span> 
+     </span>
+     <span class='ocr_line' id='line_1_6' title="bbox 383 532 1345 574; baseline 0.001 -7; x_size 43; x_descenders 7; x_ascenders 11"><span class='ocrx_word' id='word_1_50' title='bbox 383 532 635 574; x_wconf 88'>FORWARD,</span> <span class='ocrx_word' id='word_1_51' title='bbox 652 532 865 574; x_wconf 89'>REWIND,</span> <span class='ocrx_word' id='word_1_52' title='bbox 882 532 956 568; x_wconf 91'>and</span> <span class='ocrx_word' id='word_1_53' title='bbox 971 532 1163 568; x_wconf 90'>LOCATE</span> <span class='ocrx_word' id='word_1_54' title='bbox 1177 532 1345 568; x_wconf 87'>controls.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_3' title="bbox 349 589 2136 685">
+    <p class='ocr_par' id='par_1_5' lang='eng' title="bbox 349 589 2136 685">
+     <span class='ocr_line' id='line_1_7' title="bbox 349 589 2136 634; baseline 0.001 -9; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_55' title='bbox 349 597 368 615; x_wconf 67'><strong>0</strong></span> <span class='ocrx_word' id='word_1_56' title='bbox 383 590 482 625; x_wconf 90'>Each</span> <span class='ocrx_word' id='word_1_57' title='bbox 496 589 539 625; x_wconf 87'>of</span> <span class='ocrx_word' id='word_1_58' title='bbox 549 590 611 626; x_wconf 92'>the</span> <span class='ocrx_word' id='word_1_59' title='bbox 625 590 686 625; x_wconf 91'>100</span> <span class='ocrx_word' id='word_1_60' title='bbox 700 601 895 634; x_wconf 89'>sequences</span> <span class='ocrx_word' id='word_1_61' title='bbox 909 590 1075 626; x_wconf 88'>contains</span> <span class='ocrx_word' id='word_1_62' title='bbox 1090 591 1135 626; x_wconf 89'>32</span> <span class='ocrx_word' id='word_1_63' title='bbox 1149 590 1418 633; x_wconf 87'>simultaneous,</span> <span class='ocrx_word' id='word_1_64' title='bbox 1435 590 1661 634; x_wconf 89'>polyphonic</span> <span class='ocrx_word' id='word_1_65' title='bbox 1674 590 1802 626; x_wconf 85'>tracks.</span> <span class='ocrx_word' id='word_1_66' title='bbox 1819 590 1921 626; x_wconf 89'>Each</span> <span class='ocrx_word' id='word_1_67' title='bbox 1934 590 2035 626; x_wconf 88'>track</span> <span class='ocrx_word' id='word_1_68' title='bbox 2050 600 2136 634; x_wconf 89'>may</span> 
+     </span>
+     <span class='ocr_line' id='line_1_8' title="bbox 383 639 2022 685; baseline 0.001 -10; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_69' title='bbox 383 639 428 675; x_wconf 95'>be</span> <span class='ocrx_word' id='word_1_70' title='bbox 442 639 607 684; x_wconf 88'>assigned</span> <span class='ocrx_word' id='word_1_71' title='bbox 621 645 659 676; x_wconf 92'>to</span> <span class='ocrx_word' id='word_1_72' title='bbox 674 650 745 676; x_wconf 91'>one</span> <span class='ocrx_word' id='word_1_73' title='bbox 759 640 801 676; x_wconf 84'>of</span> <span class='ocrx_word' id='word_1_74' title='bbox 813 641 851 676; x_wconf 92'>16</span> <span class='ocrx_word' id='word_1_75' title='bbox 864 641 980 675; x_wconf 90'>MIDI</span> <span class='ocrx_word' id='word_1_76' title='bbox 996 640 1176 676; x_wconf 88'>channels.</span> <span class='ocrx_word' id='word_1_77' title='bbox 1194 640 1498 684; x_wconf 90'>Simultaneously</span> <span class='ocrx_word' id='word_1_78' title='bbox 1510 640 1609 685; x_wconf 90'>plays</span> <span class='ocrx_word' id='word_1_79' title='bbox 1624 651 1674 684; x_wconf 91'>up</span> <span class='ocrx_word' id='word_1_80' title='bbox 1688 645 1727 676; x_wconf 90'>to</span> <span class='ocrx_word' id='word_1_81' title='bbox 1742 641 1780 676; x_wconf 94'>16</span> <span class='ocrx_word' id='word_1_82' title='bbox 1794 640 2022 684; x_wconf 87'>polyphonic</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_4' title="bbox 384 689 632 734">
+    <p class='ocr_par' id='par_1_6' lang='eng' title="bbox 384 689 632 734">
+     <span class='ocr_line' id='line_1_9' title="bbox 384 689 632 734; baseline 0.004 -9; x_size 46; x_descenders 9; x_ascenders 12"><span class='ocrx_word' id='word_1_83' title='bbox 384 689 618 734; x_wconf 84'>synthesizers</span> <span class='ocrx_word' id='word_1_84' title='bbox 625 690 632 726; x_wconf 85'><strong>!</strong></span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_5' title="bbox 349 748 2117 793">
+    <p class='ocr_par' id='par_1_7' lang='eng' title="bbox 349 748 2117 793">
+     <span class='ocr_line' id='line_1_10' title="bbox 349 748 2117 793; baseline 0 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_85' title='bbox 349 755 367 774; x_wconf 66'><strong>0</strong></span> <span class='ocrx_word' id='word_1_86' title='bbox 383 748 573 784; x_wconf 87'>Ultra—fast</span> <span class='ocrx_word' id='word_1_87' title='bbox 588 749 606 784; x_wconf 90'>3</span> <span class='ocrx_word' id='word_1_88' title='bbox 616 749 652 784; x_wconf 86'>1/2</span> <span class='ocrx_word' id='word_1_89' title='bbox 663 749 677 764; x_wconf 87'><em>”</em></span> <span class='ocrx_word' id='word_1_90' title='bbox 694 748 775 784; x_wconf 91'>disk</span> <span class='ocrx_word' id='word_1_91' title='bbox 790 748 887 784; x_wconf 89'>drive</span> <span class='ocrx_word' id='word_1_92' title='bbox 901 754 1012 785; x_wconf 85'>stores</span> <span class='ocrx_word' id='word_1_93' title='bbox 1027 748 1194 793; x_wconf 87'>complex</span> <span class='ocrx_word' id='word_1_94' title='bbox 1208 760 1316 793; x_wconf 87'>songs</span> <span class='ocrx_word' id='word_1_95' title='bbox 1330 748 1368 784; x_wconf 87'>in</span> <span class='ocrx_word' id='word_1_96' title='bbox 1382 748 1535 785; x_wconf 87'>seconds</span> <span class='ocrx_word' id='word_1_97' title='bbox 1550 748 1624 784; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_98' title='bbox 1638 748 1746 784; x_wconf 87'>holds</span> <span class='ocrx_word' id='word_1_99' title='bbox 1761 759 1844 784; x_wconf 85'>over</span> <span class='ocrx_word' id='word_1_100' title='bbox 1859 749 2000 791; x_wconf 88'>110,000</span> <span class='ocrx_word' id='word_1_101' title='bbox 2013 753 2117 784; x_wconf 89'>notes</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_6' title="bbox 383 797 554 842">
+    <p class='ocr_par' id='par_1_8' lang='eng' title="bbox 383 797 554 842">
+     <span class='ocr_line' id='line_1_11' title="bbox 383 797 554 842; baseline 0 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_102' title='bbox 383 808 446 842; x_wconf 82'>per</span> <span class='ocrx_word' id='word_1_103' title='bbox 459 797 554 834; x_wconf 86'>disk!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_7' title="bbox 349 855 2030 1016">
+    <p class='ocr_par' id='par_1_9' lang='eng' title="bbox 349 855 2030 1016">
+     <span class='ocr_line' id='line_1_12' title="bbox 350 855 1638 900; baseline 0.001 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_104' title='bbox 350 863 367 881; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_105' title='bbox 383 856 464 891; x_wconf 90'>One</span> <span class='ocrx_word' id='word_1_106' title='bbox 478 866 520 891; x_wconf 87'>or</span> <span class='ocrx_word' id='word_1_107' title='bbox 534 855 580 891; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_108' title='bbox 594 856 712 892; x_wconf 89'>tracks</span> <span class='ocrx_word' id='word_1_109' title='bbox 726 867 811 900; x_wconf 90'>may</span> <span class='ocrx_word' id='word_1_110' title='bbox 823 856 869 892; x_wconf 88'>be</span> <span class='ocrx_word' id='word_1_111' title='bbox 882 856 1212 892; x_wconf 90'>TRANSPOSED</span> <span class='ocrx_word' id='word_1_112' title='bbox 1227 861 1264 892; x_wconf 90'>at</span> <span class='ocrx_word' id='word_1_113' title='bbox 1277 856 1338 892; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_114' title='bbox 1351 856 1463 892; x_wconf 90'>touch</span> <span class='ocrx_word' id='word_1_115' title='bbox 1477 856 1520 892; x_wconf 89'>of</span> <span class='ocrx_word' id='word_1_116' title='bbox 1531 867 1554 892; x_wconf 90'>a</span> <span class='ocrx_word' id='word_1_117' title='bbox 1568 856 1638 900; x_wconf 89'>key.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_13' title="bbox 350 913 1535 958; baseline 0.001 -9; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_118' title='bbox 350 921 367 939; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_119' title='bbox 383 913 568 950; x_wconf 84'>Exclusive</span> <span class='ocrx_word' id='word_1_120' title='bbox 581 913 756 950; x_wconf 87'>real—time</span> <span class='ocrx_word' id='word_1_121' title='bbox 769 914 929 950; x_wconf 91'>ERASE</span> <span class='ocrx_word' id='word_1_122' title='bbox 944 914 1114 950; x_wconf 87'>function</span> <span class='ocrx_word' id='word_1_123' title='bbox 1128 914 1251 950; x_wconf 88'>makes</span> <span class='ocrx_word' id='word_1_124' title='bbox 1266 914 1400 958; x_wconf 87'>editing</span> <span class='ocrx_word' id='word_1_125' title='bbox 1414 915 1535 950; x_wconf 91'>FAST.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_14' title="bbox 349 971 2030 1016; baseline 0.001 -10; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_126' title='bbox 349 979 367 997; x_wconf 67'><strong>0</strong></span> <span class='ocrx_word' id='word_1_127' title='bbox 382 971 568 1007; x_wconf 83'>Exclusive</span> <span class='ocrx_word' id='word_1_128' title='bbox 582 972 773 1007; x_wconf 87'>REPEAT</span> <span class='ocrx_word' id='word_1_129' title='bbox 787 972 958 1008; x_wconf 86'>function</span> <span class='ocrx_word' id='word_1_130' title='bbox 971 972 1245 1016; x_wconf 89'>automatically</span> <span class='ocrx_word' id='word_1_131' title='bbox 1258 977 1396 1016; x_wconf 90'>repeats</span> <span class='ocrx_word' id='word_1_132' title='bbox 1410 983 1481 1016; x_wconf 88'>any</span> <span class='ocrx_word' id='word_1_133' title='bbox 1493 972 1578 1008; x_wconf 91'>held</span> <span class='ocrx_word' id='word_1_134' title='bbox 1592 977 1694 1008; x_wconf 91'>notes</span> <span class='ocrx_word' id='word_1_135' title='bbox 1709 977 1746 1008; x_wconf 85'>at</span> <span class='ocrx_word' id='word_1_136' title='bbox 1760 982 1783 1008; x_wconf 89'>a</span> <span class='ocrx_word' id='word_1_137' title='bbox 1796 972 2030 1016; x_wconf 87'>pre-selected</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_8' title="bbox 382 1021 689 1065">
+    <p class='ocr_par' id='par_1_10' lang='eng' title="bbox 382 1021 689 1065">
+     <span class='ocr_line' id='line_1_15' title="bbox 382 1021 689 1065; baseline 0.003 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_138' title='bbox 382 1021 564 1065; x_wconf 87'>rhythmic</span> <span class='ocrx_word' id='word_1_139' title='bbox 577 1021 689 1058; x_wconf 86'>value.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_9' title="bbox 349 1080 2174 1125">
+    <p class='ocr_par' id='par_1_11' lang='eng' title="bbox 349 1080 2174 1125">
+     <span class='ocr_line' id='line_1_16' title="bbox 349 1080 2174 1125; baseline 0.001 -11; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_140' title='bbox 349 1087 367 1105; x_wconf 73'><strong>0</strong></span> <span class='ocrx_word' id='word_1_141' title='bbox 382 1080 567 1115; x_wconf 87'>TIMING</span> <span class='ocrx_word' id='word_1_142' title='bbox 582 1080 908 1116; x_wconf 90'>CORRECTION</span> <span class='ocrx_word' id='word_1_143' title='bbox 921 1080 1041 1116; x_wconf 90'>works</span> <span class='ocrx_word' id='word_1_144' title='bbox 1056 1080 1186 1124; x_wconf 86'>during</span> <span class='ocrx_word' id='word_1_145' title='bbox 1199 1080 1378 1124; x_wconf 89'>playback</span> <span class='ocrx_word' id='word_1_146' title='bbox 1392 1080 1466 1116; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_147' title='bbox 1480 1085 1644 1124; x_wconf 87'>operates</span> <span class='ocrx_word' id='word_1_148' title='bbox 1658 1080 1814 1116; x_wconf 91'>without</span> <span class='ocrx_word' id='word_1_149' title='bbox 1831 1080 2044 1125; x_wconf 87'>‘chopping’</span> <span class='ocrx_word' id='word_1_150' title='bbox 2061 1085 2174 1116; x_wconf 88'>notes.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_10' title="bbox 349 1137 1287 1182">
+    <p class='ocr_par' id='par_1_12' lang='eng' title="bbox 349 1137 1287 1182">
+     <span class='ocr_line' id='line_1_17' title="bbox 349 1137 1287 1182; baseline 0.001 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_151' title='bbox 349 1145 367 1163; x_wconf 73'><strong>0</strong></span> <span class='ocrx_word' id='word_1_152' title='bbox 382 1137 560 1182; x_wconf 85'>Optional</span> <span class='ocrx_word' id='word_1_153' title='bbox 575 1138 739 1174; x_wconf 89'>SMPTE</span> <span class='ocrx_word' id='word_1_154' title='bbox 752 1138 839 1174; x_wconf 89'>time</span> <span class='ocrx_word' id='word_1_155' title='bbox 853 1138 945 1174; x_wconf 89'>code</span> <span class='ocrx_word' id='word_1_156' title='bbox 959 1138 1287 1182; x_wconf 88'>synchronization.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_11' title="bbox 349 1195 874 1240">
+    <p class='ocr_par' id='par_1_13' lang='eng' title="bbox 349 1195 874 1240">
+     <span class='ocr_line' id='line_1_18' title="bbox 349 1195 874 1240; baseline 0 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_157' title='bbox 349 1203 367 1222; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_158' title='bbox 382 1195 560 1240; x_wconf 87'>Optional</span> <span class='ocrx_word' id='word_1_159' title='bbox 573 1201 709 1233; x_wconf 86'>remote</span> <span class='ocrx_word' id='word_1_160' title='bbox 723 1196 874 1233; x_wconf 87'>control.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_12' title="bbox 346 1288 1239 1491">
+    <p class='ocr_par' id='par_1_14' lang='eng' title="bbox 346 1288 749 1329">
+     <span class='ocr_line' id='line_1_19' title="bbox 346 1288 749 1329; baseline 0.002 -9; x_size 42; x_descenders 9; x_ascenders 10"><span class='ocrx_word' id='word_1_161' title='bbox 346 1288 535 1329; x_wconf 82'>Recording</span> <span class='ocrx_word' id='word_1_162' title='bbox 547 1298 567 1321; x_wconf 73'>3</span> <span class='ocrx_word' id='word_1_163' title='bbox 579 1288 749 1328; x_wconf 86'>Sequence</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_15' lang='eng' title="bbox 346 1339 1239 1491">
+     <span class='ocr_line' id='line_1_20' title="bbox 383 1339 1239 1373; baseline 0 -7; x_size 33; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_164' title='bbox 383 1340 420 1366; x_wconf 87'>To</span> <span class='ocrx_word' id='word_1_165' title='bbox 430 1339 524 1366; x_wconf 85'>record</span> <span class='ocrx_word' id='word_1_166' title='bbox 535 1347 551 1366; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_167' title='bbox 562 1347 704 1373; x_wconf 80'>sequence,</span> <span class='ocrx_word' id='word_1_168' title='bbox 716 1340 815 1373; x_wconf 91'>simply</span> <span class='ocrx_word' id='word_1_169' title='bbox 825 1348 898 1373; x_wconf 88'>press</span> <span class='ocrx_word' id='word_1_170' title='bbox 910 1340 1065 1367; x_wconf 90'>RECORD</span> <span class='ocrx_word' id='word_1_171' title='bbox 1075 1340 1131 1367; x_wconf 87'>and</span> <span class='ocrx_word' id='word_1_172' title='bbox 1142 1341 1239 1372; x_wconf 87'>PLAY,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_21' title="bbox 346 1378 1205 1412; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_173' title='bbox 346 1379 411 1406; x_wconf 91'>then</span> <span class='ocrx_word' id='word_1_174' title='bbox 422 1378 483 1412; x_wconf 89'>play</span> <span class='ocrx_word' id='word_1_175' title='bbox 493 1387 562 1412; x_wconf 87'>your</span> <span class='ocrx_word' id='word_1_176' title='bbox 572 1379 659 1405; x_wconf 89'>MIDI</span> <span class='ocrx_word' id='word_1_177' title='bbox 671 1379 810 1412; x_wconf 86'>keyboard</span> <span class='ocrx_word' id='word_1_178' title='bbox 821 1379 848 1406; x_wconf 90'>in</span> <span class='ocrx_word' id='word_1_179' title='bbox 858 1379 923 1406; x_wconf 90'>time</span> <span class='ocrx_word' id='word_1_180' title='bbox 934 1384 963 1406; x_wconf 90'>to</span> <span class='ocrx_word' id='word_1_181' title='bbox 974 1379 1019 1406; x_wconf 92'>the</span> <span class='ocrx_word' id='word_1_182' title='bbox 1030 1379 1205 1412; x_wconf 88'>Sequencer’s</span> 
+     </span>
+     <span class='ocr_line' id='line_1_22' title="bbox 347 1417 1232 1452; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_183' title='bbox 347 1417 414 1445; x_wconf 86'>click</span> <span class='ocrx_word' id='word_1_184' title='bbox 425 1418 506 1445; x_wconf 84'>track.</span> <span class='ocrx_word' id='word_1_185' title='bbox 519 1418 609 1445; x_wconf 85'>When</span> <span class='ocrx_word' id='word_1_186' title='bbox 620 1419 665 1445; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_187' title='bbox 676 1426 810 1452; x_wconf 86'>sequence</span> <span class='ocrx_word' id='word_1_188' title='bbox 821 1419 901 1452; x_wconf 88'>loops</span> <span class='ocrx_word' id='word_1_189' title='bbox 912 1419 983 1446; x_wconf 86'>back</span> <span class='ocrx_word' id='word_1_190' title='bbox 995 1419 1101 1446; x_wconf 90'>around</span> <span class='ocrx_word' id='word_1_191' title='bbox 1112 1423 1141 1446; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_192' title='bbox 1152 1419 1201 1446; x_wconf 87'>bar</span> <span class='ocrx_word' id='word_1_193' title='bbox 1213 1419 1232 1450; x_wconf 88'>1,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_23' title="bbox 346 1457 1223 1491; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_194' title='bbox 346 1457 430 1490; x_wconf 84'>you’ll</span> <span class='ocrx_word' id='word_1_195' title='bbox 441 1457 506 1485; x_wconf 93'>hear</span> <span class='ocrx_word' id='word_1_196' title='bbox 517 1458 590 1485; x_wconf 79'>what</span> <span class='ocrx_word' id='word_1_197' title='bbox 600 1466 654 1491; x_wconf 88'>you</span> <span class='ocrx_word' id='word_1_198' title='bbox 666 1458 865 1491; x_wconf 77'>played—only</span> <span class='ocrx_word' id='word_1_199' title='bbox 876 1458 910 1485; x_wconf 90'>all</span> <span class='ocrx_word' id='word_1_200' title='bbox 921 1458 1018 1491; x_wconf 84'>timing</span> <span class='ocrx_word' id='word_1_201' title='bbox 1029 1466 1115 1485; x_wconf 84'>errors</span> <span class='ocrx_word' id='word_1_202' title='bbox 1126 1458 1179 1485; x_wconf 91'>will</span> <span class='ocrx_word' id='word_1_203' title='bbox 1189 1458 1223 1485; x_wconf 91'>be</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_13' title="bbox 346 1497 1245 1531">
+    <p class='ocr_par' id='par_1_16' lang='eng' title="bbox 346 1497 1245 1531">
+     <span class='ocr_line' id='line_1_24' title="bbox 346 1497 1245 1531; baseline 0.001 -7; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_204' title='bbox 346 1497 494 1524; x_wconf 86'>corrected!</span> <span class='ocrx_word' id='word_1_205' title='bbox 508 1497 628 1530; x_wconf 85'>(Timing</span> <span class='ocrx_word' id='word_1_206' title='bbox 638 1497 791 1525; x_wconf 82'>correction</span> <span class='ocrx_word' id='word_1_207' title='bbox 802 1506 864 1531; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_208' title='bbox 875 1498 909 1525; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_209' title='bbox 920 1497 1047 1531; x_wconf 81'>adjusted</span> <span class='ocrx_word' id='word_1_210' title='bbox 1058 1505 1089 1525; x_wconf 92'>or</span> <span class='ocrx_word' id='word_1_211' title='bbox 1099 1497 1245 1531; x_wconf 87'>defeated).</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_14' title="bbox 346 1536 1205 1965">
+    <p class='ocr_par' id='par_1_17' lang='eng' title="bbox 346 1536 1186 1610">
+     <span class='ocr_line' id='line_1_25' title="bbox 346 1536 1186 1570; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_212' title='bbox 346 1537 409 1569; x_wconf 90'>Any</span> <span class='ocrx_word' id='word_1_213' title='bbox 419 1536 571 1563; x_wconf 86'>additional</span> <span class='ocrx_word' id='word_1_214' title='bbox 582 1541 659 1564; x_wconf 85'>notes</span> <span class='ocrx_word' id='word_1_215' title='bbox 670 1537 766 1570; x_wconf 87'>played</span> <span class='ocrx_word' id='word_1_216' title='bbox 777 1537 830 1563; x_wconf 88'>will</span> <span class='ocrx_word' id='word_1_217' title='bbox 840 1537 874 1564; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_218' title='bbox 885 1537 976 1564; x_wconf 88'>added</span> <span class='ocrx_word' id='word_1_219' title='bbox 986 1537 1045 1564; x_wconf 89'>into</span> <span class='ocrx_word' id='word_1_220' title='bbox 1055 1537 1101 1564; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_221' title='bbox 1111 1537 1186 1564; x_wconf 82'>track</span> 
+     </span>
