aeneas

aeneas is a Python library and a set of tools to automagically synchronize audio and text.

• Version: 1.2.0
• Date: 2015-09-27
• Developed by: ReadBeyond
• Lead Developer: Alberto Pettarin
• License: the GNU Affero General Public License Version 3 (AGPL v3)
• Contact: [email protected]

Goal

aeneas automatically generates a synchronization map between a list of text fragments and an audio file containing the narration of the (same) text.

For example, given this text file and this audio file, aeneas computes the following abstract map:

[00:00:00.000, 00:00:02.680] <=> 1                                                      
[00:00:02.680, 00:00:05.480] <=> From fairest creatures we desire increase,            
[00:00:05.480, 00:00:08.640] <=> That thereby beauty's rose might never die,           
[00:00:08.640, 00:00:11.960] <=> But as the riper should by time decease,              
[00:00:11.960, 00:00:15.279] <=> His tender heir might bear his memory:                
[00:00:15.279, 00:00:18.519] <=> But thou contracted to thine own bright eyes,         
[00:00:18.519, 00:00:22.760] <=> Feed'st thy light's flame with self-substantial fuel, 
[00:00:22.760, 00:00:25.719] <=> Making a famine where abundance lies,                 
[00:00:25.719, 00:00:31.239] <=> Thy self thy foe, to thy sweet self too cruel:        
[00:00:31.239, 00:00:34.280] <=> Thou that art now the world's fresh ornament,         
[00:00:34.280, 00:00:36.960] <=> And only herald to the gaudy spring,                  
[00:00:36.960, 00:00:40.640] <=> Within thine own bud buriest thy content,             
[00:00:40.640, 00:00:43.600] <=> And tender churl mak'st waste in niggarding:          
[00:00:43.600, 00:00:48.000] <=> Pity the world, or else this glutton be,              
[00:00:48.000, 00:00:53.280] <=> To eat the world's due, by the grave and thee.  

The map can be output to file in several formats: SMIL for EPUB 3, SRT/TTML/VTT for closed captioning, JSON/RBSE for Web usage, or raw CSV/SSV/TSV/TXT/XML for further processing.

System Requirements, Supported Platforms and Installation

System Requirements

1. a reasonably recent machine (recommended 4 GB RAM, 2 GHz 64bit CPU)
2. ffmpeg and ffprobe executables available in your $PATH
3. espeak executable available in your $PATH
4. Python 2.7.x
5. Python modules BeautifulSoup, lxml, numpy, and scikits.audiolab
6. (Optional but strongly suggested) Python C headers to compile the Python C extensions

Depending on the format(s) of audio files you work with, you might need to install additional audio codecs for ffmpeg. Similarly, you might need to install additional voices for espeak, depending on the language(s) you work on. (Installing all the codecs and all the voices available might be a good idea.)

If installing the above dependencies proves difficult on your OS, consider using the Vagrant box created by aeneas-vagrant.

Supported Platforms

aeneas has been developed and tested on Debian 64bit, which is the only supported OS at the moment.

However, aeneas has been confirmed to work on other Linux distributions (Ubuntu, Slackware), on Mac OS X (with developer tools installed) and on Windows Vista/7/8.1/10.

Whatever your OS is, make sure ffmpeg, ffprobe (which is part of ffmpeg distribution), and espeak are properly installed and callable by the subprocess Python module. A way to ensure the latter consists in adding these three executables to your $PATH.

If installing aeneas natively on your OS proves difficult, you can use VirtualBox and Vagrant to run aeneas inside a virtualized Debian image, using aeneas-vagrant.

Installation

Linux and Mac OS X

1. If you are a user of a deb-based Linux distribution (e.g., Debian or Ubuntu), you can install all the dependencies by running the provided install_dependencies.sh script

   $ sudo bash install_dependencies.sh

2. If you have another Linux distribution or Mac OS X, just make sure you have ffmpeg, ffprobe (part of the ffmpeg package), and espeak installed and available on your command line. You also need Python 2.x and its "developer" package containing the C headers.

