ass2rythmo 1.0.0

Generate rythmo band from Ass subtitles

ass2rythmo is a converter which generate rythmo band out of subtitles in ASS (Advanced SubStation) format.

Rythmo bands are used to provide synchronized script for voice actors when dubbing a video.

The text scrolls at a constant speed but has a variable width to adapt to the speech flow that needs to be recorded to be synchronised with lips on screen.

Installation

Ass2Rythmo can be installed from pip:

pip install ass2rythmo

Alternatively, Ass2Rythmo can be run directly from sources after a git pull:

git clone https://gitlab.com/zeograd/ass2rythmo.git
cd ass2rythmo && python setup.py install

Usage

First, generate a timed ASS subtitle file including only the part having to be said by actors.

Then, generate the rythmo band as as collection of transparent png files:

ass2rythmo my_video_script.ass -R 30

or as a transparent movie if you have ffmpeg available:

ass2rythmo my_video_script.ass -R 30 -r rythmo_band.mov

or as opaque movie, still with ffmpeg available:

ass2rythmo my_video_script.ass -R 30 -r rythmo_band.mp4

Don’t forget to specify the frame rate of the video on which the png will be superposed to ensure a correct synchronisation.

By default, this will generate png file in the current directory named 00000.png, 00001.png, etc… You can change the directory in which the png files are generated by providing a pattern to the -r argument including percent sign (/tmp/%05d.png for instance, for generating png in the /tmp directory and padding the frame number on 5 digits.)

If you specify a pattern without any percent sign, it will be considered as a single movie file output filename. The result will be encoded with FFV1 codec and mov container if the extension of the output file is mov.

If you use another extension, the default ffmpeg rule will apply. Typically you may want a h264 encoded movie in a mp4 container by asking for an output filename ending in mp4.

Then you can burn the png sequence on top of your movie using your favorite video editor. You can also use ffmpeg to perform a quick overlaying, with a command similar to:

ffmpeg -i input.mp4 -framerate 30 -i %05d.png -filter_complex "overlay=:main_h-overlay_h" output.mp4

to do the same with a transparent mov rythmo band, use something like:

ffmpeg -i input.mp4 -i rythmo_band.mov -filter_complex "overlay=:main_h-overlay_h" output.mp4

Script creation

Use any ASS compatible tool, like Aegisub.

You can use the regular { } to include comments or alternatively [ ] to include content which won’t appear in the rythmo band (but still be present if you burn the ass file onto a video, like for actor direction hints).

The bundled font was chosen so that you can include unicode character to show which kind of intonation change you’d like (using unicode arrows).

You can fine tune a single sentence timing using the karaoke tags. You don’t have to split your script for the sake of introducing a pause or changing the pace between 2 parts of a sentence. In this case, you can preview the speed at which every part of your sentence will be rendered using /kf instead of the default /k tags (both will be accepted by ass2rythmo, /ko is also working)

Obviously, take care of synchronizing carefully your subtitles with the lips on screen to ensure that recording according to the rythmo band will be synchronized with the images.

You can disable rendering rythmo band for an actor by including [norythmo] in its name.

CLI

The CLI ass2rythmo includes sensible default values. You should be careful about setting the correct frame rate with -R option.

You may want to alter the generated width to use fully your available room with -W and also change the scroll speed with -D.

Other options are less often useful and documented below.

usage: ass2rythmo [-h] [-d] [-v] [-W TARGET_WIDTH] [-H TARGET_HEIGHT]

[-R TARGET_FPS] [-F FONT_FILENAME] [-D SCROLL_DURATION] [-r RESULT_FILENAME_FORMAT] [-s SYNC_POINT_RATIO] ass_filename

GPL v3+ 2018 Olivier Jolly

positional arguments:

ass_filename input file as Ass format

optional arguments:

-h, --help

show this help message and exit

-d, --debug

print debugging info [default: False]

-v, --version

show program’s version number and exit

-W TARGET_WIDTH, --target_width TARGET_WIDTH

Width of the generated png files [default: 1920]

-H TARGET_HEIGHT, --target_height TARGET_HEIGHT

Height of the generated png files. Note that it’s not meant to support low heights at this time [default: 380]

-R TARGET_FPS, --target_fps TARGET_FPS

Frame per second to generate. Make sure it matches your subbed video ! [default: 30]

-F FONT_FILENAME, --font_face FONT_FILENAME

Filename of the font used to generate text [default: [install path]/DejaVuSans.ttf]

-D SCROLL_DURATION, --scroll_duration SCROLL_DURATION

Duration of width traversal in ms [default: 4000]

-r RESULT_FILENAME_FORMAT, --result_filename_format RESULT_FILENAME_FORMAT

Filename template for generated files. May include directory part but won’t create any. Including % in the format will generate a list of png files else will generate a transparent movie with the help of ffmpeg [default: %05d.png]

-s SYNC_POINT_RATIO, --sync_point_ratio SYNC_POINT_RATIO

Position of the sync point on screen, 0 is leftmost, 1 is rightmost [default: 0.15]

Convert ASS subtitles to rythmo band as list of transparent png or single transparent movie

LIMITATIONS

This program is quite young and not extensively tested. Notably, it was only run under linux so far, but no OS specificity were used.

The rendering is done on 2 lines maximum. So with 3 actors speaking at the same time, there will be some overlayed text above each other.

Some artefacts may appear on the leftmost part of short text chunk for some characters (like a “I” to be pronounced very quickly at the beginning of a text chunk).

LICENSE

The ass2rythmo code itself is licensed under GPL V3+ and the bundled DejaVu font is licensed under its own free License (cf LICENCE-dejavu.txt).