+     <span class='ocr_line' id='line_1_26' title="bbox 347 1575 1052 1610; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_222' title='bbox 347 1575 495 1609; x_wconf 86'>—existing</span> <span class='ocrx_word' id='word_1_223' title='bbox 505 1580 582 1603; x_wconf 89'>notes</span> <span class='ocrx_word' id='word_1_224' title='bbox 593 1584 637 1603; x_wconf 90'>are</span> <span class='ocrx_word' id='word_1_225' title='bbox 648 1580 696 1603; x_wconf 88'>not</span> <span class='ocrx_word' id='word_1_226' title='bbox 707 1576 800 1603; x_wconf 90'>erased</span> <span class='ocrx_word' id='word_1_227' title='bbox 811 1576 890 1603; x_wconf 87'>while</span> <span class='ocrx_word' id='word_1_228' title='bbox 900 1576 1052 1610; x_wconf 87'>recording!</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_18' lang='eng' title="bbox 346 1616 1205 1965">
+     <span class='ocr_line' id='line_1_27' title="bbox 384 1616 1199 1648; baseline 0.001 -6; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_229' title='bbox 384 1616 471 1642; x_wconf 88'>FAST</span> <span class='ocrx_word' id='word_1_230' title='bbox 481 1616 671 1648; x_wconf 88'>FORWARD,</span> <span class='ocrx_word' id='word_1_231' title='bbox 684 1617 844 1648; x_wconf 87'>REWIND,</span> <span class='ocrx_word' id='word_1_232' title='bbox 857 1616 912 1643; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_233' title='bbox 924 1616 1068 1643; x_wconf 78'>LOCATE</span> <span class='ocrx_word' id='word_1_234' title='bbox 1079 1616 1199 1643; x_wconf 87'>controls</span> 
+     </span>
+     <span class='ocr_line' id='line_1_28' title="bbox 346 1655 1202 1689; baseline 0 -7; x_size 34; x_descenders 6; x_ascenders 9"><span class='ocrx_word' id='word_1_235' title='bbox 346 1663 409 1688; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_236' title='bbox 419 1655 453 1682; x_wconf 94'>be</span> <span class='ocrx_word' id='word_1_237' title='bbox 463 1655 530 1682; x_wconf 90'>used</span> <span class='ocrx_word' id='word_1_238' title='bbox 541 1659 569 1682; x_wconf 90'>at</span> <span class='ocrx_word' id='word_1_239' title='bbox 580 1663 632 1688; x_wconf 86'>any</span> <span class='ocrx_word' id='word_1_240' title='bbox 642 1655 707 1683; x_wconf 88'>time</span> <span class='ocrx_word' id='word_1_241' title='bbox 717 1659 746 1683; x_wconf 88'>to</span> <span class='ocrx_word' id='word_1_242' title='bbox 757 1655 867 1689; x_wconf 88'>quickly</span> <span class='ocrx_word' id='word_1_243' title='bbox 878 1664 968 1683; x_wconf 87'>access</span> <span class='ocrx_word' id='word_1_244' title='bbox 979 1664 1032 1689; x_wconf 87'>any</span> <span class='ocrx_word' id='word_1_245' title='bbox 1042 1655 1164 1683; x_wconf 90'>location</span> <span class='ocrx_word' id='word_1_246' title='bbox 1174 1655 1202 1682; x_wconf 94'>in</span> 
+     </span>
+     <span class='ocr_line' id='line_1_29' title="bbox 346 1694 1204 1728; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_247' title='bbox 346 1702 414 1727; x_wconf 88'>your</span> <span class='ocrx_word' id='word_1_248' title='bbox 424 1702 558 1727; x_wconf 87'>sequence</span> <span class='ocrx_word' id='word_1_249' title='bbox 570 1694 612 1721; x_wconf 84'>for</span> <span class='ocrx_word' id='word_1_250' title='bbox 623 1695 847 1728; x_wconf 84'>spot—recording.</span> <span class='ocrx_word' id='word_1_251' title='bbox 860 1696 897 1722; x_wconf 89'>To</span> <span class='ocrx_word' id='word_1_252' title='bbox 908 1695 1028 1722; x_wconf 82'>overdub</span> <span class='ocrx_word' id='word_1_253' title='bbox 1039 1703 1056 1722; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_254' title='bbox 1066 1703 1125 1722; x_wconf 89'>new</span> <span class='ocrx_word' id='word_1_255' title='bbox 1135 1699 1204 1728; x_wconf 88'>part,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_30' title="bbox 347 1733 1150 1768; baseline 0.001 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_256' title='bbox 347 1733 426 1761; x_wconf 88'>select</span> <span class='ocrx_word' id='word_1_257' title='bbox 437 1741 453 1761; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_258' title='bbox 463 1733 589 1761; x_wconf 86'>different</span> <span class='ocrx_word' id='word_1_259' title='bbox 599 1734 674 1761; x_wconf 85'>track</span> <span class='ocrx_word' id='word_1_260' title='bbox 685 1734 741 1761; x_wconf 87'>and</span> <span class='ocrx_word' id='word_1_261' title='bbox 752 1738 818 1761; x_wconf 83'>start</span> <span class='ocrx_word' id='word_1_262' title='bbox 828 1734 1086 1768; x_wconf 63'>recording~while</span> <span class='ocrx_word' id='word_1_263' title='bbox 1096 1742 1150 1767; x_wconf 89'>you</span> 
+     </span>
+     <span class='ocr_line' id='line_1_31' title="bbox 346 1773 1203 1808; baseline 0.001 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_264' title='bbox 346 1774 448 1806; x_wconf 86'>record,</span> <span class='ocrx_word' id='word_1_265' title='bbox 460 1774 506 1801; x_wconf 87'>the</span> <span class='ocrx_word' id='word_1_266' title='bbox 517 1773 662 1801; x_wconf 64'>first‘track</span> <span class='ocrx_word' id='word_1_267' title='bbox 673 1774 726 1801; x_wconf 91'>will</span> <span class='ocrx_word' id='word_1_268' title='bbox 736 1774 799 1807; x_wconf 89'>play</span> <span class='ocrx_word' id='word_1_269' title='bbox 809 1774 836 1801; x_wconf 91'>in</span> <span class='ocrx_word' id='word_1_270' title='bbox 847 1774 949 1808; x_wconf 84'>perfect</span> <span class='ocrx_word' id='word_1_271' title='bbox 961 1782 1026 1808; x_wconf 87'>sync</span> <span class='ocrx_word' id='word_1_272' title='bbox 1037 1774 1137 1808; x_wconf 89'>(unless</span> <span class='ocrx_word' id='word_1_273' title='bbox 1148 1782 1203 1807; x_wconf 93'>you</span> 
+     </span>
+     <span class='ocr_line' id='line_1_32' title="bbox 346 1813 1191 1847; baseline 0.002 -8; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_274' title='bbox 346 1813 454 1840; x_wconf 86'>MUTE</span> <span class='ocrx_word' id='word_1_275' title='bbox 464 1813 492 1845; x_wconf 88'>it,</span> <span class='ocrx_word' id='word_1_276' title='bbox 505 1821 537 1840; x_wconf 84'>or</span> <span class='ocrx_word' id='word_1_277' title='bbox 547 1813 642 1840; x_wconf 88'>SOLO</span> <span class='ocrx_word' id='word_1_278' title='bbox 653 1814 769 1841; x_wconf 87'>another</span> <span class='ocrx_word' id='word_1_279' title='bbox 779 1814 875 1847; x_wconf 86'>track).</span> <span class='ocrx_word' id='word_1_280' title='bbox 889 1814 920 1840; x_wconf 86'>In</span> <span class='ocrx_word' id='word_1_281' title='bbox 930 1813 984 1841; x_wconf 86'>this</span> <span class='ocrx_word' id='word_1_282' title='bbox 995 1822 1057 1847; x_wconf 87'>way,</span> <span class='ocrx_word' id='word_1_283' title='bbox 1069 1822 1106 1847; x_wconf 88'>up</span> <span class='ocrx_word' id='word_1_284' title='bbox 1117 1818 1146 1841; x_wconf 77'>to</span> <span class='ocrx_word' id='word_1_285' title='bbox 1158 1814 1191 1841; x_wconf 90'>32</span> 
+     </span>
+     <span class='ocr_line' id='line_1_33' title="bbox 346 1852 1205 1885; baseline 0.001 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_286' title='bbox 346 1852 434 1879; x_wconf 84'>tracks</span> <span class='ocrx_word' id='word_1_287' title='bbox 445 1860 508 1885; x_wconf 90'>may</span> <span class='ocrx_word' id='word_1_288' title='bbox 518 1853 552 1879; x_wconf 93'>be</span> <span class='ocrx_word' id='word_1_289' title='bbox 562 1853 748 1880; x_wconf 90'>overdubbed!</span> <span class='ocrx_word' id='word_1_290' title='bbox 761 1853 808 1880; x_wconf 92'>All</span> <span class='ocrx_word' id='word_1_291' title='bbox 819 1854 905 1880; x_wconf 81'>MIDI</span> <span class='ocrx_word' id='word_1_292' title='bbox 917 1853 1011 1880; x_wconf 81'>effects</span> <span class='ocrx_word' id='word_1_293' title='bbox 1022 1861 1067 1880; x_wconf 86'>are</span> <span class='ocrx_word' id='word_1_294' title='bbox 1076 1853 1205 1880; x_wconf 87'>recorded</span> 
+     </span>
+     <span class='ocr_line' id='line_1_34' title="bbox 346 1891 1168 1926; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_295' title='bbox 346 1891 485 1925; x_wconf 84'>including</span> <span class='ocrx_word' id='word_1_296' title='bbox 495 1891 570 1925; x_wconf 87'>pitch</span> <span class='ocrx_word' id='word_1_297' title='bbox 580 1892 663 1924; x_wconf 84'>bend,</span> <span class='ocrx_word' id='word_1_298' title='bbox 675 1892 859 1924; x_wconf 86'>modulation,</span> <span class='ocrx_word' id='word_1_299' title='bbox 872 1892 991 1926; x_wconf 83'>velocity,</span> <span class='ocrx_word' id='word_1_300' title='bbox 1004 1892 1168 1924; x_wconf 85'>aftertouch,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_35' title="bbox 346 1931 895 1965; baseline 0.002 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_301' title='bbox 346 1931 448 1958; x_wconf 88'>sustain</span> <span class='ocrx_word' id='word_1_302' title='bbox 459 1931 549 1965; x_wconf 88'>pedal,</span> <span class='ocrx_word' id='word_1_303' title='bbox 561 1932 617 1959; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_304' title='bbox 628 1940 756 1965; x_wconf 88'>program</span> <span class='ocrx_word' id='word_1_305' title='bbox 766 1932 895 1965; x_wconf 87'>changes!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_15' title="bbox 346 1998 1212 2163">
+    <p class='ocr_par' id='par_1_19' lang='eng' title="bbox 346 1998 479 2040">
+     <span class='ocr_line' id='line_1_36' title="bbox 346 1998 479 2040; baseline -0.008 -9; x_size 43; x_descenders 10; x_ascenders 11"><span class='ocrx_word' id='word_1_306' title='bbox 346 1998 479 2040; x_wconf 87'>Editing</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_20' lang='eng' title="bbox 346 2050 1212 2163">
+     <span class='ocr_line' id='line_1_37' title="bbox 383 2050 1186 2084; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_307' title='bbox 383 2050 419 2076; x_wconf 89'>To</span> <span class='ocrx_word' id='word_1_308' title='bbox 430 2057 503 2076; x_wconf 87'>erase</span> <span class='ocrx_word' id='word_1_309' title='bbox 514 2058 530 2077; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_310' title='bbox 540 2058 634 2083; x_wconf 88'>wrong</span> <span class='ocrx_word' id='word_1_311' title='bbox 644 2054 717 2082; x_wconf 84'>note,</span> <span class='ocrx_word' id='word_1_312' title='bbox 729 2050 829 2083; x_wconf 89'>simply</span> <span class='ocrx_word' id='word_1_313' title='bbox 839 2050 905 2077; x_wconf 88'>hold</span> <span class='ocrx_word' id='word_1_314' title='bbox 916 2051 1037 2077; x_wconf 85'>ERASE</span> <span class='ocrx_word' id='word_1_315' title='bbox 1048 2050 1103 2077; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_316' title='bbox 1113 2059 1186 2084; x_wconf 89'>press</span> 
+     </span>
+     <span class='ocr_line' id='line_1_38' title="bbox 346 2089 1212 2124; baseline 0.002 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_317' title='bbox 346 2089 391 2116; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_318' title='bbox 402 2094 465 2117; x_wconf 86'>note</span> <span class='ocrx_word' id='word_1_319' title='bbox 475 2094 504 2117; x_wconf 86'>to</span> <span class='ocrx_word' id='word_1_320' title='bbox 515 2090 549 2117; x_wconf 93'>be</span> <span class='ocrx_word' id='word_1_321' title='bbox 559 2090 652 2117; x_wconf 89'>erased</span> <span class='ocrx_word' id='word_1_322' title='bbox 661 2090 718 2123; x_wconf 91'>just</span> <span class='ocrx_word' id='word_1_323' title='bbox 729 2090 822 2117; x_wconf 85'>before</span> <span class='ocrx_word' id='word_1_324' title='bbox 833 2090 852 2117; x_wconf 89'>it</span> <span class='ocrx_word' id='word_1_325' title='bbox 862 2090 937 2124; x_wconf 87'>plays</span> <span class='ocrx_word' id='word_1_326' title='bbox 947 2090 975 2117; x_wconf 92'>in</span> <span class='ocrx_word' id='word_1_327' title='bbox 986 2090 1032 2118; x_wconf 87'>the</span> <span class='ocrx_word' id='word_1_328' title='bbox 1043 2098 1212 2124; x_wconf 78'>sequence—</span> 
+     </span>
+     <span class='ocr_line' id='line_1_39' title="bbox 346 2129 1134 2163; baseline 0.003 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_329' title='bbox 346 2129 425 2156; x_wconf 92'>when</span> <span class='ocrx_word' id='word_1_330' title='bbox 435 2129 531 2162; x_wconf 91'>played</span> <span class='ocrx_word' id='word_1_331' title='bbox 542 2129 621 2161; x_wconf 82'>back,</span> <span class='ocrx_word' id='word_1_332' title='bbox 633 2129 652 2156; x_wconf 84'>it</span> <span class='ocrx_word' id='word_1_333' title='bbox 663 2129 716 2156; x_wconf 89'>will</span> <span class='ocrx_word' id='word_1_334' title='bbox 726 2130 760 2157; x_wconf 88'>be</span> <span class='ocrx_word' id='word_1_335' title='bbox 771 2138 849 2163; x_wconf 86'>gone.</span> <span class='ocrx_word' id='word_1_336' title='bbox 861 2131 947 2157; x_wconf 86'>Notes</span> <span class='ocrx_word' id='word_1_337' title='bbox 958 2138 1021 2163; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_338' title='bbox 1031 2130 1089 2157; x_wconf 87'>also</span> <span class='ocrx_word' id='word_1_339' title='bbox 1100 2130 1134 2157; x_wconf 91'>be</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_16' title="bbox 345 2168 1228 2242">
+    <p class='ocr_par' id='par_1_21' lang='eng' title="bbox 345 2168 1228 2242">
+     <span class='ocr_line' id='line_1_40' title="bbox 346 2168 1220 2203; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_340' title='bbox 346 2168 445 2200; x_wconf 86'>added,</span> <span class='ocrx_word' id='word_1_341' title='bbox 457 2168 560 2200; x_wconf 87'>erased,</span> <span class='ocrx_word' id='word_1_342' title='bbox 572 2177 604 2196; x_wconf 93'>or</span> <span class='ocrx_word' id='word_1_343' title='bbox 614 2169 739 2202; x_wconf 82'>changed</span> <span class='ocrx_word' id='word_1_344' title='bbox 749 2169 829 2203; x_wconf 86'>using</span> <span class='ocrx_word' id='word_1_345' title='bbox 839 2169 885 2196; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_346' title='bbox 896 2170 1031 2196; x_wconf 89'>SINGLE</span> <span class='ocrx_word' id='word_1_347' title='bbox 1042 2170 1131 2196; x_wconf 87'>STEP</span> <span class='ocrx_word' id='word_1_348' title='bbox 1143 2169 1220 2196; x_wconf 82'>func-</span> 
+     </span>
+     <span class='ocr_line' id='line_1_41' title="bbox 345 2207 1228 2242; baseline 0.002 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_349' title='bbox 345 2207 412 2234; x_wconf 87'>tion.</span> <span class='ocrx_word' id='word_1_350' title='bbox 424 2208 461 2235; x_wconf 88'>To</span> <span class='ocrx_word' id='word_1_351' title='bbox 472 2208 592 2235; x_wconf 88'>overdub</span> <span class='ocrx_word' id='word_1_352' title='bbox 603 2212 680 2235; x_wconf 84'>notes</span> <span class='ocrx_word' id='word_1_353' title='bbox 691 2212 718 2235; x_wconf 89'>at</span> <span class='ocrx_word' id='word_1_354' title='bbox 729 2208 841 2242; x_wconf 83'>specific</span> <span class='ocrx_word' id='word_1_355' title='bbox 851 2209 943 2242; x_wconf 89'>points</span> <span class='ocrx_word' id='word_1_356' title='bbox 955 2208 1049 2236; x_wconf 87'>within</span> <span class='ocrx_word' id='word_1_357' title='bbox 1060 2217 1076 2236; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_358' title='bbox 1086 2216 1228 2242; x_wconf 86'>sequence,</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_17' title="bbox 345 2293 705 2326">
+    <p class='ocr_par' id='par_1_22' lang='eng' title="bbox 345 2293 705 2326">
+     <span class='ocr_line' id='line_1_42' title="bbox 345 2293 705 2326; baseline 0.003 -1; x_size 38.38298; x_descenders 5.3829789; x_ascenders 10"><span class='ocrx_word' id='word_1_359' title='bbox 345 2293 541 2326; x_wconf 86'>Additional</span> <span class='ocrx_word' id='word_1_360' title='bbox 553 2293 705 2326; x_wconf 83'>Features</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_18' title="bbox 1295 1289 2165 1593">
+    <p class='ocr_par' id='par_1_23' lang='eng' title="bbox 1296 1289 2165 1362">
+     <span class='ocr_line' id='line_1_43' title="bbox 1296 1289 2165 1322; baseline 0.001 -7; x_size 35.040108; x_descenders 6.9264708; x_ascenders 8.613636"><span class='ocrx_word' id='word_1_361' title='bbox 1296 1289 1394 1322; x_wconf 91'>simply</span> <span class='ocrx_word' id='word_1_362' title='bbox 1404 1297 1452 1316; x_wconf 87'>use</span> <span class='ocrx_word' id='word_1_363' title='bbox 1463 1290 1615 1321; x_wconf 86'>LOCATE,</span> <span class='ocrx_word' id='word_1_364' title='bbox 1628 1290 1716 1316; x_wconf 88'>FAST</span> <span class='ocrx_word' id='word_1_365' title='bbox 1726 1289 1917 1321; x_wconf 86'>FORWARD,</span> <span class='ocrx_word' id='word_1_366' title='bbox 1929 1297 1960 1317; x_wconf 92'>or</span> <span class='ocrx_word' id='word_1_367' title='bbox 1972 1290 2126 1316; x_wconf 91'>REWIND</span> <span class='ocrx_word' id='word_1_368' title='bbox 2136 1294 2165 1317; x_wconf 82'>to</span> 
+     </span>
+     <span class='ocr_line' id='line_1_44' title="bbox 1297 1328 2033 1362; baseline 0.001 -7; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_369' title='bbox 1297 1328 1356 1355; x_wconf 77'>find</span> <span class='ocrx_word' id='word_1_370' title='bbox 1366 1329 1412 1355; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_371' title='bbox 1423 1328 1527 1356; x_wconf 85'>desired</span> <span class='ocrx_word' id='word_1_372' title='bbox 1537 1329 1587 1356; x_wconf 88'>bar</span> <span class='ocrx_word' id='word_1_373' title='bbox 1598 1329 1720 1361; x_wconf 86'>number,</span> <span class='ocrx_word' id='word_1_374' title='bbox 1732 1329 1798 1356; x_wconf 78'>then</span> <span class='ocrx_word' id='word_1_375' title='bbox 1809 1333 1874 1356; x_wconf 87'>start</span> <span class='ocrx_word' id='word_1_376' title='bbox 1885 1329 2033 1362; x_wconf 85'>recording.</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_24' lang='eng' title="bbox 1295 1368 2160 1593">
+     <span class='ocr_line' id='line_1_45' title="bbox 1332 1368 2160 1402; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_377' title='bbox 1332 1368 1391 1395; x_wconf 90'>The</span> <span class='ocrx_word' id='word_1_378' title='bbox 1402 1368 1650 1395; x_wconf 86'>INSERT/COPY</span> <span class='ocrx_word' id='word_1_379' title='bbox 1662 1368 1789 1395; x_wconf 81'>function</span> <span class='ocrx_word' id='word_1_380' title='bbox 1800 1368 1893 1395; x_wconf 90'>allows</span> <span class='ocrx_word' id='word_1_381' title='bbox 1904 1376 1958 1402; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_382' title='bbox 1968 1373 1997 1395; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_383' title='bbox 2008 1377 2087 1396; x_wconf 91'>move</span> <span class='ocrx_word' id='word_1_384' title='bbox 2097 1369 2160 1396; x_wconf 89'>bars</span> 
+     </span>
+     <span class='ocr_line' id='line_1_46' title="bbox 1297 1407 2151 1441; baseline 0.002 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_385' title='bbox 1297 1407 1369 1434; x_wconf 86'>from</span> <span class='ocrx_word' id='word_1_386' title='bbox 1380 1415 1433 1434; x_wconf 88'>one</span> <span class='ocrx_word' id='word_1_387' title='bbox 1443 1407 1565 1435; x_wconf 85'>location</span> <span class='ocrx_word' id='word_1_388' title='bbox 1576 1411 1605 1434; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_389' title='bbox 1616 1407 1796 1435; x_wconf 85'>another—in</span> <span class='ocrx_word' id='word_1_390' title='bbox 1806 1408 1852 1435; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_391' title='bbox 1863 1416 1937 1435; x_wconf 90'>same</span> <span class='ocrx_word' id='word_1_392' title='bbox 1948 1416 2083 1441; x_wconf 88'>sequence</span> <span class='ocrx_word' id='word_1_393' title='bbox 2093 1416 2125 1435; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_394' title='bbox 2135 1416 2151 1435; x_wconf 88'>a</span> 
+     </span>
+     <span class='ocr_line' id='line_1_47' title="bbox 1296 1447 2160 1481; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_395' title='bbox 1296 1447 1421 1474; x_wconf 81'>different</span> <span class='ocrx_word' id='word_1_396' title='bbox 1431 1455 1491 1474; x_wconf 89'>one.</span> <span class='ocrx_word' id='word_1_397' title='bbox 1504 1448 1555 1474; x_wconf 87'>For</span> <span class='ocrx_word' id='word_1_398' title='bbox 1565 1448 1698 1481; x_wconf 84'>example,</span> <span class='ocrx_word' id='word_1_399' title='bbox 1710 1456 1764 1481; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_400' title='bbox 1775 1448 1863 1481; x_wconf 84'>might</span> <span class='ocrx_word' id='word_1_401' title='bbox 1873 1448 1955 1475; x_wconf 86'>insert</span> <span class='ocrx_word' id='word_1_402' title='bbox 1966 1456 1983 1475; x_wconf 92'>a</span> <span class='ocrx_word' id='word_1_403' title='bbox 1993 1456 2064 1481; x_wconf 88'>copy</span> <span class='ocrx_word' id='word_1_404' title='bbox 2074 1448 2106 1475; x_wconf 90'>of</span> <span class='ocrx_word' id='word_1_405' title='bbox 2115 1448 2160 1475; x_wconf 90'>the</span> 
+     </span>