3. Run the following commands:

   $ git clone https://github.com/readbeyond/aeneas.git
   $ cd aeneas
   $ pip install -r requirements.txt
   $ python setup.py build_ext --inplace
   $ python check_dependencies.py

If the last command prints a success message, you have all the required dependencies installed and you can confidently run aeneas in production.

Windows

Please read the installation instructions contained in the "Using aeneas for Audio-Text Synchronization" PDF, based on these directions, written by Richard Margetts.

Usage

1. Install aeneas as described above. (Only the first time!)

2. Open a command prompt/shell/terminal and go to the root directory of the aeneas repository, that is, the one containing this README.md file.

3. To compute a synchronization map map.json for a pair (audio.mp3, text.txt in plain format), you can run:

   $ python -m aeneas.tools.execute_task audio.mp3 text.txt "task_language=en|os_task_file_format=json|is_text_type=plain" map.json

   The third parameter (the configuration string) can specify several parameters/options. See the documentation for details.

4. To compute a synchronization map map.smil for a pair (audio.mp3, page.xhtml containing fragments marked by id attributes like f001), you can run:

   $ python -m aeneas.tools.execute_task audio.mp3 page.xhtml "task_language=en|os_task_file_format=smil|os_task_file_smil_audio_ref=audio.mp3|os_task_file_smil_page_ref=page.xhtml|is_text_type=unparsed|is_text_unparsed_id_regex=f[0-9]+|is_text_unparsed_id_sort=numeric" map.smil

5. If you have several tasks to run, you can create a job container and a configuration file, and run them all at once:

   $ python -m aeneas.tools.execute_job job.zip /tmp/

   File job.zip should contain a config.txt or config.xml configuration file, providing aeneas with all the information needed to parse the input assets and format the output sync map files. See the documentation for details.

You might want to run execute_task or execute_job without arguments to get an usage message and some examples:

$ python -m aeneas.tools.execute_task
$ python -m aeneas.tools.execute_job

See the documentation for an introduction to the concepts of task and job, and for a list of the available options.

Documentation

Online: http://www.readbeyond.it/aeneas/docs/

Generated from the source (requires sphinx):

$ git clone https://github.com/readbeyond/aeneas.git
$ cd aeneas/docs
$ make html

Tutorial: A Practical Introduction To The aeneas Package

Mailing list: https://groups.google.com/d/forum/aeneas-forced-alignment

Changelog: http://www.readbeyond.it/aeneas/docs/changelog.html

Supported Features

• Input text files in plain, parsed, subtitles, or unparsed format
• Text extraction from XML (e.g., XHTML) files using id and class attributes
• Arbitrary text fragment granularity (single word, subphrase, phrase, paragraph, etc.)
• Input audio file formats: all those supported by ffmpeg
• Batch processing
• Output sync map formats: CSV, JSON, SMIL, SSV, TSV, TTML, TXT, VTT, XML
• Tested languages: BG, CA, CY, DA, DE, EL, EN, ES, ET, FA, FI, FR, GA, GRC, HR, HU, IS, IT, LA, LT, LV, NL, NO, RO, RU, PL, PT, SK, SR, SV, SW, TR, UK
• Robust against misspelled/mispronounced words, local rearrangements of words, background noise/sporadic spikes
• Code suitable for a Web app deployment (e.g., on-demand AWS instances)
• Adjustable splitting times, including a max character/second constraint for CC applications
• Automated detection of audio head/tail
• MFCC and DTW computed as Python C extensions to reduce the processing time

Limitations and Missing Features

• Audio should match the text: large portions of spurious text or audio might produce a wrong sync map
• Audio is assumed to be spoken: not suitable/YMMV for song captioning
• No protection against memory trashing if you feed extremely long audio files

TODO List

• Improving robustness against music in background
• Isolate non-speech intervals (music, prolonged silence)
• Automated text fragmentation based on audio analysis
• Auto-tuning DTW parameters
• Reporting the alignment score
• Improving (removing?) dependency from espeak, ffmpeg, ffprobe executables
• Multilevel sync map granularity (e.g., multilevel SMIL output)
• Supporting input text encodings other than UTF-8
• Better documentation
• Testing other approaches, like HMM
• Publishing the package on PyPI

How Does This Thing Work?

