Python-based tool for audio file track splitting

The purpose of this program is to provide a convenient way to perform audio file track-splitting, using simple 'YouTube style' timestamps.

E.g.

00:55:45 Pink Elephants on Parade
01:20:54 Circle of Life
02:54:01 Colours of the Wind

I created this script, because of my personal hobby of taking concert audio, and splitting it into tracks, so that I could re-package and enjoy it as an 'album'.

Pipeline

In short, this script:

1. Takes the audio + timestamps file, and generates a cue sheet (an intermediate file containing track splitting information)
2. Uses the cue sheet to perform the split using FFmpeg (see Requirements section)
3. Outputs the tracks to directory with the format: f"{artist}_{album}_tracksplit" - where 'artist' and 'album' are user-provided (see Usage section)

If you're curious, feel free to inspect the tracksplit.py docstrings, for more insight into my workflow.

I've also included some 'legacy' functions which didn't make the cut, but I think are interesting anyway, in the Archive folder.

Future

FFmpeg is actually way more capable than just processing audio files. For this reason, I have planned video splitting capabilities for a future release (see TODO.md for more).

Usage

$ python3 tracksplit.py --timestamps 'timestamps.txt' --audio 'concert_audio.mp3' --artist 'Sensible Clown Conglomerate' --album 'The Light Side of the Sun'

• --timestaps: takes a text file containing 'YouTube style' timestamps
• --audio: takes the source audio file to be split into individual tracks
• --artist and --album: respective 'artist' and 'album' metadata will be embedded during splitting process
• --only-cue: provided, to bypass FFmpeg, and obtain the cue sheet file for your own audio splitting workflow (see Alternative workflows section)

Requirements

There are a whopping 2 (two) dependencies for core audio splitting process:

1. FFmpeg - a multimedia framework which in their own words can handle 'pretty much anything that humans and machines have created'
2. ffcuesplitter - and a Python package which interfaces the with FFmpeg, using a source audio file specific cue sheet, generated during the pipeline

HOWEVER, if you just want an audio source-file specific cue file you can BYPASS these two dependencies entirely, by providing the argument: --only-cue

FFmpeg

If you're on macOS and use homebrew:

brew install ffmpeg

Since we're using Python, here's the conda installation:

conda install -c conda-forge ffmpeg

Otherwise, official platform-specific installations on FFmpeg's website.

ffcuespliter

Python dependency that interfaces with FFmpeg for the splitting process. See the official repo for more info on how it interfaces with FFmpeg.

Install through pip:

python3 -m pip install ffcuesplitter 

Alternative workflows

Apart form FFmpeg, there are other cue-sheet based ways to split the audio track:

• mp3splt - frame-accurate splitting of mp3, ogg vorbis, FLAC and other audio formats without requiring decoding or re-encoding.
• CUEtools - a tool for lossless audio/CUE sheet format conversion
• foobar2000 - one can use the 'Converter' component to split the source audio file, by dragging the respective cue file into foobar2000, highlighting, right-clicking and selecting 'Convert'