+     <span class='ocr_line' id='line_1_48' title="bbox 1297 1486 2087 1521; baseline 0 -7; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_406' title='bbox 1297 1486 1355 1514; x_wconf 82'>first</span> <span class='ocrx_word' id='word_1_407' title='bbox 1366 1495 1439 1514; x_wconf 89'>verse</span> <span class='ocrx_word' id='word_1_408' title='bbox 1449 1487 1571 1514; x_wconf 89'>between</span> <span class='ocrx_word' id='word_1_409' title='bbox 1582 1487 1628 1514; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_410' title='bbox 1638 1487 1740 1514; x_wconf 87'>second</span> <span class='ocrx_word' id='word_1_411' title='bbox 1751 1487 1852 1514; x_wconf 83'>chorus</span> <span class='ocrx_word' id='word_1_412' title='bbox 1863 1487 1919 1514; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_413' title='bbox 1929 1487 1975 1514; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_414' title='bbox 1985 1487 2087 1521; x_wconf 85'>bridge.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_49' title="bbox 1296 1527 2047 1560; baseline 0.003 -8; x_size 32; x_descenders 6; x_ascenders 7"><span class='ocrx_word' id='word_1_415' title='bbox 1296 1527 1441 1553; x_wconf 88'>DELETE</span> <span class='ocrx_word' id='word_1_416' title='bbox 1453 1527 1546 1553; x_wconf 90'>BARS</span> <span class='ocrx_word' id='word_1_417' title='bbox 1557 1531 1681 1559; x_wconf 84'>operates</span> <span class='ocrx_word' id='word_1_418' title='bbox 1691 1527 1737 1553; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_419' title='bbox 1748 1535 1823 1554; x_wconf 90'>same</span> <span class='ocrx_word' id='word_1_420' title='bbox 1833 1535 1891 1560; x_wconf 93'>way</span> <span class='ocrx_word' id='word_1_421' title='bbox 1901 1531 1930 1554; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_422' title='bbox 1940 1535 2047 1554; x_wconf 90'>remove</span> 
+     </span>
+     <span class='ocr_line' id='line_1_50' title="bbox 1295 1565 1577 1593; baseline 0.004 -1; x_size 34.748871; x_descenders 6.7488689; x_ascenders 9"><span class='ocrx_word' id='word_1_423' title='bbox 1295 1565 1441 1592; x_wconf 88'>unwanted</span> <span class='ocrx_word' id='word_1_424' title='bbox 1452 1565 1577 1593; x_wconf 84'>sections.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_19' title="bbox 1295 1638 2215 1960">
+    <p class='ocr_par' id='par_1_25' lang='eng' title="bbox 1296 1638 1589 1680">
+     <span class='ocr_line' id='line_1_51' title="bbox 1296 1638 1589 1680; baseline 0 -9; x_size 42; x_descenders 9; x_ascenders 10"><span class='ocrx_word' id='word_1_425' title='bbox 1296 1638 1452 1680; x_wconf 86'>Creating</span> <span class='ocrx_word' id='word_1_426' title='bbox 1465 1648 1485 1671; x_wconf 85'>a</span> <span class='ocrx_word' id='word_1_427' title='bbox 1497 1639 1589 1680; x_wconf 87'>Song</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_26' lang='eng' title="bbox 1295 1690 2215 1960">
+     <span class='ocr_line' id='line_1_52' title="bbox 1333 1690 2146 1723; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_428' title='bbox 1333 1690 1394 1717; x_wconf 88'>One</span> <span class='ocrx_word' id='word_1_429' title='bbox 1404 1698 1462 1723; x_wconf 93'>way</span> <span class='ocrx_word' id='word_1_430' title='bbox 1472 1694 1500 1717; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_431' title='bbox 1511 1694 1598 1717; x_wconf 88'>create</span> <span class='ocrx_word' id='word_1_432' title='bbox 1608 1698 1625 1717; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_433' title='bbox 1635 1698 1704 1723; x_wconf 83'>song</span> <span class='ocrx_word' id='word_1_434' title='bbox 1715 1690 1736 1717; x_wconf 90'>is</span> <span class='ocrx_word' id='word_1_435' title='bbox 1747 1694 1776 1717; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_436' title='bbox 1787 1690 1880 1717; x_wconf 86'>record</span> <span class='ocrx_word' id='word_1_437' title='bbox 1891 1690 1958 1717; x_wconf 89'>each</span> <span class='ocrx_word' id='word_1_438' title='bbox 1969 1690 2044 1717; x_wconf 87'>track</span> <span class='ocrx_word' id='word_1_439' title='bbox 2056 1690 2090 1717; x_wconf 89'>all</span> <span class='ocrx_word' id='word_1_440' title='bbox 2101 1690 2146 1717; x_wconf 88'>the</span> 
+     </span>
+     <span class='ocr_line' id='line_1_53' title="bbox 1295 1729 2121 1762; baseline 0.001 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_441' title='bbox 1295 1737 1353 1762; x_wconf 87'>way</span> <span class='ocrx_word' id='word_1_442' title='bbox 1362 1729 1481 1762; x_wconf 83'>through</span> <span class='ocrx_word' id='word_1_443' title='bbox 1493 1729 1541 1762; x_wconf 90'>(up</span> <span class='ocrx_word' id='word_1_444' title='bbox 1552 1733 1581 1756; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_445' title='bbox 1592 1730 1644 1756; x_wconf 89'>999</span> <span class='ocrx_word' id='word_1_446' title='bbox 1654 1729 1738 1762; x_wconf 90'>bars).</span> <span class='ocrx_word' id='word_1_447' title='bbox 1751 1729 1878 1757; x_wconf 89'>Another</span> <span class='ocrx_word' id='word_1_448' title='bbox 1888 1737 1945 1762; x_wconf 89'>way</span> <span class='ocrx_word' id='word_1_449' title='bbox 1956 1729 1977 1757; x_wconf 90'>is</span> <span class='ocrx_word' id='word_1_450' title='bbox 1987 1733 2016 1757; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_451' title='bbox 2027 1729 2121 1757; x_wconf 86'>record</span> 
+     </span>
+     <span class='ocr_line' id='line_1_54' title="bbox 1296 1768 2066 1802; baseline 0 -6; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_452' title='bbox 1296 1768 1362 1796; x_wconf 86'>each</span> <span class='ocrx_word' id='word_1_453' title='bbox 1373 1768 1448 1796; x_wconf 87'>basic</span> <span class='ocrx_word' id='word_1_454' title='bbox 1458 1768 1562 1796; x_wconf 83'>section</span> <span class='ocrx_word' id='word_1_455' title='bbox 1574 1769 1666 1802; x_wconf 89'>(verse,</span> <span class='ocrx_word' id='word_1_456' title='bbox 1679 1769 1788 1801; x_wconf 86'>chorus,</span> <span class='ocrx_word' id='word_1_457' title='bbox 1800 1769 1865 1802; x_wconf 83'>etc.)</span> <span class='ocrx_word' id='word_1_458' title='bbox 1876 1768 1904 1795; x_wconf 91'>in</span> <span class='ocrx_word' id='word_1_459' title='bbox 1914 1768 2066 1796; x_wconf 86'>individual</span> 
+     </span>
+     <span class='ocr_line' id='line_1_55' title="bbox 1296 1808 2215 1841; baseline 0 -6; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_460' title='bbox 1296 1816 1451 1841; x_wconf 85'>sequences,</span> <span class='ocrx_word' id='word_1_461' title='bbox 1463 1808 1528 1835; x_wconf 86'>then</span> <span class='ocrx_word' id='word_1_462' title='bbox 1538 1816 1587 1835; x_wconf 90'>use</span> <span class='ocrx_word' id='word_1_463' title='bbox 1597 1808 1643 1835; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_464' title='bbox 1653 1809 1799 1835; x_wconf 89'>CREATE</span> <span class='ocrx_word' id='word_1_465' title='bbox 1810 1808 1911 1835; x_wconf 89'>SONG</span> <span class='ocrx_word' id='word_1_466' title='bbox 1923 1808 2050 1836; x_wconf 85'>function</span> <span class='ocrx_word' id='word_1_467' title='bbox 2060 1812 2089 1835; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_468' title='bbox 2103 1808 2215 1836; x_wconf 87'>“chain”</span> 
+     </span>
+     <span class='ocr_line' id='line_1_56' title="bbox 1295 1847 2135 1881; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_469' title='bbox 1295 1848 1370 1874; x_wconf 91'>them</span> <span class='ocrx_word' id='word_1_470' title='bbox 1381 1848 1508 1881; x_wconf 83'>together.</span> <span class='ocrx_word' id='word_1_471' title='bbox 1521 1848 1667 1875; x_wconf 87'>CREATE</span> <span class='ocrx_word' id='word_1_472' title='bbox 1678 1848 1779 1875; x_wconf 91'>SONG</span> <span class='ocrx_word' id='word_1_473' title='bbox 1789 1847 1842 1874; x_wconf 92'>will</span> <span class='ocrx_word' id='word_1_474' title='bbox 1853 1848 1918 1875; x_wconf 86'>then</span> <span class='ocrx_word' id='word_1_475' title='bbox 1929 1848 2135 1881; x_wconf 81'>automatically</span> 
+     </span>
+     <span class='ocr_line' id='line_1_57' title="bbox 1296 1887 2162 1920; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_476' title='bbox 1296 1895 1366 1920; x_wconf 89'>copy</span> <span class='ocrx_word' id='word_1_477' title='bbox 1377 1887 1412 1914; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_478' title='bbox 1422 1887 1468 1914; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_479' title='bbox 1478 1891 1552 1920; x_wconf 86'>parts</span> <span class='ocrx_word' id='word_1_480' title='bbox 1563 1887 1621 1914; x_wconf 86'>into</span> <span class='ocrx_word' id='word_1_481' title='bbox 1632 1895 1649 1914; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_482' title='bbox 1659 1895 1718 1914; x_wconf 91'>new</span> <span class='ocrx_word' id='word_1_483' title='bbox 1729 1895 1870 1920; x_wconf 83'>sequence.</span> <span class='ocrx_word' id='word_1_484' title='bbox 1884 1887 1911 1914; x_wconf 84'>If</span> <span class='ocrx_word' id='word_1_485' title='bbox 1919 1887 2033 1919; x_wconf 89'>desired,</span> <span class='ocrx_word' id='word_1_486' title='bbox 2045 1895 2099 1920; x_wconf 92'>you</span> <span class='ocrx_word' id='word_1_487' title='bbox 2110 1895 2162 1914; x_wconf 85'>can</span> 
+     </span>
+     <span class='ocr_line' id='line_1_58' title="bbox 1296 1926 2169 1960; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_488' title='bbox 1296 1934 1361 1953; x_wconf 89'>even</span> <span class='ocrx_word' id='word_1_489' title='bbox 1372 1930 1411 1953; x_wconf 90'>set</span> <span class='ocrx_word' id='word_1_490' title='bbox 1422 1926 1468 1953; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_491' title='bbox 1478 1926 1528 1953; x_wconf 89'>last</span> <span class='ocrx_word' id='word_1_492' title='bbox 1540 1926 1590 1953; x_wconf 88'>few</span> <span class='ocrx_word' id='word_1_493' title='bbox 1601 1927 1664 1953; x_wconf 89'>bars</span> <span class='ocrx_word' id='word_1_494' title='bbox 1675 1931 1704 1954; x_wconf 85'>to</span> <span class='ocrx_word' id='word_1_495' title='bbox 1715 1931 1806 1960; x_wconf 89'>repeat</span> <span class='ocrx_word' id='word_1_496' title='bbox 1816 1926 1955 1960; x_wconf 84'>infinitely,</span> <span class='ocrx_word' id='word_1_497' title='bbox 1968 1926 2011 1954; x_wconf 86'>for</span> <span class='ocrx_word' id='word_1_498' title='bbox 2022 1935 2038 1954; x_wconf 92'>a</span> <span class='ocrx_word' id='word_1_499' title='bbox 2049 1927 2169 1954; x_wconf 83'>fadeout.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_20' title="bbox 1293 2000 2179 2242">
+    <p class='ocr_par' id='par_1_27' lang='eng' title="bbox 1294 2000 1948 2040">
+     <span class='ocr_line' id='line_1_59' title="bbox 1294 2000 1948 2040; baseline 0.002 -8; x_size 40; x_descenders 8; x_ascenders 10"><span class='ocrx_word' id='word_1_500' title='bbox 1294 2000 1532 2040; x_wconf 84'>Composition</span> <span class='ocrx_word' id='word_1_501' title='bbox 1544 2000 1699 2033; x_wconf 85'>Without</span> <span class='ocrx_word' id='word_1_502' title='bbox 1711 2000 1948 2040; x_wconf 85'>Compromise</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_28' lang='eng' title="bbox 1293 2051 2179 2242">
+     <span class='ocr_line' id='line_1_60' title="bbox 1331 2051 2179 2085; baseline 0 -7; x_size 34; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_503' title='bbox 1331 2051 1389 2078; x_wconf 90'>The</span> <span class='ocrx_word' id='word_1_504' title='bbox 1399 2051 1563 2085; x_wconf 83'>technology</span> <span class='ocrx_word' id='word_1_505' title='bbox 1573 2059 1627 2084; x_wconf 92'>you</span> <span class='ocrx_word' id='word_1_506' title='bbox 1638 2060 1686 2079; x_wconf 86'>use</span> <span class='ocrx_word' id='word_1_507' title='bbox 1697 2051 1797 2078; x_wconf 89'>should</span> <span class='ocrx_word' id='word_1_508' title='bbox 1808 2059 1887 2078; x_wconf 87'>never</span> <span class='ocrx_word' id='word_1_509' title='bbox 1897 2052 1932 2079; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_510' title='bbox 1942 2059 1973 2079; x_wconf 89'>so</span> <span class='ocrx_word' id='word_1_511' title='bbox 1984 2051 2110 2085; x_wconf 85'>complex</span> <span class='ocrx_word' id='word_1_512' title='bbox 2120 2052 2179 2078; x_wconf 86'>that</span> 
+     </span>
+     <span class='ocr_line' id='line_1_61' title="bbox 1294 2090 2157 2124; baseline 0 -6; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_513' title='bbox 1294 2090 1313 2118; x_wconf 86'>it</span> <span class='ocrx_word' id='word_1_514' title='bbox 1323 2091 1459 2118; x_wconf 87'>interferes</span> <span class='ocrx_word' id='word_1_515' title='bbox 1470 2091 1535 2118; x_wconf 84'>with</span> <span class='ocrx_word' id='word_1_516' title='bbox 1545 2091 1591 2118; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_517' title='bbox 1602 2091 1715 2118; x_wconf 90'>creative</span> <span class='ocrx_word' id='word_1_518' title='bbox 1725 2099 1841 2124; x_wconf 83'>process.</span> <span class='ocrx_word' id='word_1_519' title='bbox 1854 2091 1947 2118; x_wconf 78'>That’s</span> <span class='ocrx_word' id='word_1_520' title='bbox 1957 2091 2086 2124; x_wconf 86'>precisely</span> <span class='ocrx_word' id='word_1_521' title='bbox 2096 2091 2157 2124; x_wconf 89'>why</span> 
+     </span>
+     <span class='ocr_line' id='line_1_62' title="bbox 1293 2130 2156 2164; baseline 0 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_522' title='bbox 1293 2130 1339 2157; x_wconf 90'>the</span> <span class='ocrx_word' id='word_1_523' title='bbox 1350 2130 1576 2164; x_wconf 80'>LinnSequencer</span> <span class='ocrx_word' id='word_1_524' title='bbox 1586 2130 1607 2157; x_wconf 89'>is</span> <span class='ocrx_word' id='word_1_525' title='bbox 1619 2130 1747 2164; x_wconf 87'>designed</span> <span class='ocrx_word' id='word_1_526' title='bbox 1758 2134 1787 2157; x_wconf 87'>to</span> <span class='ocrx_word' id='word_1_527' title='bbox 1798 2130 1834 2157; x_wconf 89'>let</span> <span class='ocrx_word' id='word_1_528' title='bbox 1844 2138 1899 2163; x_wconf 93'>you</span> <span class='ocrx_word' id='word_1_529' title='bbox 1910 2138 2051 2164; x_wconf 88'>compose,</span> <span class='ocrx_word' id='word_1_530' title='bbox 2063 2130 2156 2157; x_wconf 87'>record</span> 
+     </span>
+     <span class='ocr_line' id='line_1_63' title="bbox 1294 2169 2145 2203; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_531' title='bbox 1294 2170 1349 2197; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_532' title='bbox 1360 2169 1414 2197; x_wconf 83'>edit</span> <span class='ocrx_word' id='word_1_533' title='bbox 1425 2170 1503 2197; x_wconf 89'>while</span> <span class='ocrx_word' id='word_1_534' title='bbox 1514 2170 1643 2203; x_wconf 90'>devoting</span> <span class='ocrx_word' id='word_1_535' title='bbox 1653 2178 1721 2203; x_wconf 92'>your</span> <span class='ocrx_word' id='word_1_536' title='bbox 1731 2170 1882 2197; x_wconf 88'>undivided</span> <span class='ocrx_word' id='word_1_537' title='bbox 1893 2170 2027 2197; x_wconf 86'>attention</span> <span class='ocrx_word' id='word_1_538' title='bbox 2037 2174 2067 2197; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_539' title='bbox 2077 2178 2145 2203; x_wconf 90'>your</span> 
+     </span>
+     <span class='ocr_line' id='line_1_64' title="bbox 1294 2209 2126 2242; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_540' title='bbox 1294 2209 1388 2236; x_wconf 88'>music.</span> <span class='ocrx_word' id='word_1_541' title='bbox 1402 2209 1451 2236; x_wconf 89'>See</span> <span class='ocrx_word' id='word_1_542' title='bbox 1460 2217 1529 2242; x_wconf 92'>your</span> <span class='ocrx_word' id='word_1_543' title='bbox 1540 2209 1611 2236; x_wconf 87'>Linn</span> <span class='ocrx_word' id='word_1_544' title='bbox 1621 2209 1712 2236; x_wconf 86'>dealer</span> <span class='ocrx_word' id='word_1_545' title='bbox 1722 2209 1806 2242; x_wconf 85'>today</span> <span class='ocrx_word' id='word_1_546' title='bbox 1817 2209 1859 2236; x_wconf 87'>for</span> <span class='ocrx_word' id='word_1_547' title='bbox 1870 2217 1887 2236; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_548' title='bbox 1897 2209 2126 2236; x_wconf 76'>demonstration!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_21' title="bbox 347 2343 2171 2378">
+    <p class='ocr_par' id='par_1_29' lang='eng' title="bbox 347 2343 2171 2378">
+     <span class='ocr_line' id='line_1_65' title="bbox 347 2343 2171 2378; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_549' title='bbox 347 2350 361 2363; x_wconf 69'><strong>0</strong></span> <span class='ocrx_word' id='word_1_550' title='bbox 373 2343 483 2377; x_wconf 83'>Simple,</span> <span class='ocrx_word' id='word_1_551' title='bbox 495 2352 559 2377; x_wconf 89'>easy</span> <span class='ocrx_word' id='word_1_552' title='bbox 569 2348 598 2371; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_553' title='bbox 609 2344 683 2371; x_wconf 89'>learn</span> <span class='ocrx_word' id='word_1_554' title='bbox 694 2344 922 2378; x_wconf 80'>operation—the</span> <span class='ocrx_word' id='word_1_555' title='bbox 934 2345 967 2372; x_wconf 88'>32</span> <span class='ocrx_word' id='word_1_556' title='bbox 978 2345 1116 2372; x_wconf 86'>character</span> <span class='ocrx_word' id='word_1_557' title='bbox 1126 2345 1200 2372; x_wconf 92'>LCD</span> <span class='ocrx_word' id='word_1_558' title='bbox 1211 2345 1316 2378; x_wconf 90'>display</span> <span class='ocrx_word' id='word_1_559' title='bbox 1326 2345 1424 2378; x_wconf 88'>clearly</span> <span class='ocrx_word' id='word_1_560' title='bbox 1434 2345 1528 2378; x_wconf 86'>guides</span> <span class='ocrx_word' id='word_1_561' title='bbox 1539 2353 1594 2378; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_562' title='bbox 1604 2345 1724 2378; x_wconf 87'>through</span> <span class='ocrx_word' id='word_1_563' title='bbox 1735 2344 1770 2371; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_564' title='bbox 1781 2344 1947 2377; x_wconf 86'>operations.</span> <span class='ocrx_word' id='word_1_565' title='bbox 1961 2344 1989 2371; x_wconf 88'>If</span> <span class='ocrx_word' id='word_1_566' title='bbox 1997 2344 2112 2376; x_wconf 88'>needed,</span> <span class='ocrx_word' id='word_1_567' title='bbox 2125 2345 2171 2371; x_wconf 88'>the</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_22' title="bbox 373 2381 1083 2415">
+    <p class='ocr_par' id='par_1_30' lang='eng' title="bbox 373 2381 1083 2415">
+     <span class='ocr_line' id='line_1_66' title="bbox 373 2381 1083 2415; baseline 0.003 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_568' title='bbox 373 2381 472 2407; x_wconf 85'>HELP</span> <span class='ocrx_word' id='word_1_569' title='bbox 483 2381 583 2408; x_wconf 85'>button</span> <span class='ocrx_word' id='word_1_570' title='bbox 594 2381 711 2415; x_wconf 87'>displays</span> <span class='ocrx_word' id='word_1_571' title='bbox 722 2382 875 2409; x_wconf 89'>additional</span> <span class='ocrx_word' id='word_1_572' title='bbox 886 2382 1083 2415; x_wconf 87'>explanations.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_23' title="bbox 347 2427 2145 2507">
+    <p class='ocr_par' id='par_1_31' lang='eng' title="bbox 347 2427 2145 2507">
+     <span class='ocr_line' id='line_1_67' title="bbox 347 2427 1468 2461; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_573' title='bbox 347 2432 361 2446; x_wconf 69'><strong>0</strong></span> <span class='ocrx_word' id='word_1_574' title='bbox 373 2427 612 2454; x_wconf 88'>Non-destructive</span> <span class='ocrx_word' id='word_1_575' title='bbox 622 2427 914 2461; x_wconf 85'>recording—existing</span> <span class='ocrx_word' id='word_1_576' title='bbox 924 2432 1002 2455; x_wconf 90'>notes</span> <span class='ocrx_word' id='word_1_577' title='bbox 1013 2436 1057 2455; x_wconf 88'>are</span> <span class='ocrx_word' id='word_1_578' title='bbox 1068 2432 1116 2455; x_wconf 84'>not</span> <span class='ocrx_word' id='word_1_579' title='bbox 1127 2428 1220 2455; x_wconf 87'>erased</span> <span class='ocrx_word' id='word_1_580' title='bbox 1231 2428 1309 2455; x_wconf 91'>while</span> <span class='ocrx_word' id='word_1_581' title='bbox 1319 2428 1468 2461; x_wconf 87'>recording.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_68' title="bbox 347 2473 2145 2507; baseline 0.001 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_582' title='bbox 347 2478 361 2492; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_583' title='bbox 372 2473 433 2500; x_wconf 90'>Two</span> <span class='ocrx_word' id='word_1_584' title='bbox 444 2473 689 2500; x_wconf 89'>FOOTSWITCH</span> <span class='ocrx_word' id='word_1_585' title='bbox 701 2474 837 2501; x_wconf 79'>INPUTS</span> <span class='ocrx_word' id='word_1_586' title='bbox 848 2481 910 2507; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_587' title='bbox 921 2473 955 2500; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_588' title='bbox 966 2473 1091 2507; x_wconf 85'>assigned</span> <span class='ocrx_word' id='word_1_589' title='bbox 1101 2478 1130 2501; x_wconf 81'>to</span> <span class='ocrx_word' id='word_1_590' title='bbox 1141 2474 1271 2507; x_wconf 84'>remotely</span> <span class='ocrx_word' id='word_1_591' title='bbox 1281 2473 1387 2501; x_wconf 86'>control</span> <span class='ocrx_word' id='word_1_592' title='bbox 1398 2482 1481 2507; x_wconf 86'>many</span> <span class='ocrx_word' id='word_1_593' title='bbox 1491 2473 1523 2501; x_wconf 90'>of</span> <span class='ocrx_word' id='word_1_594' title='bbox 1531 2474 1578 2501; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_595' title='bbox 1588 2473 1750 2506; x_wconf 82'>commonly</span> <span class='ocrx_word' id='word_1_596' title='bbox 1760 2473 1827 2500; x_wconf 88'>used</span> <span class='ocrx_word' id='word_1_597' title='bbox 1840 2473 1991 2505; x_wconf 77'>functions,</span> <span class='ocrx_word' id='word_1_598' title='bbox 2003 2473 2145 2507; x_wconf 84'>including</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_24' title="bbox 372 2510 1090 2543">