One Word Explanation

Math.

One Sentence Explanation (Layman Edition)

A good deal of math and computer science, a handful of software engineering and some optimization tricks.

One Sentence Explanation (Pro Edition)

Using the Sakoe-Chiba Band Dynamic Time Warping (DTW) algorithm to align the Mel-frequency cepstral coefficients (MFCCs) representation of the given (real) audio wave and the audio wave obtained by synthesizing the text fragments with a TTS engine, eventually mapping the computed alignment back onto the (real) time domain.

Extended Explanation

To be written. Eventually. Some day.

License

aeneas is released under the terms of the GNU Affero General Public License Version 3. See the LICENSE file for details.

The code for computing the MFCCs aeneas/mfcc.py is a verbatim copy from the CMU Sphinx III project.

Audio files contained in the unit tests aeneas/tests/res/ directory are adapted from recordings produced by the LibriVox Project and they are in the public domain.

Text files contained in the unit tests aeneas/tests/res/ directory are adapted from files produced by the Project Gutenberg and they are in the public domain.

No copy rights were harmed in the making of this project.

Supporting and Contributing

Sponsors

• July 2015: Michele Gianella generously supported the development of the boundary adjustment code (v1.0.4)

• August 2015: Michele Gianella partially sponsored the port of the MFCC/DTW code to C (v1.1.0)

• September 2015: friends in West Africa partially sponsored the development of the head/tail detection code (v1.2.0)

Supporting

Would you like supporting the development of aeneas?

We are open to accept sponsorships to

• fix bugs,
• add new features,
• improve the quality and the performance of the code,
• port the code to other languages/platforms,
• support of third party installations, and
• improve the documentation.

Feel free to get in touch.

Contributing

If you are able to contribute code directly, that's great!

Please do not work on the master branch. Instead, please create a new branch, and open a pull request from there. I will be glad to have a look at it!

Please make your code consistent with the existing code base style (see the Google Python Style Guide), and test your contributed code against the unit tests before opening the pull request. Ideally, add some unit tests on the code written by you.

Please note that, by opening a pull request, you automatically agree to apply the AGPL v3 license to the code you contribute.

If you think you found a bug, please use the GitHub issue tracker to file a bug report.

Development History

Early 2012: Nicola Montecchio and Alberto Pettarin co-developed an initial experimental package to align audio and text, intended to be run locally to compute Media Overlay (SMIL) files for EPUB 3 Audio-eBooks

Late 2012-June 2013: Alberto Pettarin continued engineering and tuning the alignment tool, making it faster and memory efficient, writing the I/O functions for batch processing of multiple audio/text pairs, and started producing the first EPUB 3 Audio-eBooks with Media Overlays (SMIL files) computed automatically by this package

July 2013: incorporation of ReadBeyond Srl

July 2013-March 2014: development of ReadBeyond Sync, a SaaS version of this package, exposing the alignment function via APIs and a Web application

March 2014: launch of ReadBeyond Sync beta

April 2015: ReadBeyond Sync beta ended

May 2015: release of this package on GitHub

August 2015: release of v1.1.0, including Python C extensions to speed the computation of audio/text alignment up

September 2015: release of v1.2.0, including code to automatically detect the audio head/tail

Acknowledgments

Many thanks to Nicola Montecchio, who suggested using MFCCs and DTW, and co-developed the first experimental code for aligning audio and text.

Paolo Bertasi, who developed the APIs and Web application for ReadBeyond Sync, helped shaping the structure of this package for its asynchronous usage.

All the mighty GitHub contributors, and the members of the Google Group.