jiamingz42/condense-audio

Create condensed audio + subtitle from video + subtitle for language learning. Inspired by subs2srs.

Motivation

LingQ allows user to upload subtitle (e.g. srt, ass, vtt) along with audio as a lesson. Then user can read the subtitle sentence by sentence and can also choose to play the audio for this sentence.

User can also listen to this audio for passive learning. To increase the efficient, I created this program to only extract those audio segments where there are conversations. There are similiar programs that can extract condensed audio but it doesn't adjust the subtitle accordingly. This program does both.

Usage

python main.py --sin sample.vtt --sout out.vtt --vin sample.mp4 --out combined.mp3 --keep-tmpdir

Test

Check if all the code conforms to the type hinting

fd py -X mypy

Run all tests

pytest gives out more readable error than the vanilla unittest module

PYTHONPATH=. $(pyenv which pytest)

Formating

fd py -X autoflake --in-place --remove-unused-variables --remove-all-unused-imports

TODO

1. Can post translation to LingQ
2. Can exapnd the caption endtime
3. Can join short caption