+    <p class='ocr_par' id='par_1_32' lang='eng' title="bbox 372 2510 1090 2543">
+     <span class='ocr_line' id='line_1_69' title="bbox 372 2510 1090 2543; baseline 0.001 -6; x_size 35.625; x_descenders 8.90625; x_ascenders 8.90625"><span class='ocrx_word' id='word_1_599' title='bbox 372 2510 500 2542; x_wconf 88'>ERASE,</span> <span class='ocrx_word' id='word_1_600' title='bbox 513 2511 660 2542; x_wconf 86'>REPEAT,</span> <span class='ocrx_word' id='word_1_601' title='bbox 673 2511 883 2543; x_wconf 80'>PLAY/STOP,</span> <span class='ocrx_word' id='word_1_602' title='bbox 896 2519 927 2538; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_603' title='bbox 939 2511 1090 2538; x_wconf 83'>LOCATE.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_25' title="bbox 347 2556 1768 2590">
+    <p class='ocr_par' id='par_1_33' lang='eng' title="bbox 347 2556 1768 2590">
+     <span class='ocr_line' id='line_1_70' title="bbox 347 2556 1768 2590; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_604' title='bbox 347 2561 361 2575; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_605' title='bbox 372 2556 433 2583; x_wconf 91'>Two</span> <span class='ocrx_word' id='word_1_606' title='bbox 443 2556 612 2583; x_wconf 83'>TRIGGER</span> <span class='ocrx_word' id='word_1_607' title='bbox 623 2556 797 2584; x_wconf 84'>OUTPUTS</span> <span class='ocrx_word' id='word_1_608' title='bbox 808 2565 871 2590; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_609' title='bbox 881 2557 915 2584; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_610' title='bbox 925 2557 1119 2590; x_wconf 84'>programmed</span> <span class='ocrx_word' id='word_1_611' title='bbox 1129 2561 1158 2584; x_wconf 93'>to</span> <span class='ocrx_word' id='word_1_612' title='bbox 1169 2561 1270 2590; x_wconf 87'>output</span> <span class='ocrx_word' id='word_1_613' title='bbox 1281 2557 1370 2590; x_wconf 88'>pulses</span> <span class='ocrx_word' id='word_1_614' title='bbox 1381 2561 1409 2584; x_wconf 92'>at</span> <span class='ocrx_word' id='word_1_615' title='bbox 1419 2565 1472 2590; x_wconf 90'>any</span> <span class='ocrx_word' id='word_1_616' title='bbox 1483 2557 1598 2584; x_wconf 86'>selected</span> <span class='ocrx_word' id='word_1_617' title='bbox 1608 2561 1673 2584; x_wconf 89'>note</span> <span class='ocrx_word' id='word_1_618' title='bbox 1683 2556 1768 2583; x_wconf 89'>value.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_26' title="bbox 347 2601 1226 2635">
+    <p class='ocr_par' id='par_1_34' lang='eng' title="bbox 347 2601 1226 2635">
+     <span class='ocr_line' id='line_1_71' title="bbox 347 2601 1226 2635; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_619' title='bbox 347 2607 361 2620; x_wconf 65'><strong>0</strong></span> <span class='ocrx_word' id='word_1_620' title='bbox 372 2601 434 2628; x_wconf 88'>Will</span> <span class='ocrx_word' id='word_1_621' title='bbox 445 2609 510 2634; x_wconf 89'>sync</span> <span class='ocrx_word' id='word_1_622' title='bbox 521 2605 549 2628; x_wconf 84'>to</span> <span class='ocrx_word' id='word_1_623' title='bbox 561 2602 690 2629; x_wconf 88'>standard</span> <span class='ocrx_word' id='word_1_624' title='bbox 701 2602 864 2629; x_wconf 86'>LinnDrum</span> <span class='ocrx_word' id='word_1_625' title='bbox 875 2610 907 2629; x_wconf 91'>or</span> <span class='ocrx_word' id='word_1_626' title='bbox 918 2602 989 2629; x_wconf 89'>Linn</span> <span class='ocrx_word' id='word_1_627' title='bbox 1000 2603 1069 2629; x_wconf 90'>9000</span> <span class='ocrx_word' id='word_1_628' title='bbox 1080 2610 1145 2635; x_wconf 87'>sync</span> <span class='ocrx_word' id='word_1_629' title='bbox 1155 2607 1226 2630; x_wconf 89'>tone.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_27' title="bbox 347 2648 2100 2727">
+    <p class='ocr_par' id='par_1_35' lang='eng' title="bbox 347 2648 2100 2727">
+     <span class='ocr_line' id='line_1_72' title="bbox 347 2648 1664 2682; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_630' title='bbox 347 2654 360 2667; x_wconf 75'>&#39;</span> <span class='ocrx_word' id='word_1_631' title='bbox 372 2648 483 2675; x_wconf 89'>Utilizes</span> <span class='ocrx_word' id='word_1_632' title='bbox 493 2648 564 2680; x_wconf 75'>ultra</span> <span class='ocrx_word' id='word_1_633' title='bbox 573 2648 744 2682; x_wconf 86'>high—speed,</span> <span class='ocrx_word' id='word_1_634' title='bbox 757 2649 772 2676; x_wconf 91'>8</span> <span class='ocrx_word' id='word_1_635' title='bbox 783 2649 862 2675; x_wconf 89'>MHz</span> <span class='ocrx_word' id='word_1_636' title='bbox 873 2649 954 2676; x_wconf 88'>80186</span> <span class='ocrx_word' id='word_1_637' title='bbox 965 2649 994 2676; x_wconf 91'>16</span> <span class='ocrx_word' id='word_1_638' title='bbox 1004 2648 1043 2676; x_wconf 90'>bit</span> <span class='ocrx_word' id='word_1_639' title='bbox 1054 2653 1197 2682; x_wconf 88'>computer</span> <span class='ocrx_word' id='word_1_640' title='bbox 1208 2649 1350 2682; x_wconf 90'>internally</span> <span class='ocrx_word' id='word_1_641' title='bbox 1361 2649 1403 2676; x_wconf 87'>for</span> <span class='ocrx_word' id='word_1_642' title='bbox 1414 2649 1502 2676; x_wconf 87'>FAST</span> <span class='ocrx_word' id='word_1_643' title='bbox 1512 2648 1664 2682; x_wconf 81'>operation.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_73' title="bbox 347 2694 2100 2727; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_644' title='bbox 347 2699 361 2713; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_645' title='bbox 372 2694 504 2721; x_wconf 88'>TEMPO</span> <span class='ocrx_word' id='word_1_646' title='bbox 515 2702 578 2727; x_wconf 92'>may</span> <span class='ocrx_word' id='word_1_647' title='bbox 589 2694 623 2721; x_wconf 90'>be</span> <span class='ocrx_word' id='word_1_648' title='bbox 633 2694 764 2727; x_wconf 87'>specified</span> <span class='ocrx_word' id='word_1_649' title='bbox 774 2694 802 2721; x_wconf 89'>in</span> <span class='ocrx_word' id='word_1_650' title='bbox 814 2695 1172 2722; x_wconf 84'>BEATS-PER—MINUTE</span> <span class='ocrx_word' id='word_1_651' title='bbox 1183 2703 1215 2722; x_wconf 93'>or</span> <span class='ocrx_word' id='word_1_652' title='bbox 1225 2695 1567 2722; x_wconf 84'>FRAMES-PER—BEAT</span> <span class='ocrx_word' id='word_1_653' title='bbox 1577 2698 1605 2721; x_wconf 89'>at</span> <span class='ocrx_word' id='word_1_654' title='bbox 1616 2695 1659 2726; x_wconf 91'>24,</span> <span class='ocrx_word' id='word_1_655' title='bbox 1672 2695 1716 2726; x_wconf 88'>25,</span> <span class='ocrx_word' id='word_1_656' title='bbox 1728 2702 1760 2721; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_657' title='bbox 1772 2695 1806 2721; x_wconf 88'>30</span> <span class='ocrx_word' id='word_1_658' title='bbox 1817 2694 1917 2721; x_wconf 86'>frames</span> <span class='ocrx_word' id='word_1_659' title='bbox 1928 2702 1976 2727; x_wconf 89'>per</span> <span class='ocrx_word' id='word_1_660' title='bbox 1987 2694 2100 2726; x_wconf 87'>second,</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_28' title="bbox 373 2731 652 2765">
+    <p class='ocr_par' id='par_1_36' lang='eng' title="bbox 373 2731 652 2765">
+     <span class='ocr_line' id='line_1_74' title="bbox 373 2731 652 2765; baseline 0 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_661' title='bbox 373 2731 449 2764; x_wconf 87'>(even</span> <span class='ocrx_word' id='word_1_662' title='bbox 460 2731 529 2765; x_wconf 88'>drop</span> <span class='ocrx_word' id='word_1_663' title='bbox 541 2731 652 2764; x_wconf 81'>frame!)</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_29' title="bbox 347 2777 2174 2811">
+    <p class='ocr_par' id='par_1_37' lang='eng' title="bbox 347 2777 2174 2811">
+     <span class='ocr_line' id='line_1_75' title="bbox 347 2777 2174 2811; baseline 0.001 -8; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_664' title='bbox 347 2782 360 2796; x_wconf 72'><strong>0</strong></span> <span class='ocrx_word' id='word_1_665' title='bbox 372 2777 504 2804; x_wconf 82'>TEMPO</span> <span class='ocrx_word' id='word_1_666' title='bbox 515 2785 578 2810; x_wconf 91'>may</span> <span class='ocrx_word' id='word_1_667' title='bbox 588 2777 622 2804; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_668' title='bbox 633 2778 741 2804; x_wconf 84'>entered</span> <span class='ocrx_word' id='word_1_669' title='bbox 751 2777 934 2811; x_wconf 80'>numerically,</span> <span class='ocrx_word' id='word_1_670' title='bbox 946 2777 1101 2811; x_wconf 85'>adjustable</span> <span class='ocrx_word' id='word_1_671' title='bbox 1111 2777 1139 2804; x_wconf 92'>in</span> <span class='ocrx_word' id='word_1_672' title='bbox 1149 2778 1239 2805; x_wconf 88'>tenths</span> <span class='ocrx_word' id='word_1_673' title='bbox 1250 2778 1282 2805; x_wconf 85'>of</span> <span class='ocrx_word' id='word_1_674' title='bbox 1290 2786 1307 2805; x_wconf 94'>a</span> <span class='ocrx_word' id='word_1_675' title='bbox 1317 2777 1567 2805; x_wconf 85'>Beat-Per-Minute</span> <span class='ocrx_word' id='word_1_676' title='bbox 1577 2777 1748 2809; x_wconf 85'>increments,</span> <span class='ocrx_word' id='word_1_677' title='bbox 1760 2785 1792 2804; x_wconf 87'>or</span> <span class='ocrx_word' id='word_1_678' title='bbox 1803 2777 1839 2810; x_wconf 94'>by</span> <span class='ocrx_word' id='word_1_679' title='bbox 1849 2777 1964 2811; x_wconf 84'>tapping</span> <span class='ocrx_word' id='word_1_680' title='bbox 1975 2781 2086 2810; x_wconf 85'>quarter</span> <span class='ocrx_word' id='word_1_681' title='bbox 2097 2781 2174 2804; x_wconf 85'>notes</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_30' title="bbox 372 2815 808 2842">
+    <p class='ocr_par' id='par_1_38' lang='eng' title="bbox 372 2815 808 2842">
+     <span class='ocr_line' id='line_1_76' title="bbox 372 2815 808 2842; baseline 0.002 -1; x_size 32.351353; x_descenders 5.3513513; x_ascenders 9"><span class='ocrx_word' id='word_1_682' title='bbox 372 2822 410 2841; x_wconf 93'>on</span> <span class='ocrx_word' id='word_1_683' title='bbox 420 2815 466 2842; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_684' title='bbox 476 2815 545 2841; x_wconf 91'>TAP</span> <span class='ocrx_word' id='word_1_685' title='bbox 556 2815 689 2842; x_wconf 85'>TEMPO</span> <span class='ocrx_word' id='word_1_686' title='bbox 699 2815 808 2842; x_wconf 83'>button.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_31' title="bbox 347 2861 1792 2940">
+    <p class='ocr_par' id='par_1_39' lang='eng' title="bbox 347 2861 1792 2940">
+     <span class='ocr_line' id='line_1_77' title="bbox 347 2861 1792 2895; baseline 0.001 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_687' title='bbox 347 2866 360 2880; x_wconf 68'><strong>0</strong></span> <span class='ocrx_word' id='word_1_688' title='bbox 372 2861 504 2887; x_wconf 87'>TEMPO</span> <span class='ocrx_word' id='word_1_689' title='bbox 515 2861 696 2888; x_wconf 86'>CHANGES</span> <span class='ocrx_word' id='word_1_690' title='bbox 707 2869 771 2894; x_wconf 92'>may</span> <span class='ocrx_word' id='word_1_691' title='bbox 781 2861 815 2888; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_692' title='bbox 825 2861 1019 2894; x_wconf 85'>programmed</span> <span class='ocrx_word' id='word_1_693' title='bbox 1030 2861 1087 2888; x_wconf 88'>into</span> <span class='ocrx_word' id='word_1_694' title='bbox 1099 2869 1115 2888; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_695' title='bbox 1126 2870 1268 2895; x_wconf 87'>sequence,</span> <span class='ocrx_word' id='word_1_696' title='bbox 1280 2861 1344 2888; x_wconf 89'>with</span> <span class='ocrx_word' id='word_1_697' title='bbox 1356 2862 1467 2888; x_wconf 88'>smooth</span> <span class='ocrx_word' id='word_1_698' title='bbox 1478 2861 1635 2888; x_wconf 85'>transitions</span> <span class='ocrx_word' id='word_1_699' title='bbox 1646 2861 1670 2887; x_wconf 90'>if</span> <span class='ocrx_word' id='word_1_700' title='bbox 1679 2861 1792 2888; x_wconf 89'>desired.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_78' title="bbox 347 2906 1507 2940; baseline 0.002 -8; x_size 33; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_701' title='bbox 347 2911 360 2925; x_wconf 74'><strong><em>°</em></strong></span> <span class='ocrx_word' id='word_1_702' title='bbox 371 2906 434 2938; x_wconf 92'>Any</span> <span class='ocrx_word' id='word_1_703' title='bbox 444 2906 539 2932; x_wconf 85'>TIME</span> <span class='ocrx_word' id='word_1_704' title='bbox 550 2906 763 2933; x_wconf 88'>SIGNATURE</span> <span class='ocrx_word' id='word_1_705' title='bbox 773 2915 836 2939; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_706' title='bbox 846 2907 880 2933; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_707' title='bbox 891 2907 968 2938; x_wconf 89'>used,</span> <span class='ocrx_word' id='word_1_708' title='bbox 980 2907 1036 2934; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_709' title='bbox 1046 2915 1109 2940; x_wconf 91'>may</span> <span class='ocrx_word' id='word_1_710' title='bbox 1120 2907 1154 2934; x_wconf 90'>be</span> <span class='ocrx_word' id='word_1_711' title='bbox 1164 2907 1288 2940; x_wconf 84'>changed</span> <span class='ocrx_word' id='word_1_712' title='bbox 1299 2907 1393 2934; x_wconf 83'>within</span> <span class='ocrx_word' id='word_1_713' title='bbox 1404 2915 1420 2934; x_wconf 89'>a</span> <span class='ocrx_word' id='word_1_714' title='bbox 1431 2915 1507 2940; x_wconf 87'>song.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_32' title="bbox 1540 2970 1878 3140">
+    <p class='ocr_par' id='par_1_40' lang='eng' title="bbox 1540 2970 1878 3140">
+     <span class='ocr_line' id='line_1_79' title="bbox 1540 2970 1800 3089; baseline 0 -14; x_size 100.5; x_descenders 16.5; x_ascenders 28.499998"><span class='ocrx_word' id='word_1_715' title='bbox 1540 2970 1800 3089; x_wconf 74'>E1111</span> 
+     </span>
+     <span class='ocr_line' id='line_1_80' title="bbox 1554 3108 1878 3140; baseline 0 -5; x_size 32; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_716' title='bbox 1554 3108 1624 3135; x_wconf 91'>Linn</span> <span class='ocrx_word' id='word_1_717' title='bbox 1635 3108 1808 3140; x_wconf 87'>Electronics,</span> <span class='ocrx_word' id='word_1_718' title='bbox 1822 3109 1878 3136; x_wconf 84'>Inc.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_33' title="bbox 1554 3150 2188 3226">
+    <p class='ocr_par' id='par_1_41' lang='eng' title="bbox 1554 3150 2188 3226">
+     <span class='ocr_line' id='line_1_81' title="bbox 1557 3150 2163 3182; baseline 0 -5; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_719' title='bbox 1557 3150 1638 3177; x_wconf 88'>18720</span> <span class='ocrx_word' id='word_1_720' title='bbox 1648 3150 1761 3177; x_wconf 89'>Oxnard</span> <span class='ocrx_word' id='word_1_721' title='bbox 1772 3150 1866 3182; x_wconf 81'>Street,</span> <span class='ocrx_word' id='word_1_722' title='bbox 1878 3150 2006 3182; x_wconf 86'>Tarzana,</span> <span class='ocrx_word' id='word_1_723' title='bbox 2019 3150 2071 3177; x_wconf 93'>CA</span> <span class='ocrx_word' id='word_1_724' title='bbox 2082 3150 2163 3177; x_wconf 88'>91356</span> 
+     </span>
+     <span class='ocr_line' id='line_1_82' title="bbox 1554 3192 2188 3226; baseline 0 -7; x_size 33.5; x_descenders 5.5; x_ascenders 8.5"><span class='ocrx_word' id='word_1_725' title='bbox 1554 3193 1622 3226; x_wconf 89'>(818)</span> <span class='ocrx_word' id='word_1_726' title='bbox 1633 3193 1755 3219; x_wconf 60'>7088131</span> <span class='ocrx_word' id='word_1_727' title='bbox 1765 3193 1788 3219; x_wconf 91'>T</span> <span class='ocrx_word' id='word_1_728' title='bbox 1789 3193 1888 3219; x_wconf 87'>ELEX</span> <span class='ocrx_word' id='word_1_729' title='bbox 1899 3192 2022 3219; x_wconf 82'>#298949</span> <span class='ocrx_word' id='word_1_730' title='bbox 2033 3193 2125 3219; x_wconf 88'>LINN</span> <span class='ocrx_word' id='word_1_731' title='bbox 2135 3193 2188 3219; x_wconf 89'>UR</span> 
+     </span>
+    </p>
+   </div>
+  </div>
+ </body>
+</html>

tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/cardinal/__-l__eng__000001.ocr.png__000001__hocr__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state—of—the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+0 Operation is similar to multi—track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+0 Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers !
+
+0 Ultra—fast 3 1/2 ” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+0 One or all tracks may be TRANSPOSED at the touch of a key.
+0 Exclusive real—time ERASE function makes editing FAST.
+0 Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+0 TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+0 Optional SMPTE time code synchronization.
+
+0 Optional remote control.
+
+Recording 3 Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+—existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot—recording. To overdub a new part,
+select a different track and start recording~while you
+record, the first‘track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections.
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+0 Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+0 Non-destructive recording—existing notes are not erased while recording.
+0 Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+0 Two TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+0 Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+' Utilizes ultra high—speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+0 TEMPO may be specified in BEATS-PER—MINUTE or FRAMES-PER—BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+0 TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+0 TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+° Any TIME SIGNATURE may be used, and may be changed within a song.
+
+E1111
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 7088131 T ELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/cardinal/__-l__eng__000002.ocr.png__000002.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state—of—the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+0 Operation is similar to multi—track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+0 Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers !
+
+0 Ultra—fast 3 1/2 ” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+0 One or all tracks may be TRANSPOSED at the touch of a key.
+0 Exclusive real—time ERASE function makes editing FAST.
+0 Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+0 TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+0 Optional SMPTE time code synchronization.
+
+0 Optional remote control.
+
+Recording 3 Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+—existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot—recording. To overdub a new part,
+select a different track and start recording~while you
+record, the first‘track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections.
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+0 Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+0 Non-destructive recording—existing notes are not erased while recording.
+0 Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+0 Two TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+0 Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+' Utilizes ultra high—speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+0 TEMPO may be specified in BEATS-PER—MINUTE or FRAMES-PER—BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+0 TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+0 TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+° Any TIME SIGNATURE may be used, and may be changed within a song.
+
+E1111
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 7088131 T ELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/hocr.bin

@@ -0,0 +1,335 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
+    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ <head>
+  <title></title>
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
+  <meta name='ocr-system' content='tesseract 3.05.01' />
+  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
+</head>
+<body>
+  <div class='ocr_page' id='page_1' title='image "/var/folders/37/78_114p552q16vv6vmgm5kr00000gn/T/com.github.ocrmypdf.wahjk6fa/000002.ocr.png"; bbox 0 0 2550 3300; ppageno 0'>
+   <div class='ocr_carea' id='block_1_1' title="bbox 582 131 1968 303">
+    <p class='ocr_par' id='par_1_1' lang='eng' title="bbox 582 131 1968 303">
+     <span class='ocr_line' id='line_1_1' title="bbox 882 131 1657 217; baseline 0.001 -17; x_size 85; x_descenders 16; x_ascenders 19"><span class='ocrx_word' id='word_1_1' title='bbox 882 132 1036 202; x_wconf 86'>The</span> <span class='ocrx_word' id='word_1_2' title='bbox 1061 131 1657 217; x_wconf 84'>LinnSequencer</span> 
+     </span>
+     <span class='ocr_line' id='line_1_2' title="bbox 582 215 1968 303; baseline 0 -17; x_size 87; x_descenders 16; x_ascenders 21"><span class='ocrx_word' id='word_1_3' title='bbox 582 215 674 286; x_wconf 90'>32</span> <span class='ocrx_word' id='word_1_4' title='bbox 697 218 923 288; x_wconf 84'>Track</span> <span class='ocrx_word' id='word_1_5' title='bbox 948 218 1181 287; x_wconf 94'><strong>MIDI</strong></span> <span class='ocrx_word' id='word_1_6' title='bbox 1208 217 1575 303; x_wconf 89'>Sequence</span> <span class='ocrx_word' id='word_1_7' title='bbox 1600 218 1968 288; x_wconf 83'>Recorder</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_2' title="bbox 347 380 2188 574">
+    <p class='ocr_par' id='par_1_2' lang='eng' title="bbox 347 380 2188 423">
+     <span class='ocr_line' id='line_1_3' title="bbox 347 380 2188 423; baseline -0.001 -12; x_size 38; x_descenders 8; x_ascenders 10"><span class='ocrx_word' id='word_1_8' title='bbox 347 380 412 410; x_wconf 86'>The</span> <span class='ocrx_word' id='word_1_9' title='bbox 424 380 676 417; x_wconf 85'>LinnSequencer</span> <span class='ocrx_word' id='word_1_10' title='bbox 688 380 712 411; x_wconf 87'>is</span> <span class='ocrx_word' id='word_1_11' title='bbox 724 390 743 411; x_wconf 90'>a</span> <span class='ocrx_word' id='word_1_12' title='bbox 754 381 1005 423; x_wconf 79'>state—of—the-art</span> <span class='ocrx_word' id='word_1_13' title='bbox 1017 380 1226 418; x_wconf 86'>composition</span> <span class='ocrx_word' id='word_1_14' title='bbox 1238 381 1299 411; x_wconf 88'>and</span> <span class='ocrx_word' id='word_1_15' title='bbox 1311 380 1525 418; x_wconf 85'>performance</span> <span class='ocrx_word' id='word_1_16' title='bbox 1536 380 1602 411; x_wconf 85'>tool</span> <span class='ocrx_word' id='word_1_17' title='bbox 1615 380 1663 411; x_wconf 82'>for</span> <span class='ocrx_word' id='word_1_18' title='bbox 1674 381 1725 410; x_wconf 94'>the</span> <span class='ocrx_word' id='word_1_19' title='bbox 1737 380 1940 417; x_wconf 85'>professional</span> <span class='ocrx_word' id='word_1_20' title='bbox 1952 380 2112 411; x_wconf 87'>musician.</span> <span class='ocrx_word' id='word_1_21' title='bbox 2127 381 2152 410; x_wconf 90'>It</span> <span class='ocrx_word' id='word_1_22' title='bbox 2164 380 2188 410; x_wconf 90'>is</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_3' lang='eng' title="bbox 347 430 1988 468">
+     <span class='ocr_line' id='line_1_4' title="bbox 347 430 1988 468; baseline 0 -8; x_size 37; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_23' title='bbox 347 430 507 467; x_wconf 86'>extremely</span> <span class='ocrx_word' id='word_1_24' title='bbox 518 430 677 467; x_wconf 84'>powerful,</span> <span class='ocrx_word' id='word_1_25' title='bbox 691 435 739 467; x_wconf 91'>yet</span> <span class='ocrx_word' id='word_1_26' title='bbox 752 430 924 468; x_wconf 87'>amazingly</span> <span class='ocrx_word' id='word_1_27' title='bbox 936 430 1044 468; x_wconf 85'>simple</span> <span class='ocrx_word' id='word_1_28' title='bbox 1055 435 1087 461; x_wconf 87'>to</span> <span class='ocrx_word' id='word_1_29' title='bbox 1099 431 1183 461; x_wconf 88'>learn</span> <span class='ocrx_word' id='word_1_30' title='bbox 1195 431 1257 461; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_31' title='bbox 1269 440 1329 461; x_wconf 89'>use.</span> <span class='ocrx_word' id='word_1_32' title='bbox 1344 431 1393 461; x_wconf 87'>It’s</span> <span class='ocrx_word' id='word_1_33' title='bbox 1406 440 1499 467; x_wconf 91'>many</span> <span class='ocrx_word' id='word_1_34' title='bbox 1510 430 1701 461; x_wconf 88'>remarkable</span> <span class='ocrx_word' id='word_1_35' title='bbox 1714 430 1843 460; x_wconf 85'>features</span> <span class='ocrx_word' id='word_1_36' title='bbox 1856 430 1988 461; x_wconf 87'>include:</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_4' lang='eng' title="bbox 350 482 2093 574">
+     <span class='ocr_line' id='line_1_5' title="bbox 350 482 2093 527; baseline 0 -9; x_size 43; x_descenders 7; x_ascenders 11"><span class='ocrx_word' id='word_1_37' title='bbox 350 490 368 508; x_wconf 68'><strong>0</strong></span> <span class='ocrx_word' id='word_1_38' title='bbox 383 482 585 526; x_wconf 88'>Operation</span> <span class='ocrx_word' id='word_1_39' title='bbox 598 482 627 518; x_wconf 89'>is</span> <span class='ocrx_word' id='word_1_40' title='bbox 641 482 776 518; x_wconf 88'>similar</span> <span class='ocrx_word' id='word_1_41' title='bbox 789 488 829 518; x_wconf 88'>to</span> <span class='ocrx_word' id='word_1_42' title='bbox 843 482 1062 519; x_wconf 86'>multi—track</span> <span class='ocrx_word' id='word_1_43' title='bbox 1076 488 1160 527; x_wconf 92'>tape</span> <span class='ocrx_word' id='word_1_44' title='bbox 1173 483 1336 519; x_wconf 88'>recorder</span> <span class='ocrx_word' id='word_1_45' title='bbox 1350 482 1436 518; x_wconf 87'>with</span> <span class='ocrx_word' id='word_1_46' title='bbox 1451 483 1580 525; x_wconf 90'>PLAY,</span> <span class='ocrx_word' id='word_1_47' title='bbox 1598 483 1724 525; x_wconf 90'>STOP,</span> <span class='ocrx_word' id='word_1_48' title='bbox 1741 483 1957 525; x_wconf 89'>RECORD,</span> <span class='ocrx_word' id='word_1_49' title='bbox 1974 483 2093 518; x_wconf 88'>FAST</span> 
+     </span>
+     <span class='ocr_line' id='line_1_6' title="bbox 383 532 1345 574; baseline 0.001 -7; x_size 43; x_descenders 7; x_ascenders 11"><span class='ocrx_word' id='word_1_50' title='bbox 383 532 635 574; x_wconf 88'>FORWARD,</span> <span class='ocrx_word' id='word_1_51' title='bbox 652 532 865 574; x_wconf 89'>REWIND,</span> <span class='ocrx_word' id='word_1_52' title='bbox 882 532 956 568; x_wconf 91'>and</span> <span class='ocrx_word' id='word_1_53' title='bbox 971 532 1163 568; x_wconf 90'>LOCATE</span> <span class='ocrx_word' id='word_1_54' title='bbox 1177 532 1345 568; x_wconf 87'>controls.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_3' title="bbox 349 589 2136 685">
+    <p class='ocr_par' id='par_1_5' lang='eng' title="bbox 349 589 2136 685">
+     <span class='ocr_line' id='line_1_7' title="bbox 349 589 2136 634; baseline 0.001 -9; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_55' title='bbox 349 597 368 615; x_wconf 67'><strong>0</strong></span> <span class='ocrx_word' id='word_1_56' title='bbox 383 590 482 625; x_wconf 90'>Each</span> <span class='ocrx_word' id='word_1_57' title='bbox 496 589 539 625; x_wconf 87'>of</span> <span class='ocrx_word' id='word_1_58' title='bbox 549 590 611 626; x_wconf 92'>the</span> <span class='ocrx_word' id='word_1_59' title='bbox 625 590 686 625; x_wconf 91'>100</span> <span class='ocrx_word' id='word_1_60' title='bbox 700 601 895 634; x_wconf 89'>sequences</span> <span class='ocrx_word' id='word_1_61' title='bbox 909 590 1075 626; x_wconf 88'>contains</span> <span class='ocrx_word' id='word_1_62' title='bbox 1090 591 1135 626; x_wconf 89'>32</span> <span class='ocrx_word' id='word_1_63' title='bbox 1149 590 1418 633; x_wconf 87'>simultaneous,</span> <span class='ocrx_word' id='word_1_64' title='bbox 1435 590 1661 634; x_wconf 89'>polyphonic</span> <span class='ocrx_word' id='word_1_65' title='bbox 1674 590 1802 626; x_wconf 85'>tracks.</span> <span class='ocrx_word' id='word_1_66' title='bbox 1819 590 1921 626; x_wconf 89'>Each</span> <span class='ocrx_word' id='word_1_67' title='bbox 1934 590 2035 626; x_wconf 88'>track</span> <span class='ocrx_word' id='word_1_68' title='bbox 2050 600 2136 634; x_wconf 89'>may</span> 
+     </span>
+     <span class='ocr_line' id='line_1_8' title="bbox 383 639 2022 685; baseline 0.001 -10; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_69' title='bbox 383 639 428 675; x_wconf 95'>be</span> <span class='ocrx_word' id='word_1_70' title='bbox 442 639 607 684; x_wconf 88'>assigned</span> <span class='ocrx_word' id='word_1_71' title='bbox 621 645 659 676; x_wconf 92'>to</span> <span class='ocrx_word' id='word_1_72' title='bbox 674 650 745 676; x_wconf 91'>one</span> <span class='ocrx_word' id='word_1_73' title='bbox 759 640 801 676; x_wconf 84'>of</span> <span class='ocrx_word' id='word_1_74' title='bbox 813 641 851 676; x_wconf 92'>16</span> <span class='ocrx_word' id='word_1_75' title='bbox 864 641 980 675; x_wconf 90'>MIDI</span> <span class='ocrx_word' id='word_1_76' title='bbox 996 640 1176 676; x_wconf 88'>channels.</span> <span class='ocrx_word' id='word_1_77' title='bbox 1194 640 1498 684; x_wconf 90'>Simultaneously</span> <span class='ocrx_word' id='word_1_78' title='bbox 1510 640 1609 685; x_wconf 90'>plays</span> <span class='ocrx_word' id='word_1_79' title='bbox 1624 651 1674 684; x_wconf 91'>up</span> <span class='ocrx_word' id='word_1_80' title='bbox 1688 645 1727 676; x_wconf 90'>to</span> <span class='ocrx_word' id='word_1_81' title='bbox 1742 641 1780 676; x_wconf 94'>16</span> <span class='ocrx_word' id='word_1_82' title='bbox 1794 640 2022 684; x_wconf 87'>polyphonic</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_4' title="bbox 384 689 632 734">
+    <p class='ocr_par' id='par_1_6' lang='eng' title="bbox 384 689 632 734">
+     <span class='ocr_line' id='line_1_9' title="bbox 384 689 632 734; baseline 0.004 -9; x_size 46; x_descenders 9; x_ascenders 12"><span class='ocrx_word' id='word_1_83' title='bbox 384 689 618 734; x_wconf 84'>synthesizers</span> <span class='ocrx_word' id='word_1_84' title='bbox 625 690 632 726; x_wconf 85'><strong>!</strong></span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_5' title="bbox 349 748 2117 793">
+    <p class='ocr_par' id='par_1_7' lang='eng' title="bbox 349 748 2117 793">
+     <span class='ocr_line' id='line_1_10' title="bbox 349 748 2117 793; baseline 0 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_85' title='bbox 349 755 367 774; x_wconf 66'><strong>0</strong></span> <span class='ocrx_word' id='word_1_86' title='bbox 383 748 573 784; x_wconf 87'>Ultra—fast</span> <span class='ocrx_word' id='word_1_87' title='bbox 588 749 606 784; x_wconf 90'>3</span> <span class='ocrx_word' id='word_1_88' title='bbox 616 749 652 784; x_wconf 86'>1/2</span> <span class='ocrx_word' id='word_1_89' title='bbox 663 749 677 764; x_wconf 87'><em>”</em></span> <span class='ocrx_word' id='word_1_90' title='bbox 694 748 775 784; x_wconf 91'>disk</span> <span class='ocrx_word' id='word_1_91' title='bbox 790 748 887 784; x_wconf 89'>drive</span> <span class='ocrx_word' id='word_1_92' title='bbox 901 754 1012 785; x_wconf 85'>stores</span> <span class='ocrx_word' id='word_1_93' title='bbox 1027 748 1194 793; x_wconf 87'>complex</span> <span class='ocrx_word' id='word_1_94' title='bbox 1208 760 1316 793; x_wconf 87'>songs</span> <span class='ocrx_word' id='word_1_95' title='bbox 1330 748 1368 784; x_wconf 87'>in</span> <span class='ocrx_word' id='word_1_96' title='bbox 1382 748 1535 785; x_wconf 87'>seconds</span> <span class='ocrx_word' id='word_1_97' title='bbox 1550 748 1624 784; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_98' title='bbox 1638 748 1746 784; x_wconf 87'>holds</span> <span class='ocrx_word' id='word_1_99' title='bbox 1761 759 1844 784; x_wconf 85'>over</span> <span class='ocrx_word' id='word_1_100' title='bbox 1859 749 2000 791; x_wconf 88'>110,000</span> <span class='ocrx_word' id='word_1_101' title='bbox 2013 753 2117 784; x_wconf 89'>notes</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_6' title="bbox 383 797 554 842">
+    <p class='ocr_par' id='par_1_8' lang='eng' title="bbox 383 797 554 842">
+     <span class='ocr_line' id='line_1_11' title="bbox 383 797 554 842; baseline 0 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_102' title='bbox 383 808 446 842; x_wconf 82'>per</span> <span class='ocrx_word' id='word_1_103' title='bbox 459 797 554 834; x_wconf 86'>disk!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_7' title="bbox 349 855 2030 1016">
+    <p class='ocr_par' id='par_1_9' lang='eng' title="bbox 349 855 2030 1016">
+     <span class='ocr_line' id='line_1_12' title="bbox 350 855 1638 900; baseline 0.001 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_104' title='bbox 350 863 367 881; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_105' title='bbox 383 856 464 891; x_wconf 90'>One</span> <span class='ocrx_word' id='word_1_106' title='bbox 478 866 520 891; x_wconf 87'>or</span> <span class='ocrx_word' id='word_1_107' title='bbox 534 855 580 891; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_108' title='bbox 594 856 712 892; x_wconf 89'>tracks</span> <span class='ocrx_word' id='word_1_109' title='bbox 726 867 811 900; x_wconf 90'>may</span> <span class='ocrx_word' id='word_1_110' title='bbox 823 856 869 892; x_wconf 88'>be</span> <span class='ocrx_word' id='word_1_111' title='bbox 882 856 1212 892; x_wconf 90'>TRANSPOSED</span> <span class='ocrx_word' id='word_1_112' title='bbox 1227 861 1264 892; x_wconf 90'>at</span> <span class='ocrx_word' id='word_1_113' title='bbox 1277 856 1338 892; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_114' title='bbox 1351 856 1463 892; x_wconf 90'>touch</span> <span class='ocrx_word' id='word_1_115' title='bbox 1477 856 1520 892; x_wconf 89'>of</span> <span class='ocrx_word' id='word_1_116' title='bbox 1531 867 1554 892; x_wconf 90'>a</span> <span class='ocrx_word' id='word_1_117' title='bbox 1568 856 1638 900; x_wconf 89'>key.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_13' title="bbox 350 913 1535 958; baseline 0.001 -9; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_118' title='bbox 350 921 367 939; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_119' title='bbox 383 913 568 950; x_wconf 84'>Exclusive</span> <span class='ocrx_word' id='word_1_120' title='bbox 581 913 756 950; x_wconf 87'>real—time</span> <span class='ocrx_word' id='word_1_121' title='bbox 769 914 929 950; x_wconf 91'>ERASE</span> <span class='ocrx_word' id='word_1_122' title='bbox 944 914 1114 950; x_wconf 87'>function</span> <span class='ocrx_word' id='word_1_123' title='bbox 1128 914 1251 950; x_wconf 88'>makes</span> <span class='ocrx_word' id='word_1_124' title='bbox 1266 914 1400 958; x_wconf 87'>editing</span> <span class='ocrx_word' id='word_1_125' title='bbox 1414 915 1535 950; x_wconf 91'>FAST.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_14' title="bbox 349 971 2030 1016; baseline 0.001 -10; x_size 44; x_descenders 8; x_ascenders 11"><span class='ocrx_word' id='word_1_126' title='bbox 349 979 367 997; x_wconf 67'><strong>0</strong></span> <span class='ocrx_word' id='word_1_127' title='bbox 382 971 568 1007; x_wconf 83'>Exclusive</span> <span class='ocrx_word' id='word_1_128' title='bbox 582 972 773 1007; x_wconf 87'>REPEAT</span> <span class='ocrx_word' id='word_1_129' title='bbox 787 972 958 1008; x_wconf 86'>function</span> <span class='ocrx_word' id='word_1_130' title='bbox 971 972 1245 1016; x_wconf 89'>automatically</span> <span class='ocrx_word' id='word_1_131' title='bbox 1258 977 1396 1016; x_wconf 90'>repeats</span> <span class='ocrx_word' id='word_1_132' title='bbox 1410 983 1481 1016; x_wconf 88'>any</span> <span class='ocrx_word' id='word_1_133' title='bbox 1493 972 1578 1008; x_wconf 91'>held</span> <span class='ocrx_word' id='word_1_134' title='bbox 1592 977 1694 1008; x_wconf 91'>notes</span> <span class='ocrx_word' id='word_1_135' title='bbox 1709 977 1746 1008; x_wconf 85'>at</span> <span class='ocrx_word' id='word_1_136' title='bbox 1760 982 1783 1008; x_wconf 89'>a</span> <span class='ocrx_word' id='word_1_137' title='bbox 1796 972 2030 1016; x_wconf 87'>pre-selected</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_8' title="bbox 382 1021 689 1065">
+    <p class='ocr_par' id='par_1_10' lang='eng' title="bbox 382 1021 689 1065">
+     <span class='ocr_line' id='line_1_15' title="bbox 382 1021 689 1065; baseline 0.003 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_138' title='bbox 382 1021 564 1065; x_wconf 87'>rhythmic</span> <span class='ocrx_word' id='word_1_139' title='bbox 577 1021 689 1058; x_wconf 86'>value.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_9' title="bbox 349 1080 2174 1125">
+    <p class='ocr_par' id='par_1_11' lang='eng' title="bbox 349 1080 2174 1125">
+     <span class='ocr_line' id='line_1_16' title="bbox 349 1080 2174 1125; baseline 0.001 -11; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_140' title='bbox 349 1087 367 1105; x_wconf 73'><strong>0</strong></span> <span class='ocrx_word' id='word_1_141' title='bbox 382 1080 567 1115; x_wconf 87'>TIMING</span> <span class='ocrx_word' id='word_1_142' title='bbox 582 1080 908 1116; x_wconf 90'>CORRECTION</span> <span class='ocrx_word' id='word_1_143' title='bbox 921 1080 1041 1116; x_wconf 90'>works</span> <span class='ocrx_word' id='word_1_144' title='bbox 1056 1080 1186 1124; x_wconf 86'>during</span> <span class='ocrx_word' id='word_1_145' title='bbox 1199 1080 1378 1124; x_wconf 89'>playback</span> <span class='ocrx_word' id='word_1_146' title='bbox 1392 1080 1466 1116; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_147' title='bbox 1480 1085 1644 1124; x_wconf 87'>operates</span> <span class='ocrx_word' id='word_1_148' title='bbox 1658 1080 1814 1116; x_wconf 91'>without</span> <span class='ocrx_word' id='word_1_149' title='bbox 1831 1080 2044 1125; x_wconf 87'>‘chopping’</span> <span class='ocrx_word' id='word_1_150' title='bbox 2061 1085 2174 1116; x_wconf 88'>notes.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_10' title="bbox 349 1137 1287 1182">
+    <p class='ocr_par' id='par_1_12' lang='eng' title="bbox 349 1137 1287 1182">
+     <span class='ocr_line' id='line_1_17' title="bbox 349 1137 1287 1182; baseline 0.001 -9; x_size 45; x_descenders 9; x_ascenders 11"><span class='ocrx_word' id='word_1_151' title='bbox 349 1145 367 1163; x_wconf 73'><strong>0</strong></span> <span class='ocrx_word' id='word_1_152' title='bbox 382 1137 560 1182; x_wconf 85'>Optional</span> <span class='ocrx_word' id='word_1_153' title='bbox 575 1138 739 1174; x_wconf 89'>SMPTE</span> <span class='ocrx_word' id='word_1_154' title='bbox 752 1138 839 1174; x_wconf 89'>time</span> <span class='ocrx_word' id='word_1_155' title='bbox 853 1138 945 1174; x_wconf 89'>code</span> <span class='ocrx_word' id='word_1_156' title='bbox 959 1138 1287 1182; x_wconf 88'>synchronization.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_11' title="bbox 349 1195 874 1240">
+    <p class='ocr_par' id='par_1_13' lang='eng' title="bbox 349 1195 874 1240">
+     <span class='ocr_line' id='line_1_18' title="bbox 349 1195 874 1240; baseline 0 -8; x_size 45; x_descenders 8; x_ascenders 12"><span class='ocrx_word' id='word_1_157' title='bbox 349 1203 367 1222; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_158' title='bbox 382 1195 560 1240; x_wconf 87'>Optional</span> <span class='ocrx_word' id='word_1_159' title='bbox 573 1201 709 1233; x_wconf 86'>remote</span> <span class='ocrx_word' id='word_1_160' title='bbox 723 1196 874 1233; x_wconf 87'>control.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_12' title="bbox 346 1288 1239 1491">
+    <p class='ocr_par' id='par_1_14' lang='eng' title="bbox 346 1288 749 1329">
+     <span class='ocr_line' id='line_1_19' title="bbox 346 1288 749 1329; baseline 0.002 -9; x_size 42; x_descenders 9; x_ascenders 10"><span class='ocrx_word' id='word_1_161' title='bbox 346 1288 535 1329; x_wconf 82'>Recording</span> <span class='ocrx_word' id='word_1_162' title='bbox 547 1298 567 1321; x_wconf 73'>3</span> <span class='ocrx_word' id='word_1_163' title='bbox 579 1288 749 1328; x_wconf 86'>Sequence</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_15' lang='eng' title="bbox 346 1339 1239 1491">
+     <span class='ocr_line' id='line_1_20' title="bbox 383 1339 1239 1373; baseline 0 -7; x_size 33; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_164' title='bbox 383 1340 420 1366; x_wconf 87'>To</span> <span class='ocrx_word' id='word_1_165' title='bbox 430 1339 524 1366; x_wconf 85'>record</span> <span class='ocrx_word' id='word_1_166' title='bbox 535 1347 551 1366; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_167' title='bbox 562 1347 704 1373; x_wconf 80'>sequence,</span> <span class='ocrx_word' id='word_1_168' title='bbox 716 1340 815 1373; x_wconf 91'>simply</span> <span class='ocrx_word' id='word_1_169' title='bbox 825 1348 898 1373; x_wconf 88'>press</span> <span class='ocrx_word' id='word_1_170' title='bbox 910 1340 1065 1367; x_wconf 90'>RECORD</span> <span class='ocrx_word' id='word_1_171' title='bbox 1075 1340 1131 1367; x_wconf 87'>and</span> <span class='ocrx_word' id='word_1_172' title='bbox 1142 1341 1239 1372; x_wconf 87'>PLAY,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_21' title="bbox 346 1378 1205 1412; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_173' title='bbox 346 1379 411 1406; x_wconf 91'>then</span> <span class='ocrx_word' id='word_1_174' title='bbox 422 1378 483 1412; x_wconf 89'>play</span> <span class='ocrx_word' id='word_1_175' title='bbox 493 1387 562 1412; x_wconf 87'>your</span> <span class='ocrx_word' id='word_1_176' title='bbox 572 1379 659 1405; x_wconf 89'>MIDI</span> <span class='ocrx_word' id='word_1_177' title='bbox 671 1379 810 1412; x_wconf 86'>keyboard</span> <span class='ocrx_word' id='word_1_178' title='bbox 821 1379 848 1406; x_wconf 90'>in</span> <span class='ocrx_word' id='word_1_179' title='bbox 858 1379 923 1406; x_wconf 90'>time</span> <span class='ocrx_word' id='word_1_180' title='bbox 934 1384 963 1406; x_wconf 90'>to</span> <span class='ocrx_word' id='word_1_181' title='bbox 974 1379 1019 1406; x_wconf 92'>the</span> <span class='ocrx_word' id='word_1_182' title='bbox 1030 1379 1205 1412; x_wconf 88'>Sequencer’s</span> 
+     </span>
+     <span class='ocr_line' id='line_1_22' title="bbox 347 1417 1232 1452; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_183' title='bbox 347 1417 414 1445; x_wconf 86'>click</span> <span class='ocrx_word' id='word_1_184' title='bbox 425 1418 506 1445; x_wconf 84'>track.</span> <span class='ocrx_word' id='word_1_185' title='bbox 519 1418 609 1445; x_wconf 85'>When</span> <span class='ocrx_word' id='word_1_186' title='bbox 620 1419 665 1445; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_187' title='bbox 676 1426 810 1452; x_wconf 86'>sequence</span> <span class='ocrx_word' id='word_1_188' title='bbox 821 1419 901 1452; x_wconf 88'>loops</span> <span class='ocrx_word' id='word_1_189' title='bbox 912 1419 983 1446; x_wconf 86'>back</span> <span class='ocrx_word' id='word_1_190' title='bbox 995 1419 1101 1446; x_wconf 90'>around</span> <span class='ocrx_word' id='word_1_191' title='bbox 1112 1423 1141 1446; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_192' title='bbox 1152 1419 1201 1446; x_wconf 87'>bar</span> <span class='ocrx_word' id='word_1_193' title='bbox 1213 1419 1232 1450; x_wconf 88'>1,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_23' title="bbox 346 1457 1223 1491; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_194' title='bbox 346 1457 430 1490; x_wconf 84'>you’ll</span> <span class='ocrx_word' id='word_1_195' title='bbox 441 1457 506 1485; x_wconf 93'>hear</span> <span class='ocrx_word' id='word_1_196' title='bbox 517 1458 590 1485; x_wconf 79'>what</span> <span class='ocrx_word' id='word_1_197' title='bbox 600 1466 654 1491; x_wconf 88'>you</span> <span class='ocrx_word' id='word_1_198' title='bbox 666 1458 865 1491; x_wconf 77'>played—only</span> <span class='ocrx_word' id='word_1_199' title='bbox 876 1458 910 1485; x_wconf 90'>all</span> <span class='ocrx_word' id='word_1_200' title='bbox 921 1458 1018 1491; x_wconf 84'>timing</span> <span class='ocrx_word' id='word_1_201' title='bbox 1029 1466 1115 1485; x_wconf 84'>errors</span> <span class='ocrx_word' id='word_1_202' title='bbox 1126 1458 1179 1485; x_wconf 91'>will</span> <span class='ocrx_word' id='word_1_203' title='bbox 1189 1458 1223 1485; x_wconf 91'>be</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_13' title="bbox 346 1497 1245 1531">
+    <p class='ocr_par' id='par_1_16' lang='eng' title="bbox 346 1497 1245 1531">
+     <span class='ocr_line' id='line_1_24' title="bbox 346 1497 1245 1531; baseline 0.001 -7; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_204' title='bbox 346 1497 494 1524; x_wconf 86'>corrected!</span> <span class='ocrx_word' id='word_1_205' title='bbox 508 1497 628 1530; x_wconf 85'>(Timing</span> <span class='ocrx_word' id='word_1_206' title='bbox 638 1497 791 1525; x_wconf 82'>correction</span> <span class='ocrx_word' id='word_1_207' title='bbox 802 1506 864 1531; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_208' title='bbox 875 1498 909 1525; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_209' title='bbox 920 1497 1047 1531; x_wconf 81'>adjusted</span> <span class='ocrx_word' id='word_1_210' title='bbox 1058 1505 1089 1525; x_wconf 92'>or</span> <span class='ocrx_word' id='word_1_211' title='bbox 1099 1497 1245 1531; x_wconf 87'>defeated).</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_14' title="bbox 346 1536 1205 1965">
+    <p class='ocr_par' id='par_1_17' lang='eng' title="bbox 346 1536 1186 1610">
+     <span class='ocr_line' id='line_1_25' title="bbox 346 1536 1186 1570; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_212' title='bbox 346 1537 409 1569; x_wconf 90'>Any</span> <span class='ocrx_word' id='word_1_213' title='bbox 419 1536 571 1563; x_wconf 86'>additional</span> <span class='ocrx_word' id='word_1_214' title='bbox 582 1541 659 1564; x_wconf 85'>notes</span> <span class='ocrx_word' id='word_1_215' title='bbox 670 1537 766 1570; x_wconf 87'>played</span> <span class='ocrx_word' id='word_1_216' title='bbox 777 1537 830 1563; x_wconf 88'>will</span> <span class='ocrx_word' id='word_1_217' title='bbox 840 1537 874 1564; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_218' title='bbox 885 1537 976 1564; x_wconf 88'>added</span> <span class='ocrx_word' id='word_1_219' title='bbox 986 1537 1045 1564; x_wconf 89'>into</span> <span class='ocrx_word' id='word_1_220' title='bbox 1055 1537 1101 1564; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_221' title='bbox 1111 1537 1186 1564; x_wconf 82'>track</span> 
+     </span>
+     <span class='ocr_line' id='line_1_26' title="bbox 347 1575 1052 1610; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_222' title='bbox 347 1575 495 1609; x_wconf 86'>—existing</span> <span class='ocrx_word' id='word_1_223' title='bbox 505 1580 582 1603; x_wconf 89'>notes</span> <span class='ocrx_word' id='word_1_224' title='bbox 593 1584 637 1603; x_wconf 90'>are</span> <span class='ocrx_word' id='word_1_225' title='bbox 648 1580 696 1603; x_wconf 88'>not</span> <span class='ocrx_word' id='word_1_226' title='bbox 707 1576 800 1603; x_wconf 90'>erased</span> <span class='ocrx_word' id='word_1_227' title='bbox 811 1576 890 1603; x_wconf 87'>while</span> <span class='ocrx_word' id='word_1_228' title='bbox 900 1576 1052 1610; x_wconf 87'>recording!</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_18' lang='eng' title="bbox 346 1616 1205 1965">
+     <span class='ocr_line' id='line_1_27' title="bbox 384 1616 1199 1648; baseline 0.001 -6; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_229' title='bbox 384 1616 471 1642; x_wconf 88'>FAST</span> <span class='ocrx_word' id='word_1_230' title='bbox 481 1616 671 1648; x_wconf 88'>FORWARD,</span> <span class='ocrx_word' id='word_1_231' title='bbox 684 1617 844 1648; x_wconf 87'>REWIND,</span> <span class='ocrx_word' id='word_1_232' title='bbox 857 1616 912 1643; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_233' title='bbox 924 1616 1068 1643; x_wconf 78'>LOCATE</span> <span class='ocrx_word' id='word_1_234' title='bbox 1079 1616 1199 1643; x_wconf 87'>controls</span> 
+     </span>
+     <span class='ocr_line' id='line_1_28' title="bbox 346 1655 1202 1689; baseline 0 -7; x_size 34; x_descenders 6; x_ascenders 9"><span class='ocrx_word' id='word_1_235' title='bbox 346 1663 409 1688; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_236' title='bbox 419 1655 453 1682; x_wconf 94'>be</span> <span class='ocrx_word' id='word_1_237' title='bbox 463 1655 530 1682; x_wconf 90'>used</span> <span class='ocrx_word' id='word_1_238' title='bbox 541 1659 569 1682; x_wconf 90'>at</span> <span class='ocrx_word' id='word_1_239' title='bbox 580 1663 632 1688; x_wconf 86'>any</span> <span class='ocrx_word' id='word_1_240' title='bbox 642 1655 707 1683; x_wconf 88'>time</span> <span class='ocrx_word' id='word_1_241' title='bbox 717 1659 746 1683; x_wconf 88'>to</span> <span class='ocrx_word' id='word_1_242' title='bbox 757 1655 867 1689; x_wconf 88'>quickly</span> <span class='ocrx_word' id='word_1_243' title='bbox 878 1664 968 1683; x_wconf 87'>access</span> <span class='ocrx_word' id='word_1_244' title='bbox 979 1664 1032 1689; x_wconf 87'>any</span> <span class='ocrx_word' id='word_1_245' title='bbox 1042 1655 1164 1683; x_wconf 90'>location</span> <span class='ocrx_word' id='word_1_246' title='bbox 1174 1655 1202 1682; x_wconf 94'>in</span> 
+     </span>
+     <span class='ocr_line' id='line_1_29' title="bbox 346 1694 1204 1728; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_247' title='bbox 346 1702 414 1727; x_wconf 88'>your</span> <span class='ocrx_word' id='word_1_248' title='bbox 424 1702 558 1727; x_wconf 87'>sequence</span> <span class='ocrx_word' id='word_1_249' title='bbox 570 1694 612 1721; x_wconf 84'>for</span> <span class='ocrx_word' id='word_1_250' title='bbox 623 1695 847 1728; x_wconf 84'>spot—recording.</span> <span class='ocrx_word' id='word_1_251' title='bbox 860 1696 897 1722; x_wconf 89'>To</span> <span class='ocrx_word' id='word_1_252' title='bbox 908 1695 1028 1722; x_wconf 82'>overdub</span> <span class='ocrx_word' id='word_1_253' title='bbox 1039 1703 1056 1722; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_254' title='bbox 1066 1703 1125 1722; x_wconf 89'>new</span> <span class='ocrx_word' id='word_1_255' title='bbox 1135 1699 1204 1728; x_wconf 88'>part,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_30' title="bbox 347 1733 1150 1768; baseline 0.001 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_256' title='bbox 347 1733 426 1761; x_wconf 88'>select</span> <span class='ocrx_word' id='word_1_257' title='bbox 437 1741 453 1761; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_258' title='bbox 463 1733 589 1761; x_wconf 86'>different</span> <span class='ocrx_word' id='word_1_259' title='bbox 599 1734 674 1761; x_wconf 85'>track</span> <span class='ocrx_word' id='word_1_260' title='bbox 685 1734 741 1761; x_wconf 87'>and</span> <span class='ocrx_word' id='word_1_261' title='bbox 752 1738 818 1761; x_wconf 83'>start</span> <span class='ocrx_word' id='word_1_262' title='bbox 828 1734 1086 1768; x_wconf 63'>recording~while</span> <span class='ocrx_word' id='word_1_263' title='bbox 1096 1742 1150 1767; x_wconf 89'>you</span> 
+     </span>
+     <span class='ocr_line' id='line_1_31' title="bbox 346 1773 1203 1808; baseline 0.001 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_264' title='bbox 346 1774 448 1806; x_wconf 86'>record,</span> <span class='ocrx_word' id='word_1_265' title='bbox 460 1774 506 1801; x_wconf 87'>the</span> <span class='ocrx_word' id='word_1_266' title='bbox 517 1773 662 1801; x_wconf 64'>first‘track</span> <span class='ocrx_word' id='word_1_267' title='bbox 673 1774 726 1801; x_wconf 91'>will</span> <span class='ocrx_word' id='word_1_268' title='bbox 736 1774 799 1807; x_wconf 89'>play</span> <span class='ocrx_word' id='word_1_269' title='bbox 809 1774 836 1801; x_wconf 91'>in</span> <span class='ocrx_word' id='word_1_270' title='bbox 847 1774 949 1808; x_wconf 84'>perfect</span> <span class='ocrx_word' id='word_1_271' title='bbox 961 1782 1026 1808; x_wconf 87'>sync</span> <span class='ocrx_word' id='word_1_272' title='bbox 1037 1774 1137 1808; x_wconf 89'>(unless</span> <span class='ocrx_word' id='word_1_273' title='bbox 1148 1782 1203 1807; x_wconf 93'>you</span> 
+     </span>
+     <span class='ocr_line' id='line_1_32' title="bbox 346 1813 1191 1847; baseline 0.002 -8; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_274' title='bbox 346 1813 454 1840; x_wconf 86'>MUTE</span> <span class='ocrx_word' id='word_1_275' title='bbox 464 1813 492 1845; x_wconf 88'>it,</span> <span class='ocrx_word' id='word_1_276' title='bbox 505 1821 537 1840; x_wconf 84'>or</span> <span class='ocrx_word' id='word_1_277' title='bbox 547 1813 642 1840; x_wconf 88'>SOLO</span> <span class='ocrx_word' id='word_1_278' title='bbox 653 1814 769 1841; x_wconf 87'>another</span> <span class='ocrx_word' id='word_1_279' title='bbox 779 1814 875 1847; x_wconf 86'>track).</span> <span class='ocrx_word' id='word_1_280' title='bbox 889 1814 920 1840; x_wconf 86'>In</span> <span class='ocrx_word' id='word_1_281' title='bbox 930 1813 984 1841; x_wconf 86'>this</span> <span class='ocrx_word' id='word_1_282' title='bbox 995 1822 1057 1847; x_wconf 87'>way,</span> <span class='ocrx_word' id='word_1_283' title='bbox 1069 1822 1106 1847; x_wconf 88'>up</span> <span class='ocrx_word' id='word_1_284' title='bbox 1117 1818 1146 1841; x_wconf 77'>to</span> <span class='ocrx_word' id='word_1_285' title='bbox 1158 1814 1191 1841; x_wconf 90'>32</span> 
+     </span>
+     <span class='ocr_line' id='line_1_33' title="bbox 346 1852 1205 1885; baseline 0.001 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_286' title='bbox 346 1852 434 1879; x_wconf 84'>tracks</span> <span class='ocrx_word' id='word_1_287' title='bbox 445 1860 508 1885; x_wconf 90'>may</span> <span class='ocrx_word' id='word_1_288' title='bbox 518 1853 552 1879; x_wconf 93'>be</span> <span class='ocrx_word' id='word_1_289' title='bbox 562 1853 748 1880; x_wconf 90'>overdubbed!</span> <span class='ocrx_word' id='word_1_290' title='bbox 761 1853 808 1880; x_wconf 92'>All</span> <span class='ocrx_word' id='word_1_291' title='bbox 819 1854 905 1880; x_wconf 81'>MIDI</span> <span class='ocrx_word' id='word_1_292' title='bbox 917 1853 1011 1880; x_wconf 81'>effects</span> <span class='ocrx_word' id='word_1_293' title='bbox 1022 1861 1067 1880; x_wconf 86'>are</span> <span class='ocrx_word' id='word_1_294' title='bbox 1076 1853 1205 1880; x_wconf 87'>recorded</span> 
+     </span>
+     <span class='ocr_line' id='line_1_34' title="bbox 346 1891 1168 1926; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_295' title='bbox 346 1891 485 1925; x_wconf 84'>including</span> <span class='ocrx_word' id='word_1_296' title='bbox 495 1891 570 1925; x_wconf 87'>pitch</span> <span class='ocrx_word' id='word_1_297' title='bbox 580 1892 663 1924; x_wconf 84'>bend,</span> <span class='ocrx_word' id='word_1_298' title='bbox 675 1892 859 1924; x_wconf 86'>modulation,</span> <span class='ocrx_word' id='word_1_299' title='bbox 872 1892 991 1926; x_wconf 83'>velocity,</span> <span class='ocrx_word' id='word_1_300' title='bbox 1004 1892 1168 1924; x_wconf 85'>aftertouch,</span> 
+     </span>
+     <span class='ocr_line' id='line_1_35' title="bbox 346 1931 895 1965; baseline 0.002 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_301' title='bbox 346 1931 448 1958; x_wconf 88'>sustain</span> <span class='ocrx_word' id='word_1_302' title='bbox 459 1931 549 1965; x_wconf 88'>pedal,</span> <span class='ocrx_word' id='word_1_303' title='bbox 561 1932 617 1959; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_304' title='bbox 628 1940 756 1965; x_wconf 88'>program</span> <span class='ocrx_word' id='word_1_305' title='bbox 766 1932 895 1965; x_wconf 87'>changes!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_15' title="bbox 346 1998 1212 2163">
+    <p class='ocr_par' id='par_1_19' lang='eng' title="bbox 346 1998 479 2040">
+     <span class='ocr_line' id='line_1_36' title="bbox 346 1998 479 2040; baseline -0.008 -9; x_size 43; x_descenders 10; x_ascenders 11"><span class='ocrx_word' id='word_1_306' title='bbox 346 1998 479 2040; x_wconf 87'>Editing</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_20' lang='eng' title="bbox 346 2050 1212 2163">
+     <span class='ocr_line' id='line_1_37' title="bbox 383 2050 1186 2084; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_307' title='bbox 383 2050 419 2076; x_wconf 89'>To</span> <span class='ocrx_word' id='word_1_308' title='bbox 430 2057 503 2076; x_wconf 87'>erase</span> <span class='ocrx_word' id='word_1_309' title='bbox 514 2058 530 2077; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_310' title='bbox 540 2058 634 2083; x_wconf 88'>wrong</span> <span class='ocrx_word' id='word_1_311' title='bbox 644 2054 717 2082; x_wconf 84'>note,</span> <span class='ocrx_word' id='word_1_312' title='bbox 729 2050 829 2083; x_wconf 89'>simply</span> <span class='ocrx_word' id='word_1_313' title='bbox 839 2050 905 2077; x_wconf 88'>hold</span> <span class='ocrx_word' id='word_1_314' title='bbox 916 2051 1037 2077; x_wconf 85'>ERASE</span> <span class='ocrx_word' id='word_1_315' title='bbox 1048 2050 1103 2077; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_316' title='bbox 1113 2059 1186 2084; x_wconf 89'>press</span> 
+     </span>
+     <span class='ocr_line' id='line_1_38' title="bbox 346 2089 1212 2124; baseline 0.002 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_317' title='bbox 346 2089 391 2116; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_318' title='bbox 402 2094 465 2117; x_wconf 86'>note</span> <span class='ocrx_word' id='word_1_319' title='bbox 475 2094 504 2117; x_wconf 86'>to</span> <span class='ocrx_word' id='word_1_320' title='bbox 515 2090 549 2117; x_wconf 93'>be</span> <span class='ocrx_word' id='word_1_321' title='bbox 559 2090 652 2117; x_wconf 89'>erased</span> <span class='ocrx_word' id='word_1_322' title='bbox 661 2090 718 2123; x_wconf 91'>just</span> <span class='ocrx_word' id='word_1_323' title='bbox 729 2090 822 2117; x_wconf 85'>before</span> <span class='ocrx_word' id='word_1_324' title='bbox 833 2090 852 2117; x_wconf 89'>it</span> <span class='ocrx_word' id='word_1_325' title='bbox 862 2090 937 2124; x_wconf 87'>plays</span> <span class='ocrx_word' id='word_1_326' title='bbox 947 2090 975 2117; x_wconf 92'>in</span> <span class='ocrx_word' id='word_1_327' title='bbox 986 2090 1032 2118; x_wconf 87'>the</span> <span class='ocrx_word' id='word_1_328' title='bbox 1043 2098 1212 2124; x_wconf 78'>sequence—</span> 
+     </span>
+     <span class='ocr_line' id='line_1_39' title="bbox 346 2129 1134 2163; baseline 0.003 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_329' title='bbox 346 2129 425 2156; x_wconf 92'>when</span> <span class='ocrx_word' id='word_1_330' title='bbox 435 2129 531 2162; x_wconf 91'>played</span> <span class='ocrx_word' id='word_1_331' title='bbox 542 2129 621 2161; x_wconf 82'>back,</span> <span class='ocrx_word' id='word_1_332' title='bbox 633 2129 652 2156; x_wconf 84'>it</span> <span class='ocrx_word' id='word_1_333' title='bbox 663 2129 716 2156; x_wconf 89'>will</span> <span class='ocrx_word' id='word_1_334' title='bbox 726 2130 760 2157; x_wconf 88'>be</span> <span class='ocrx_word' id='word_1_335' title='bbox 771 2138 849 2163; x_wconf 86'>gone.</span> <span class='ocrx_word' id='word_1_336' title='bbox 861 2131 947 2157; x_wconf 86'>Notes</span> <span class='ocrx_word' id='word_1_337' title='bbox 958 2138 1021 2163; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_338' title='bbox 1031 2130 1089 2157; x_wconf 87'>also</span> <span class='ocrx_word' id='word_1_339' title='bbox 1100 2130 1134 2157; x_wconf 91'>be</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_16' title="bbox 345 2168 1228 2242">
+    <p class='ocr_par' id='par_1_21' lang='eng' title="bbox 345 2168 1228 2242">
+     <span class='ocr_line' id='line_1_40' title="bbox 346 2168 1220 2203; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_340' title='bbox 346 2168 445 2200; x_wconf 86'>added,</span> <span class='ocrx_word' id='word_1_341' title='bbox 457 2168 560 2200; x_wconf 87'>erased,</span> <span class='ocrx_word' id='word_1_342' title='bbox 572 2177 604 2196; x_wconf 93'>or</span> <span class='ocrx_word' id='word_1_343' title='bbox 614 2169 739 2202; x_wconf 82'>changed</span> <span class='ocrx_word' id='word_1_344' title='bbox 749 2169 829 2203; x_wconf 86'>using</span> <span class='ocrx_word' id='word_1_345' title='bbox 839 2169 885 2196; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_346' title='bbox 896 2170 1031 2196; x_wconf 89'>SINGLE</span> <span class='ocrx_word' id='word_1_347' title='bbox 1042 2170 1131 2196; x_wconf 87'>STEP</span> <span class='ocrx_word' id='word_1_348' title='bbox 1143 2169 1220 2196; x_wconf 82'>func-</span> 
+     </span>
+     <span class='ocr_line' id='line_1_41' title="bbox 345 2207 1228 2242; baseline 0.002 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_349' title='bbox 345 2207 412 2234; x_wconf 87'>tion.</span> <span class='ocrx_word' id='word_1_350' title='bbox 424 2208 461 2235; x_wconf 88'>To</span> <span class='ocrx_word' id='word_1_351' title='bbox 472 2208 592 2235; x_wconf 88'>overdub</span> <span class='ocrx_word' id='word_1_352' title='bbox 603 2212 680 2235; x_wconf 84'>notes</span> <span class='ocrx_word' id='word_1_353' title='bbox 691 2212 718 2235; x_wconf 89'>at</span> <span class='ocrx_word' id='word_1_354' title='bbox 729 2208 841 2242; x_wconf 83'>specific</span> <span class='ocrx_word' id='word_1_355' title='bbox 851 2209 943 2242; x_wconf 89'>points</span> <span class='ocrx_word' id='word_1_356' title='bbox 955 2208 1049 2236; x_wconf 87'>within</span> <span class='ocrx_word' id='word_1_357' title='bbox 1060 2217 1076 2236; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_358' title='bbox 1086 2216 1228 2242; x_wconf 86'>sequence,</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_17' title="bbox 345 2293 705 2326">
+    <p class='ocr_par' id='par_1_22' lang='eng' title="bbox 345 2293 705 2326">
+     <span class='ocr_line' id='line_1_42' title="bbox 345 2293 705 2326; baseline 0.003 -1; x_size 38.38298; x_descenders 5.3829789; x_ascenders 10"><span class='ocrx_word' id='word_1_359' title='bbox 345 2293 541 2326; x_wconf 86'>Additional</span> <span class='ocrx_word' id='word_1_360' title='bbox 553 2293 705 2326; x_wconf 83'>Features</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_18' title="bbox 1295 1289 2165 1593">
+    <p class='ocr_par' id='par_1_23' lang='eng' title="bbox 1296 1289 2165 1362">
+     <span class='ocr_line' id='line_1_43' title="bbox 1296 1289 2165 1322; baseline 0.001 -7; x_size 35.040108; x_descenders 6.9264708; x_ascenders 8.613636"><span class='ocrx_word' id='word_1_361' title='bbox 1296 1289 1394 1322; x_wconf 91'>simply</span> <span class='ocrx_word' id='word_1_362' title='bbox 1404 1297 1452 1316; x_wconf 87'>use</span> <span class='ocrx_word' id='word_1_363' title='bbox 1463 1290 1615 1321; x_wconf 86'>LOCATE,</span> <span class='ocrx_word' id='word_1_364' title='bbox 1628 1290 1716 1316; x_wconf 88'>FAST</span> <span class='ocrx_word' id='word_1_365' title='bbox 1726 1289 1917 1321; x_wconf 86'>FORWARD,</span> <span class='ocrx_word' id='word_1_366' title='bbox 1929 1297 1960 1317; x_wconf 92'>or</span> <span class='ocrx_word' id='word_1_367' title='bbox 1972 1290 2126 1316; x_wconf 91'>REWIND</span> <span class='ocrx_word' id='word_1_368' title='bbox 2136 1294 2165 1317; x_wconf 82'>to</span> 
+     </span>
+     <span class='ocr_line' id='line_1_44' title="bbox 1297 1328 2033 1362; baseline 0.001 -7; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_369' title='bbox 1297 1328 1356 1355; x_wconf 77'>find</span> <span class='ocrx_word' id='word_1_370' title='bbox 1366 1329 1412 1355; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_371' title='bbox 1423 1328 1527 1356; x_wconf 85'>desired</span> <span class='ocrx_word' id='word_1_372' title='bbox 1537 1329 1587 1356; x_wconf 88'>bar</span> <span class='ocrx_word' id='word_1_373' title='bbox 1598 1329 1720 1361; x_wconf 86'>number,</span> <span class='ocrx_word' id='word_1_374' title='bbox 1732 1329 1798 1356; x_wconf 78'>then</span> <span class='ocrx_word' id='word_1_375' title='bbox 1809 1333 1874 1356; x_wconf 87'>start</span> <span class='ocrx_word' id='word_1_376' title='bbox 1885 1329 2033 1362; x_wconf 85'>recording.</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_24' lang='eng' title="bbox 1295 1368 2160 1593">
+     <span class='ocr_line' id='line_1_45' title="bbox 1332 1368 2160 1402; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_377' title='bbox 1332 1368 1391 1395; x_wconf 90'>The</span> <span class='ocrx_word' id='word_1_378' title='bbox 1402 1368 1650 1395; x_wconf 86'>INSERT/COPY</span> <span class='ocrx_word' id='word_1_379' title='bbox 1662 1368 1789 1395; x_wconf 81'>function</span> <span class='ocrx_word' id='word_1_380' title='bbox 1800 1368 1893 1395; x_wconf 90'>allows</span> <span class='ocrx_word' id='word_1_381' title='bbox 1904 1376 1958 1402; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_382' title='bbox 1968 1373 1997 1395; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_383' title='bbox 2008 1377 2087 1396; x_wconf 91'>move</span> <span class='ocrx_word' id='word_1_384' title='bbox 2097 1369 2160 1396; x_wconf 89'>bars</span> 
+     </span>
+     <span class='ocr_line' id='line_1_46' title="bbox 1297 1407 2151 1441; baseline 0.002 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_385' title='bbox 1297 1407 1369 1434; x_wconf 86'>from</span> <span class='ocrx_word' id='word_1_386' title='bbox 1380 1415 1433 1434; x_wconf 88'>one</span> <span class='ocrx_word' id='word_1_387' title='bbox 1443 1407 1565 1435; x_wconf 85'>location</span> <span class='ocrx_word' id='word_1_388' title='bbox 1576 1411 1605 1434; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_389' title='bbox 1616 1407 1796 1435; x_wconf 85'>another—in</span> <span class='ocrx_word' id='word_1_390' title='bbox 1806 1408 1852 1435; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_391' title='bbox 1863 1416 1937 1435; x_wconf 90'>same</span> <span class='ocrx_word' id='word_1_392' title='bbox 1948 1416 2083 1441; x_wconf 88'>sequence</span> <span class='ocrx_word' id='word_1_393' title='bbox 2093 1416 2125 1435; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_394' title='bbox 2135 1416 2151 1435; x_wconf 88'>a</span> 
+     </span>
+     <span class='ocr_line' id='line_1_47' title="bbox 1296 1447 2160 1481; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_395' title='bbox 1296 1447 1421 1474; x_wconf 81'>different</span> <span class='ocrx_word' id='word_1_396' title='bbox 1431 1455 1491 1474; x_wconf 89'>one.</span> <span class='ocrx_word' id='word_1_397' title='bbox 1504 1448 1555 1474; x_wconf 87'>For</span> <span class='ocrx_word' id='word_1_398' title='bbox 1565 1448 1698 1481; x_wconf 84'>example,</span> <span class='ocrx_word' id='word_1_399' title='bbox 1710 1456 1764 1481; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_400' title='bbox 1775 1448 1863 1481; x_wconf 84'>might</span> <span class='ocrx_word' id='word_1_401' title='bbox 1873 1448 1955 1475; x_wconf 86'>insert</span> <span class='ocrx_word' id='word_1_402' title='bbox 1966 1456 1983 1475; x_wconf 92'>a</span> <span class='ocrx_word' id='word_1_403' title='bbox 1993 1456 2064 1481; x_wconf 88'>copy</span> <span class='ocrx_word' id='word_1_404' title='bbox 2074 1448 2106 1475; x_wconf 90'>of</span> <span class='ocrx_word' id='word_1_405' title='bbox 2115 1448 2160 1475; x_wconf 90'>the</span> 
+     </span>
+     <span class='ocr_line' id='line_1_48' title="bbox 1297 1486 2087 1521; baseline 0 -7; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_406' title='bbox 1297 1486 1355 1514; x_wconf 82'>first</span> <span class='ocrx_word' id='word_1_407' title='bbox 1366 1495 1439 1514; x_wconf 89'>verse</span> <span class='ocrx_word' id='word_1_408' title='bbox 1449 1487 1571 1514; x_wconf 89'>between</span> <span class='ocrx_word' id='word_1_409' title='bbox 1582 1487 1628 1514; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_410' title='bbox 1638 1487 1740 1514; x_wconf 87'>second</span> <span class='ocrx_word' id='word_1_411' title='bbox 1751 1487 1852 1514; x_wconf 83'>chorus</span> <span class='ocrx_word' id='word_1_412' title='bbox 1863 1487 1919 1514; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_413' title='bbox 1929 1487 1975 1514; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_414' title='bbox 1985 1487 2087 1521; x_wconf 85'>bridge.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_49' title="bbox 1296 1527 2047 1560; baseline 0.003 -8; x_size 32; x_descenders 6; x_ascenders 7"><span class='ocrx_word' id='word_1_415' title='bbox 1296 1527 1441 1553; x_wconf 88'>DELETE</span> <span class='ocrx_word' id='word_1_416' title='bbox 1453 1527 1546 1553; x_wconf 90'>BARS</span> <span class='ocrx_word' id='word_1_417' title='bbox 1557 1531 1681 1559; x_wconf 84'>operates</span> <span class='ocrx_word' id='word_1_418' title='bbox 1691 1527 1737 1553; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_419' title='bbox 1748 1535 1823 1554; x_wconf 90'>same</span> <span class='ocrx_word' id='word_1_420' title='bbox 1833 1535 1891 1560; x_wconf 93'>way</span> <span class='ocrx_word' id='word_1_421' title='bbox 1901 1531 1930 1554; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_422' title='bbox 1940 1535 2047 1554; x_wconf 90'>remove</span> 
+     </span>
+     <span class='ocr_line' id='line_1_50' title="bbox 1295 1565 1577 1593; baseline 0.004 -1; x_size 34.748871; x_descenders 6.7488689; x_ascenders 9"><span class='ocrx_word' id='word_1_423' title='bbox 1295 1565 1441 1592; x_wconf 88'>unwanted</span> <span class='ocrx_word' id='word_1_424' title='bbox 1452 1565 1577 1593; x_wconf 84'>sections.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_19' title="bbox 1295 1638 2215 1960">
+    <p class='ocr_par' id='par_1_25' lang='eng' title="bbox 1296 1638 1589 1680">
+     <span class='ocr_line' id='line_1_51' title="bbox 1296 1638 1589 1680; baseline 0 -9; x_size 42; x_descenders 9; x_ascenders 10"><span class='ocrx_word' id='word_1_425' title='bbox 1296 1638 1452 1680; x_wconf 86'>Creating</span> <span class='ocrx_word' id='word_1_426' title='bbox 1465 1648 1485 1671; x_wconf 85'>a</span> <span class='ocrx_word' id='word_1_427' title='bbox 1497 1639 1589 1680; x_wconf 87'>Song</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_26' lang='eng' title="bbox 1295 1690 2215 1960">
+     <span class='ocr_line' id='line_1_52' title="bbox 1333 1690 2146 1723; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_428' title='bbox 1333 1690 1394 1717; x_wconf 88'>One</span> <span class='ocrx_word' id='word_1_429' title='bbox 1404 1698 1462 1723; x_wconf 93'>way</span> <span class='ocrx_word' id='word_1_430' title='bbox 1472 1694 1500 1717; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_431' title='bbox 1511 1694 1598 1717; x_wconf 88'>create</span> <span class='ocrx_word' id='word_1_432' title='bbox 1608 1698 1625 1717; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_433' title='bbox 1635 1698 1704 1723; x_wconf 83'>song</span> <span class='ocrx_word' id='word_1_434' title='bbox 1715 1690 1736 1717; x_wconf 90'>is</span> <span class='ocrx_word' id='word_1_435' title='bbox 1747 1694 1776 1717; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_436' title='bbox 1787 1690 1880 1717; x_wconf 86'>record</span> <span class='ocrx_word' id='word_1_437' title='bbox 1891 1690 1958 1717; x_wconf 89'>each</span> <span class='ocrx_word' id='word_1_438' title='bbox 1969 1690 2044 1717; x_wconf 87'>track</span> <span class='ocrx_word' id='word_1_439' title='bbox 2056 1690 2090 1717; x_wconf 89'>all</span> <span class='ocrx_word' id='word_1_440' title='bbox 2101 1690 2146 1717; x_wconf 88'>the</span> 
+     </span>
+     <span class='ocr_line' id='line_1_53' title="bbox 1295 1729 2121 1762; baseline 0.001 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_441' title='bbox 1295 1737 1353 1762; x_wconf 87'>way</span> <span class='ocrx_word' id='word_1_442' title='bbox 1362 1729 1481 1762; x_wconf 83'>through</span> <span class='ocrx_word' id='word_1_443' title='bbox 1493 1729 1541 1762; x_wconf 90'>(up</span> <span class='ocrx_word' id='word_1_444' title='bbox 1552 1733 1581 1756; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_445' title='bbox 1592 1730 1644 1756; x_wconf 89'>999</span> <span class='ocrx_word' id='word_1_446' title='bbox 1654 1729 1738 1762; x_wconf 90'>bars).</span> <span class='ocrx_word' id='word_1_447' title='bbox 1751 1729 1878 1757; x_wconf 89'>Another</span> <span class='ocrx_word' id='word_1_448' title='bbox 1888 1737 1945 1762; x_wconf 89'>way</span> <span class='ocrx_word' id='word_1_449' title='bbox 1956 1729 1977 1757; x_wconf 90'>is</span> <span class='ocrx_word' id='word_1_450' title='bbox 1987 1733 2016 1757; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_451' title='bbox 2027 1729 2121 1757; x_wconf 86'>record</span> 
+     </span>
+     <span class='ocr_line' id='line_1_54' title="bbox 1296 1768 2066 1802; baseline 0 -6; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_452' title='bbox 1296 1768 1362 1796; x_wconf 86'>each</span> <span class='ocrx_word' id='word_1_453' title='bbox 1373 1768 1448 1796; x_wconf 87'>basic</span> <span class='ocrx_word' id='word_1_454' title='bbox 1458 1768 1562 1796; x_wconf 83'>section</span> <span class='ocrx_word' id='word_1_455' title='bbox 1574 1769 1666 1802; x_wconf 89'>(verse,</span> <span class='ocrx_word' id='word_1_456' title='bbox 1679 1769 1788 1801; x_wconf 86'>chorus,</span> <span class='ocrx_word' id='word_1_457' title='bbox 1800 1769 1865 1802; x_wconf 83'>etc.)</span> <span class='ocrx_word' id='word_1_458' title='bbox 1876 1768 1904 1795; x_wconf 91'>in</span> <span class='ocrx_word' id='word_1_459' title='bbox 1914 1768 2066 1796; x_wconf 86'>individual</span> 
+     </span>
+     <span class='ocr_line' id='line_1_55' title="bbox 1296 1808 2215 1841; baseline 0 -6; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_460' title='bbox 1296 1816 1451 1841; x_wconf 85'>sequences,</span> <span class='ocrx_word' id='word_1_461' title='bbox 1463 1808 1528 1835; x_wconf 86'>then</span> <span class='ocrx_word' id='word_1_462' title='bbox 1538 1816 1587 1835; x_wconf 90'>use</span> <span class='ocrx_word' id='word_1_463' title='bbox 1597 1808 1643 1835; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_464' title='bbox 1653 1809 1799 1835; x_wconf 89'>CREATE</span> <span class='ocrx_word' id='word_1_465' title='bbox 1810 1808 1911 1835; x_wconf 89'>SONG</span> <span class='ocrx_word' id='word_1_466' title='bbox 1923 1808 2050 1836; x_wconf 85'>function</span> <span class='ocrx_word' id='word_1_467' title='bbox 2060 1812 2089 1835; x_wconf 89'>to</span> <span class='ocrx_word' id='word_1_468' title='bbox 2103 1808 2215 1836; x_wconf 87'>“chain”</span> 
+     </span>
+     <span class='ocr_line' id='line_1_56' title="bbox 1295 1847 2135 1881; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_469' title='bbox 1295 1848 1370 1874; x_wconf 91'>them</span> <span class='ocrx_word' id='word_1_470' title='bbox 1381 1848 1508 1881; x_wconf 83'>together.</span> <span class='ocrx_word' id='word_1_471' title='bbox 1521 1848 1667 1875; x_wconf 87'>CREATE</span> <span class='ocrx_word' id='word_1_472' title='bbox 1678 1848 1779 1875; x_wconf 91'>SONG</span> <span class='ocrx_word' id='word_1_473' title='bbox 1789 1847 1842 1874; x_wconf 92'>will</span> <span class='ocrx_word' id='word_1_474' title='bbox 1853 1848 1918 1875; x_wconf 86'>then</span> <span class='ocrx_word' id='word_1_475' title='bbox 1929 1848 2135 1881; x_wconf 81'>automatically</span> 
+     </span>
+     <span class='ocr_line' id='line_1_57' title="bbox 1296 1887 2162 1920; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_476' title='bbox 1296 1895 1366 1920; x_wconf 89'>copy</span> <span class='ocrx_word' id='word_1_477' title='bbox 1377 1887 1412 1914; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_478' title='bbox 1422 1887 1468 1914; x_wconf 85'>the</span> <span class='ocrx_word' id='word_1_479' title='bbox 1478 1891 1552 1920; x_wconf 86'>parts</span> <span class='ocrx_word' id='word_1_480' title='bbox 1563 1887 1621 1914; x_wconf 86'>into</span> <span class='ocrx_word' id='word_1_481' title='bbox 1632 1895 1649 1914; x_wconf 91'>a</span> <span class='ocrx_word' id='word_1_482' title='bbox 1659 1895 1718 1914; x_wconf 91'>new</span> <span class='ocrx_word' id='word_1_483' title='bbox 1729 1895 1870 1920; x_wconf 83'>sequence.</span> <span class='ocrx_word' id='word_1_484' title='bbox 1884 1887 1911 1914; x_wconf 84'>If</span> <span class='ocrx_word' id='word_1_485' title='bbox 1919 1887 2033 1919; x_wconf 89'>desired,</span> <span class='ocrx_word' id='word_1_486' title='bbox 2045 1895 2099 1920; x_wconf 92'>you</span> <span class='ocrx_word' id='word_1_487' title='bbox 2110 1895 2162 1914; x_wconf 85'>can</span> 
+     </span>
+     <span class='ocr_line' id='line_1_58' title="bbox 1296 1926 2169 1960; baseline 0.001 -7; x_size 34; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_488' title='bbox 1296 1934 1361 1953; x_wconf 89'>even</span> <span class='ocrx_word' id='word_1_489' title='bbox 1372 1930 1411 1953; x_wconf 90'>set</span> <span class='ocrx_word' id='word_1_490' title='bbox 1422 1926 1468 1953; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_491' title='bbox 1478 1926 1528 1953; x_wconf 89'>last</span> <span class='ocrx_word' id='word_1_492' title='bbox 1540 1926 1590 1953; x_wconf 88'>few</span> <span class='ocrx_word' id='word_1_493' title='bbox 1601 1927 1664 1953; x_wconf 89'>bars</span> <span class='ocrx_word' id='word_1_494' title='bbox 1675 1931 1704 1954; x_wconf 85'>to</span> <span class='ocrx_word' id='word_1_495' title='bbox 1715 1931 1806 1960; x_wconf 89'>repeat</span> <span class='ocrx_word' id='word_1_496' title='bbox 1816 1926 1955 1960; x_wconf 84'>infinitely,</span> <span class='ocrx_word' id='word_1_497' title='bbox 1968 1926 2011 1954; x_wconf 86'>for</span> <span class='ocrx_word' id='word_1_498' title='bbox 2022 1935 2038 1954; x_wconf 92'>a</span> <span class='ocrx_word' id='word_1_499' title='bbox 2049 1927 2169 1954; x_wconf 83'>fadeout.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_20' title="bbox 1293 2000 2179 2242">
+    <p class='ocr_par' id='par_1_27' lang='eng' title="bbox 1294 2000 1948 2040">
+     <span class='ocr_line' id='line_1_59' title="bbox 1294 2000 1948 2040; baseline 0.002 -8; x_size 40; x_descenders 8; x_ascenders 10"><span class='ocrx_word' id='word_1_500' title='bbox 1294 2000 1532 2040; x_wconf 84'>Composition</span> <span class='ocrx_word' id='word_1_501' title='bbox 1544 2000 1699 2033; x_wconf 85'>Without</span> <span class='ocrx_word' id='word_1_502' title='bbox 1711 2000 1948 2040; x_wconf 85'>Compromise</span> 
+     </span>
+    </p>
+
+    <p class='ocr_par' id='par_1_28' lang='eng' title="bbox 1293 2051 2179 2242">
+     <span class='ocr_line' id='line_1_60' title="bbox 1331 2051 2179 2085; baseline 0 -7; x_size 34; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_503' title='bbox 1331 2051 1389 2078; x_wconf 90'>The</span> <span class='ocrx_word' id='word_1_504' title='bbox 1399 2051 1563 2085; x_wconf 83'>technology</span> <span class='ocrx_word' id='word_1_505' title='bbox 1573 2059 1627 2084; x_wconf 92'>you</span> <span class='ocrx_word' id='word_1_506' title='bbox 1638 2060 1686 2079; x_wconf 86'>use</span> <span class='ocrx_word' id='word_1_507' title='bbox 1697 2051 1797 2078; x_wconf 89'>should</span> <span class='ocrx_word' id='word_1_508' title='bbox 1808 2059 1887 2078; x_wconf 87'>never</span> <span class='ocrx_word' id='word_1_509' title='bbox 1897 2052 1932 2079; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_510' title='bbox 1942 2059 1973 2079; x_wconf 89'>so</span> <span class='ocrx_word' id='word_1_511' title='bbox 1984 2051 2110 2085; x_wconf 85'>complex</span> <span class='ocrx_word' id='word_1_512' title='bbox 2120 2052 2179 2078; x_wconf 86'>that</span> 
+     </span>
+     <span class='ocr_line' id='line_1_61' title="bbox 1294 2090 2157 2124; baseline 0 -6; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_513' title='bbox 1294 2090 1313 2118; x_wconf 86'>it</span> <span class='ocrx_word' id='word_1_514' title='bbox 1323 2091 1459 2118; x_wconf 87'>interferes</span> <span class='ocrx_word' id='word_1_515' title='bbox 1470 2091 1535 2118; x_wconf 84'>with</span> <span class='ocrx_word' id='word_1_516' title='bbox 1545 2091 1591 2118; x_wconf 88'>the</span> <span class='ocrx_word' id='word_1_517' title='bbox 1602 2091 1715 2118; x_wconf 90'>creative</span> <span class='ocrx_word' id='word_1_518' title='bbox 1725 2099 1841 2124; x_wconf 83'>process.</span> <span class='ocrx_word' id='word_1_519' title='bbox 1854 2091 1947 2118; x_wconf 78'>That’s</span> <span class='ocrx_word' id='word_1_520' title='bbox 1957 2091 2086 2124; x_wconf 86'>precisely</span> <span class='ocrx_word' id='word_1_521' title='bbox 2096 2091 2157 2124; x_wconf 89'>why</span> 
+     </span>
+     <span class='ocr_line' id='line_1_62' title="bbox 1293 2130 2156 2164; baseline 0 -7; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_522' title='bbox 1293 2130 1339 2157; x_wconf 90'>the</span> <span class='ocrx_word' id='word_1_523' title='bbox 1350 2130 1576 2164; x_wconf 80'>LinnSequencer</span> <span class='ocrx_word' id='word_1_524' title='bbox 1586 2130 1607 2157; x_wconf 89'>is</span> <span class='ocrx_word' id='word_1_525' title='bbox 1619 2130 1747 2164; x_wconf 87'>designed</span> <span class='ocrx_word' id='word_1_526' title='bbox 1758 2134 1787 2157; x_wconf 87'>to</span> <span class='ocrx_word' id='word_1_527' title='bbox 1798 2130 1834 2157; x_wconf 89'>let</span> <span class='ocrx_word' id='word_1_528' title='bbox 1844 2138 1899 2163; x_wconf 93'>you</span> <span class='ocrx_word' id='word_1_529' title='bbox 1910 2138 2051 2164; x_wconf 88'>compose,</span> <span class='ocrx_word' id='word_1_530' title='bbox 2063 2130 2156 2157; x_wconf 87'>record</span> 
+     </span>
+     <span class='ocr_line' id='line_1_63' title="bbox 1294 2169 2145 2203; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_531' title='bbox 1294 2170 1349 2197; x_wconf 90'>and</span> <span class='ocrx_word' id='word_1_532' title='bbox 1360 2169 1414 2197; x_wconf 83'>edit</span> <span class='ocrx_word' id='word_1_533' title='bbox 1425 2170 1503 2197; x_wconf 89'>while</span> <span class='ocrx_word' id='word_1_534' title='bbox 1514 2170 1643 2203; x_wconf 90'>devoting</span> <span class='ocrx_word' id='word_1_535' title='bbox 1653 2178 1721 2203; x_wconf 92'>your</span> <span class='ocrx_word' id='word_1_536' title='bbox 1731 2170 1882 2197; x_wconf 88'>undivided</span> <span class='ocrx_word' id='word_1_537' title='bbox 1893 2170 2027 2197; x_wconf 86'>attention</span> <span class='ocrx_word' id='word_1_538' title='bbox 2037 2174 2067 2197; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_539' title='bbox 2077 2178 2145 2203; x_wconf 90'>your</span> 
+     </span>
+     <span class='ocr_line' id='line_1_64' title="bbox 1294 2209 2126 2242; baseline 0 -6; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_540' title='bbox 1294 2209 1388 2236; x_wconf 88'>music.</span> <span class='ocrx_word' id='word_1_541' title='bbox 1402 2209 1451 2236; x_wconf 89'>See</span> <span class='ocrx_word' id='word_1_542' title='bbox 1460 2217 1529 2242; x_wconf 92'>your</span> <span class='ocrx_word' id='word_1_543' title='bbox 1540 2209 1611 2236; x_wconf 87'>Linn</span> <span class='ocrx_word' id='word_1_544' title='bbox 1621 2209 1712 2236; x_wconf 86'>dealer</span> <span class='ocrx_word' id='word_1_545' title='bbox 1722 2209 1806 2242; x_wconf 85'>today</span> <span class='ocrx_word' id='word_1_546' title='bbox 1817 2209 1859 2236; x_wconf 87'>for</span> <span class='ocrx_word' id='word_1_547' title='bbox 1870 2217 1887 2236; x_wconf 88'>a</span> <span class='ocrx_word' id='word_1_548' title='bbox 1897 2209 2126 2236; x_wconf 76'>demonstration!</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_21' title="bbox 347 2343 2171 2378">
+    <p class='ocr_par' id='par_1_29' lang='eng' title="bbox 347 2343 2171 2378">
+     <span class='ocr_line' id='line_1_65' title="bbox 347 2343 2171 2378; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_549' title='bbox 347 2350 361 2363; x_wconf 69'><strong>0</strong></span> <span class='ocrx_word' id='word_1_550' title='bbox 373 2343 483 2377; x_wconf 83'>Simple,</span> <span class='ocrx_word' id='word_1_551' title='bbox 495 2352 559 2377; x_wconf 89'>easy</span> <span class='ocrx_word' id='word_1_552' title='bbox 569 2348 598 2371; x_wconf 91'>to</span> <span class='ocrx_word' id='word_1_553' title='bbox 609 2344 683 2371; x_wconf 89'>learn</span> <span class='ocrx_word' id='word_1_554' title='bbox 694 2344 922 2378; x_wconf 80'>operation—the</span> <span class='ocrx_word' id='word_1_555' title='bbox 934 2345 967 2372; x_wconf 88'>32</span> <span class='ocrx_word' id='word_1_556' title='bbox 978 2345 1116 2372; x_wconf 86'>character</span> <span class='ocrx_word' id='word_1_557' title='bbox 1126 2345 1200 2372; x_wconf 92'>LCD</span> <span class='ocrx_word' id='word_1_558' title='bbox 1211 2345 1316 2378; x_wconf 90'>display</span> <span class='ocrx_word' id='word_1_559' title='bbox 1326 2345 1424 2378; x_wconf 88'>clearly</span> <span class='ocrx_word' id='word_1_560' title='bbox 1434 2345 1528 2378; x_wconf 86'>guides</span> <span class='ocrx_word' id='word_1_561' title='bbox 1539 2353 1594 2378; x_wconf 91'>you</span> <span class='ocrx_word' id='word_1_562' title='bbox 1604 2345 1724 2378; x_wconf 87'>through</span> <span class='ocrx_word' id='word_1_563' title='bbox 1735 2344 1770 2371; x_wconf 88'>all</span> <span class='ocrx_word' id='word_1_564' title='bbox 1781 2344 1947 2377; x_wconf 86'>operations.</span> <span class='ocrx_word' id='word_1_565' title='bbox 1961 2344 1989 2371; x_wconf 88'>If</span> <span class='ocrx_word' id='word_1_566' title='bbox 1997 2344 2112 2376; x_wconf 88'>needed,</span> <span class='ocrx_word' id='word_1_567' title='bbox 2125 2345 2171 2371; x_wconf 88'>the</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_22' title="bbox 373 2381 1083 2415">
+    <p class='ocr_par' id='par_1_30' lang='eng' title="bbox 373 2381 1083 2415">
+     <span class='ocr_line' id='line_1_66' title="bbox 373 2381 1083 2415; baseline 0.003 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_568' title='bbox 373 2381 472 2407; x_wconf 85'>HELP</span> <span class='ocrx_word' id='word_1_569' title='bbox 483 2381 583 2408; x_wconf 85'>button</span> <span class='ocrx_word' id='word_1_570' title='bbox 594 2381 711 2415; x_wconf 87'>displays</span> <span class='ocrx_word' id='word_1_571' title='bbox 722 2382 875 2409; x_wconf 89'>additional</span> <span class='ocrx_word' id='word_1_572' title='bbox 886 2382 1083 2415; x_wconf 87'>explanations.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_23' title="bbox 347 2427 2145 2507">
+    <p class='ocr_par' id='par_1_31' lang='eng' title="bbox 347 2427 2145 2507">
+     <span class='ocr_line' id='line_1_67' title="bbox 347 2427 1468 2461; baseline 0.002 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_573' title='bbox 347 2432 361 2446; x_wconf 69'><strong>0</strong></span> <span class='ocrx_word' id='word_1_574' title='bbox 373 2427 612 2454; x_wconf 88'>Non-destructive</span> <span class='ocrx_word' id='word_1_575' title='bbox 622 2427 914 2461; x_wconf 85'>recording—existing</span> <span class='ocrx_word' id='word_1_576' title='bbox 924 2432 1002 2455; x_wconf 90'>notes</span> <span class='ocrx_word' id='word_1_577' title='bbox 1013 2436 1057 2455; x_wconf 88'>are</span> <span class='ocrx_word' id='word_1_578' title='bbox 1068 2432 1116 2455; x_wconf 84'>not</span> <span class='ocrx_word' id='word_1_579' title='bbox 1127 2428 1220 2455; x_wconf 87'>erased</span> <span class='ocrx_word' id='word_1_580' title='bbox 1231 2428 1309 2455; x_wconf 91'>while</span> <span class='ocrx_word' id='word_1_581' title='bbox 1319 2428 1468 2461; x_wconf 87'>recording.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_68' title="bbox 347 2473 2145 2507; baseline 0.001 -8; x_size 35; x_descenders 7; x_ascenders 9"><span class='ocrx_word' id='word_1_582' title='bbox 347 2478 361 2492; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_583' title='bbox 372 2473 433 2500; x_wconf 90'>Two</span> <span class='ocrx_word' id='word_1_584' title='bbox 444 2473 689 2500; x_wconf 89'>FOOTSWITCH</span> <span class='ocrx_word' id='word_1_585' title='bbox 701 2474 837 2501; x_wconf 79'>INPUTS</span> <span class='ocrx_word' id='word_1_586' title='bbox 848 2481 910 2507; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_587' title='bbox 921 2473 955 2500; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_588' title='bbox 966 2473 1091 2507; x_wconf 85'>assigned</span> <span class='ocrx_word' id='word_1_589' title='bbox 1101 2478 1130 2501; x_wconf 81'>to</span> <span class='ocrx_word' id='word_1_590' title='bbox 1141 2474 1271 2507; x_wconf 84'>remotely</span> <span class='ocrx_word' id='word_1_591' title='bbox 1281 2473 1387 2501; x_wconf 86'>control</span> <span class='ocrx_word' id='word_1_592' title='bbox 1398 2482 1481 2507; x_wconf 86'>many</span> <span class='ocrx_word' id='word_1_593' title='bbox 1491 2473 1523 2501; x_wconf 90'>of</span> <span class='ocrx_word' id='word_1_594' title='bbox 1531 2474 1578 2501; x_wconf 89'>the</span> <span class='ocrx_word' id='word_1_595' title='bbox 1588 2473 1750 2506; x_wconf 82'>commonly</span> <span class='ocrx_word' id='word_1_596' title='bbox 1760 2473 1827 2500; x_wconf 88'>used</span> <span class='ocrx_word' id='word_1_597' title='bbox 1840 2473 1991 2505; x_wconf 77'>functions,</span> <span class='ocrx_word' id='word_1_598' title='bbox 2003 2473 2145 2507; x_wconf 84'>including</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_24' title="bbox 372 2510 1090 2543">
+    <p class='ocr_par' id='par_1_32' lang='eng' title="bbox 372 2510 1090 2543">
+     <span class='ocr_line' id='line_1_69' title="bbox 372 2510 1090 2543; baseline 0.001 -6; x_size 35.625; x_descenders 8.90625; x_ascenders 8.90625"><span class='ocrx_word' id='word_1_599' title='bbox 372 2510 500 2542; x_wconf 88'>ERASE,</span> <span class='ocrx_word' id='word_1_600' title='bbox 513 2511 660 2542; x_wconf 86'>REPEAT,</span> <span class='ocrx_word' id='word_1_601' title='bbox 673 2511 883 2543; x_wconf 80'>PLAY/STOP,</span> <span class='ocrx_word' id='word_1_602' title='bbox 896 2519 927 2538; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_603' title='bbox 939 2511 1090 2538; x_wconf 83'>LOCATE.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_25' title="bbox 347 2556 1768 2590">
+    <p class='ocr_par' id='par_1_33' lang='eng' title="bbox 347 2556 1768 2590">
+     <span class='ocr_line' id='line_1_70' title="bbox 347 2556 1768 2590; baseline 0.001 -8; x_size 34; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_604' title='bbox 347 2561 361 2575; x_wconf 71'><strong>0</strong></span> <span class='ocrx_word' id='word_1_605' title='bbox 372 2556 433 2583; x_wconf 91'>Two</span> <span class='ocrx_word' id='word_1_606' title='bbox 443 2556 612 2583; x_wconf 83'>TRIGGER</span> <span class='ocrx_word' id='word_1_607' title='bbox 623 2556 797 2584; x_wconf 84'>OUTPUTS</span> <span class='ocrx_word' id='word_1_608' title='bbox 808 2565 871 2590; x_wconf 89'>may</span> <span class='ocrx_word' id='word_1_609' title='bbox 881 2557 915 2584; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_610' title='bbox 925 2557 1119 2590; x_wconf 84'>programmed</span> <span class='ocrx_word' id='word_1_611' title='bbox 1129 2561 1158 2584; x_wconf 93'>to</span> <span class='ocrx_word' id='word_1_612' title='bbox 1169 2561 1270 2590; x_wconf 87'>output</span> <span class='ocrx_word' id='word_1_613' title='bbox 1281 2557 1370 2590; x_wconf 88'>pulses</span> <span class='ocrx_word' id='word_1_614' title='bbox 1381 2561 1409 2584; x_wconf 92'>at</span> <span class='ocrx_word' id='word_1_615' title='bbox 1419 2565 1472 2590; x_wconf 90'>any</span> <span class='ocrx_word' id='word_1_616' title='bbox 1483 2557 1598 2584; x_wconf 86'>selected</span> <span class='ocrx_word' id='word_1_617' title='bbox 1608 2561 1673 2584; x_wconf 89'>note</span> <span class='ocrx_word' id='word_1_618' title='bbox 1683 2556 1768 2583; x_wconf 89'>value.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_26' title="bbox 347 2601 1226 2635">
+    <p class='ocr_par' id='par_1_34' lang='eng' title="bbox 347 2601 1226 2635">
+     <span class='ocr_line' id='line_1_71' title="bbox 347 2601 1226 2635; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_619' title='bbox 347 2607 361 2620; x_wconf 65'><strong>0</strong></span> <span class='ocrx_word' id='word_1_620' title='bbox 372 2601 434 2628; x_wconf 88'>Will</span> <span class='ocrx_word' id='word_1_621' title='bbox 445 2609 510 2634; x_wconf 89'>sync</span> <span class='ocrx_word' id='word_1_622' title='bbox 521 2605 549 2628; x_wconf 84'>to</span> <span class='ocrx_word' id='word_1_623' title='bbox 561 2602 690 2629; x_wconf 88'>standard</span> <span class='ocrx_word' id='word_1_624' title='bbox 701 2602 864 2629; x_wconf 86'>LinnDrum</span> <span class='ocrx_word' id='word_1_625' title='bbox 875 2610 907 2629; x_wconf 91'>or</span> <span class='ocrx_word' id='word_1_626' title='bbox 918 2602 989 2629; x_wconf 89'>Linn</span> <span class='ocrx_word' id='word_1_627' title='bbox 1000 2603 1069 2629; x_wconf 90'>9000</span> <span class='ocrx_word' id='word_1_628' title='bbox 1080 2610 1145 2635; x_wconf 87'>sync</span> <span class='ocrx_word' id='word_1_629' title='bbox 1155 2607 1226 2630; x_wconf 89'>tone.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_27' title="bbox 347 2648 2100 2727">
+    <p class='ocr_par' id='par_1_35' lang='eng' title="bbox 347 2648 2100 2727">
+     <span class='ocr_line' id='line_1_72' title="bbox 347 2648 1664 2682; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_630' title='bbox 347 2654 360 2667; x_wconf 75'>&#39;</span> <span class='ocrx_word' id='word_1_631' title='bbox 372 2648 483 2675; x_wconf 89'>Utilizes</span> <span class='ocrx_word' id='word_1_632' title='bbox 493 2648 564 2680; x_wconf 75'>ultra</span> <span class='ocrx_word' id='word_1_633' title='bbox 573 2648 744 2682; x_wconf 86'>high—speed,</span> <span class='ocrx_word' id='word_1_634' title='bbox 757 2649 772 2676; x_wconf 91'>8</span> <span class='ocrx_word' id='word_1_635' title='bbox 783 2649 862 2675; x_wconf 89'>MHz</span> <span class='ocrx_word' id='word_1_636' title='bbox 873 2649 954 2676; x_wconf 88'>80186</span> <span class='ocrx_word' id='word_1_637' title='bbox 965 2649 994 2676; x_wconf 91'>16</span> <span class='ocrx_word' id='word_1_638' title='bbox 1004 2648 1043 2676; x_wconf 90'>bit</span> <span class='ocrx_word' id='word_1_639' title='bbox 1054 2653 1197 2682; x_wconf 88'>computer</span> <span class='ocrx_word' id='word_1_640' title='bbox 1208 2649 1350 2682; x_wconf 90'>internally</span> <span class='ocrx_word' id='word_1_641' title='bbox 1361 2649 1403 2676; x_wconf 87'>for</span> <span class='ocrx_word' id='word_1_642' title='bbox 1414 2649 1502 2676; x_wconf 87'>FAST</span> <span class='ocrx_word' id='word_1_643' title='bbox 1512 2648 1664 2682; x_wconf 81'>operation.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_73' title="bbox 347 2694 2100 2727; baseline 0.001 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_644' title='bbox 347 2699 361 2713; x_wconf 70'><strong>0</strong></span> <span class='ocrx_word' id='word_1_645' title='bbox 372 2694 504 2721; x_wconf 88'>TEMPO</span> <span class='ocrx_word' id='word_1_646' title='bbox 515 2702 578 2727; x_wconf 92'>may</span> <span class='ocrx_word' id='word_1_647' title='bbox 589 2694 623 2721; x_wconf 90'>be</span> <span class='ocrx_word' id='word_1_648' title='bbox 633 2694 764 2727; x_wconf 87'>specified</span> <span class='ocrx_word' id='word_1_649' title='bbox 774 2694 802 2721; x_wconf 89'>in</span> <span class='ocrx_word' id='word_1_650' title='bbox 814 2695 1172 2722; x_wconf 84'>BEATS-PER—MINUTE</span> <span class='ocrx_word' id='word_1_651' title='bbox 1183 2703 1215 2722; x_wconf 93'>or</span> <span class='ocrx_word' id='word_1_652' title='bbox 1225 2695 1567 2722; x_wconf 84'>FRAMES-PER—BEAT</span> <span class='ocrx_word' id='word_1_653' title='bbox 1577 2698 1605 2721; x_wconf 89'>at</span> <span class='ocrx_word' id='word_1_654' title='bbox 1616 2695 1659 2726; x_wconf 91'>24,</span> <span class='ocrx_word' id='word_1_655' title='bbox 1672 2695 1716 2726; x_wconf 88'>25,</span> <span class='ocrx_word' id='word_1_656' title='bbox 1728 2702 1760 2721; x_wconf 90'>or</span> <span class='ocrx_word' id='word_1_657' title='bbox 1772 2695 1806 2721; x_wconf 88'>30</span> <span class='ocrx_word' id='word_1_658' title='bbox 1817 2694 1917 2721; x_wconf 86'>frames</span> <span class='ocrx_word' id='word_1_659' title='bbox 1928 2702 1976 2727; x_wconf 89'>per</span> <span class='ocrx_word' id='word_1_660' title='bbox 1987 2694 2100 2726; x_wconf 87'>second,</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_28' title="bbox 373 2731 652 2765">
+    <p class='ocr_par' id='par_1_36' lang='eng' title="bbox 373 2731 652 2765">
+     <span class='ocr_line' id='line_1_74' title="bbox 373 2731 652 2765; baseline 0 -7; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_661' title='bbox 373 2731 449 2764; x_wconf 87'>(even</span> <span class='ocrx_word' id='word_1_662' title='bbox 460 2731 529 2765; x_wconf 88'>drop</span> <span class='ocrx_word' id='word_1_663' title='bbox 541 2731 652 2764; x_wconf 81'>frame!)</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_29' title="bbox 347 2777 2174 2811">
+    <p class='ocr_par' id='par_1_37' lang='eng' title="bbox 347 2777 2174 2811">
+     <span class='ocr_line' id='line_1_75' title="bbox 347 2777 2174 2811; baseline 0.001 -8; x_size 33; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_664' title='bbox 347 2782 360 2796; x_wconf 72'><strong>0</strong></span> <span class='ocrx_word' id='word_1_665' title='bbox 372 2777 504 2804; x_wconf 82'>TEMPO</span> <span class='ocrx_word' id='word_1_666' title='bbox 515 2785 578 2810; x_wconf 91'>may</span> <span class='ocrx_word' id='word_1_667' title='bbox 588 2777 622 2804; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_668' title='bbox 633 2778 741 2804; x_wconf 84'>entered</span> <span class='ocrx_word' id='word_1_669' title='bbox 751 2777 934 2811; x_wconf 80'>numerically,</span> <span class='ocrx_word' id='word_1_670' title='bbox 946 2777 1101 2811; x_wconf 85'>adjustable</span> <span class='ocrx_word' id='word_1_671' title='bbox 1111 2777 1139 2804; x_wconf 92'>in</span> <span class='ocrx_word' id='word_1_672' title='bbox 1149 2778 1239 2805; x_wconf 88'>tenths</span> <span class='ocrx_word' id='word_1_673' title='bbox 1250 2778 1282 2805; x_wconf 85'>of</span> <span class='ocrx_word' id='word_1_674' title='bbox 1290 2786 1307 2805; x_wconf 94'>a</span> <span class='ocrx_word' id='word_1_675' title='bbox 1317 2777 1567 2805; x_wconf 85'>Beat-Per-Minute</span> <span class='ocrx_word' id='word_1_676' title='bbox 1577 2777 1748 2809; x_wconf 85'>increments,</span> <span class='ocrx_word' id='word_1_677' title='bbox 1760 2785 1792 2804; x_wconf 87'>or</span> <span class='ocrx_word' id='word_1_678' title='bbox 1803 2777 1839 2810; x_wconf 94'>by</span> <span class='ocrx_word' id='word_1_679' title='bbox 1849 2777 1964 2811; x_wconf 84'>tapping</span> <span class='ocrx_word' id='word_1_680' title='bbox 1975 2781 2086 2810; x_wconf 85'>quarter</span> <span class='ocrx_word' id='word_1_681' title='bbox 2097 2781 2174 2804; x_wconf 85'>notes</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_30' title="bbox 372 2815 808 2842">
+    <p class='ocr_par' id='par_1_38' lang='eng' title="bbox 372 2815 808 2842">
+     <span class='ocr_line' id='line_1_76' title="bbox 372 2815 808 2842; baseline 0.002 -1; x_size 32.351353; x_descenders 5.3513513; x_ascenders 9"><span class='ocrx_word' id='word_1_682' title='bbox 372 2822 410 2841; x_wconf 93'>on</span> <span class='ocrx_word' id='word_1_683' title='bbox 420 2815 466 2842; x_wconf 86'>the</span> <span class='ocrx_word' id='word_1_684' title='bbox 476 2815 545 2841; x_wconf 91'>TAP</span> <span class='ocrx_word' id='word_1_685' title='bbox 556 2815 689 2842; x_wconf 85'>TEMPO</span> <span class='ocrx_word' id='word_1_686' title='bbox 699 2815 808 2842; x_wconf 83'>button.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_31' title="bbox 347 2861 1792 2940">
+    <p class='ocr_par' id='par_1_39' lang='eng' title="bbox 347 2861 1792 2940">
+     <span class='ocr_line' id='line_1_77' title="bbox 347 2861 1792 2895; baseline 0.001 -8; x_size 33; x_descenders 6; x_ascenders 8"><span class='ocrx_word' id='word_1_687' title='bbox 347 2866 360 2880; x_wconf 68'><strong>0</strong></span> <span class='ocrx_word' id='word_1_688' title='bbox 372 2861 504 2887; x_wconf 87'>TEMPO</span> <span class='ocrx_word' id='word_1_689' title='bbox 515 2861 696 2888; x_wconf 86'>CHANGES</span> <span class='ocrx_word' id='word_1_690' title='bbox 707 2869 771 2894; x_wconf 92'>may</span> <span class='ocrx_word' id='word_1_691' title='bbox 781 2861 815 2888; x_wconf 92'>be</span> <span class='ocrx_word' id='word_1_692' title='bbox 825 2861 1019 2894; x_wconf 85'>programmed</span> <span class='ocrx_word' id='word_1_693' title='bbox 1030 2861 1087 2888; x_wconf 88'>into</span> <span class='ocrx_word' id='word_1_694' title='bbox 1099 2869 1115 2888; x_wconf 87'>a</span> <span class='ocrx_word' id='word_1_695' title='bbox 1126 2870 1268 2895; x_wconf 87'>sequence,</span> <span class='ocrx_word' id='word_1_696' title='bbox 1280 2861 1344 2888; x_wconf 89'>with</span> <span class='ocrx_word' id='word_1_697' title='bbox 1356 2862 1467 2888; x_wconf 88'>smooth</span> <span class='ocrx_word' id='word_1_698' title='bbox 1478 2861 1635 2888; x_wconf 85'>transitions</span> <span class='ocrx_word' id='word_1_699' title='bbox 1646 2861 1670 2887; x_wconf 90'>if</span> <span class='ocrx_word' id='word_1_700' title='bbox 1679 2861 1792 2888; x_wconf 89'>desired.</span> 
+     </span>
+     <span class='ocr_line' id='line_1_78' title="bbox 347 2906 1507 2940; baseline 0.002 -8; x_size 33; x_descenders 7; x_ascenders 8"><span class='ocrx_word' id='word_1_701' title='bbox 347 2911 360 2925; x_wconf 74'><strong><em>°</em></strong></span> <span class='ocrx_word' id='word_1_702' title='bbox 371 2906 434 2938; x_wconf 92'>Any</span> <span class='ocrx_word' id='word_1_703' title='bbox 444 2906 539 2932; x_wconf 85'>TIME</span> <span class='ocrx_word' id='word_1_704' title='bbox 550 2906 763 2933; x_wconf 88'>SIGNATURE</span> <span class='ocrx_word' id='word_1_705' title='bbox 773 2915 836 2939; x_wconf 88'>may</span> <span class='ocrx_word' id='word_1_706' title='bbox 846 2907 880 2933; x_wconf 91'>be</span> <span class='ocrx_word' id='word_1_707' title='bbox 891 2907 968 2938; x_wconf 89'>used,</span> <span class='ocrx_word' id='word_1_708' title='bbox 980 2907 1036 2934; x_wconf 89'>and</span> <span class='ocrx_word' id='word_1_709' title='bbox 1046 2915 1109 2940; x_wconf 91'>may</span> <span class='ocrx_word' id='word_1_710' title='bbox 1120 2907 1154 2934; x_wconf 90'>be</span> <span class='ocrx_word' id='word_1_711' title='bbox 1164 2907 1288 2940; x_wconf 84'>changed</span> <span class='ocrx_word' id='word_1_712' title='bbox 1299 2907 1393 2934; x_wconf 83'>within</span> <span class='ocrx_word' id='word_1_713' title='bbox 1404 2915 1420 2934; x_wconf 89'>a</span> <span class='ocrx_word' id='word_1_714' title='bbox 1431 2915 1507 2940; x_wconf 87'>song.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_32' title="bbox 1540 2970 1878 3140">
+    <p class='ocr_par' id='par_1_40' lang='eng' title="bbox 1540 2970 1878 3140">
+     <span class='ocr_line' id='line_1_79' title="bbox 1540 2970 1800 3089; baseline 0 -14; x_size 100.5; x_descenders 16.5; x_ascenders 28.499998"><span class='ocrx_word' id='word_1_715' title='bbox 1540 2970 1800 3089; x_wconf 74'>E1111</span> 
+     </span>
+     <span class='ocr_line' id='line_1_80' title="bbox 1554 3108 1878 3140; baseline 0 -5; x_size 32; x_descenders 5; x_ascenders 9"><span class='ocrx_word' id='word_1_716' title='bbox 1554 3108 1624 3135; x_wconf 91'>Linn</span> <span class='ocrx_word' id='word_1_717' title='bbox 1635 3108 1808 3140; x_wconf 87'>Electronics,</span> <span class='ocrx_word' id='word_1_718' title='bbox 1822 3109 1878 3136; x_wconf 84'>Inc.</span> 
+     </span>
+    </p>
+   </div>
+   <div class='ocr_carea' id='block_1_33' title="bbox 1554 3150 2188 3226">
+    <p class='ocr_par' id='par_1_41' lang='eng' title="bbox 1554 3150 2188 3226">
+     <span class='ocr_line' id='line_1_81' title="bbox 1557 3150 2163 3182; baseline 0 -5; x_size 32; x_descenders 5; x_ascenders 8"><span class='ocrx_word' id='word_1_719' title='bbox 1557 3150 1638 3177; x_wconf 88'>18720</span> <span class='ocrx_word' id='word_1_720' title='bbox 1648 3150 1761 3177; x_wconf 89'>Oxnard</span> <span class='ocrx_word' id='word_1_721' title='bbox 1772 3150 1866 3182; x_wconf 81'>Street,</span> <span class='ocrx_word' id='word_1_722' title='bbox 1878 3150 2006 3182; x_wconf 86'>Tarzana,</span> <span class='ocrx_word' id='word_1_723' title='bbox 2019 3150 2071 3177; x_wconf 93'>CA</span> <span class='ocrx_word' id='word_1_724' title='bbox 2082 3150 2163 3177; x_wconf 88'>91356</span> 
+     </span>
+     <span class='ocr_line' id='line_1_82' title="bbox 1554 3192 2188 3226; baseline 0 -7; x_size 33.5; x_descenders 5.5; x_ascenders 8.5"><span class='ocrx_word' id='word_1_725' title='bbox 1554 3193 1622 3226; x_wconf 89'>(818)</span> <span class='ocrx_word' id='word_1_726' title='bbox 1633 3193 1755 3219; x_wconf 60'>7088131</span> <span class='ocrx_word' id='word_1_727' title='bbox 1765 3193 1788 3219; x_wconf 91'>T</span> <span class='ocrx_word' id='word_1_728' title='bbox 1789 3193 1888 3219; x_wconf 87'>ELEX</span> <span class='ocrx_word' id='word_1_729' title='bbox 1899 3192 2022 3219; x_wconf 82'>#298949</span> <span class='ocrx_word' id='word_1_730' title='bbox 2033 3193 2125 3219; x_wconf 88'>LINN</span> <span class='ocrx_word' id='word_1_731' title='bbox 2135 3193 2188 3219; x_wconf 89'>UR</span> 
+     </span>
+    </p>
+   </div>
+  </div>
+ </body>
+</html>

tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/cardinal/__-l__eng__000002.ocr.png__000002__hocr__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state—of—the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+0 Operation is similar to multi—track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+0 Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers !
+
+0 Ultra—fast 3 1/2 ” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+0 One or all tracks may be TRANSPOSED at the touch of a key.
+0 Exclusive real—time ERASE function makes editing FAST.
+0 Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+0 TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+0 Optional SMPTE time code synchronization.
+
+0 Optional remote control.
+
+Recording 3 Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+—existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot—recording. To overdub a new part,
+select a different track and start recording~while you
+record, the first‘track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections.
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+0 Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+0 Non-destructive recording—existing notes are not erased while recording.
+0 Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+0 Two TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+0 Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+' Utilizes ultra high—speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+0 TEMPO may be specified in BEATS-PER—MINUTE or FRAMES-PER—BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+0 TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+0 TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+° Any TIME SIGNATURE may be used, and may be changed within a song.
+
+E1111
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 7088131 T ELEX #298949 LINN UR
+

tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/stderr.bin

@@ -0,0 +1 @@
+Tesseract Open Source OCR Engine v3.05.01 with Leptonica

tests/cache/cardinal/__-l__eng__000003.ocr.png__000003.text__pdf__txt/txt.bin

@@ -0,0 +1,123 @@
+The LinnSequencer
+32 Track MIDI Sequence Recorder
+
+The LinnSequencer is a state—of—the-art composition and performance tool for the professional musician. It is
+
+extremely powerful, yet amazingly simple to learn and use. It’s many remarkable features include:
+
+0 Operation is similar to multi—track tape recorder with PLAY, STOP, RECORD, FAST
+FORWARD, REWIND, and LOCATE controls.
+
+0 Each of the 100 sequences contains 32 simultaneous, polyphonic tracks. Each track may
+be assigned to one of 16 MIDI channels. Simultaneously plays up to 16 polyphonic
+
+synthesizers !
+
+0 Ultra—fast 3 1/2 ” disk drive stores complex songs in seconds and holds over 110,000 notes
+
+per disk!
+
+0 One or all tracks may be TRANSPOSED at the touch of a key.
+0 Exclusive real—time ERASE function makes editing FAST.
+0 Exclusive REPEAT function automatically repeats any held notes at a pre-selected
+
+rhythmic value.
+
+0 TIMING CORRECTION works during playback and operates without ‘chopping’ notes.
+
+0 Optional SMPTE time code synchronization.
+
+0 Optional remote control.
+
+Recording 3 Sequence
+
+To record a sequence, simply press RECORD and PLAY,
+then play your MIDI keyboard in time to the Sequencer’s
+click track. When the sequence loops back around to bar 1,
+you’ll hear what you played—only all timing errors will be
+
+corrected! (Timing correction may be adjusted or defeated).
+
+Any additional notes played will be added into the track
+—existing notes are not erased while recording!
+
+FAST FORWARD, REWIND, and LOCATE controls
+may be used at any time to quickly access any location in
+your sequence for spot—recording. To overdub a new part,
+select a different track and start recording~while you
+record, the first‘track will play in perfect sync (unless you
+MUTE it, or SOLO another track). In this way, up to 32
+tracks may be overdubbed! All MIDI effects are recorded
+including pitch bend, modulation, velocity, aftertouch,
+sustain pedal, and program changes!
+
+Editing
+
+To erase a wrong note, simply hold ERASE and press
+the note to be erased just before it plays in the sequence—
+when played back, it will be gone. Notes may also be
+
+added, erased, or changed using the SINGLE STEP func-
+tion. To overdub notes at specific points within a sequence,
+
+Additional Features
+
+simply use LOCATE, FAST FORWARD, or REWIND to
+find the desired bar number, then start recording.
+
+The INSERT/COPY function allows you to move bars
+from one location to another—in the same sequence or a
+different one. For example, you might insert a copy of the
+first verse between the second chorus and the bridge.
+DELETE BARS operates the same way to remove
+unwanted sections.
+
+Creating a Song
+
+One way to create a song is to record each track all the
+way through (up to 999 bars). Another way is to record
+each basic section (verse, chorus, etc.) in individual
+sequences, then use the CREATE SONG function to “chain”
+them together. CREATE SONG will then automatically
+copy all the parts into a new sequence. If desired, you can
+even set the last few bars to repeat infinitely, for a fadeout.
+
+Composition Without Compromise
+
+The technology you use should never be so complex that
+it interferes with the creative process. That’s precisely why
+the LinnSequencer is designed to let you compose, record
+and edit while devoting your undivided attention to your
+music. See your Linn dealer today for a demonstration!
+
+0 Simple, easy to learn operation—the 32 character LCD display clearly guides you through all operations. If needed, the
+
+HELP button displays additional explanations.
+
+0 Non-destructive recording—existing notes are not erased while recording.
+0 Two FOOTSWITCH INPUTS may be assigned to remotely control many of the commonly used functions, including
+
+ERASE, REPEAT, PLAY/STOP, or LOCATE.
+
+0 Two TRIGGER OUTPUTS may be programmed to output pulses at any selected note value.
+
+0 Will sync to standard LinnDrum or Linn 9000 sync tone.
+
+' Utilizes ultra high—speed, 8 MHz 80186 16 bit computer internally for FAST operation.
+0 TEMPO may be specified in BEATS-PER—MINUTE or FRAMES-PER—BEAT at 24, 25, or 30 frames per second,
+
+(even drop frame!)
+
+0 TEMPO may be entered numerically, adjustable in tenths of a Beat-Per-Minute increments, or by tapping quarter notes
+
+on the TAP TEMPO button.
+
+0 TEMPO CHANGES may be programmed into a sequence, with smooth transitions if desired.
+° Any TIME SIGNATURE may be used, and may be changed within a song.
+
+E1111
+Linn Electronics, Inc.
+
+18720 Oxnard Street, Tarzana, CA 91356
+(818) 7088131 T ELEX #298949 LINN UR
+