split_audiobook

Simple python script (Python 3.5+), that wraps ffmpeg and ffprobe and enables to split large audiobook file into smaller parts. There are basically 3 methods supported:

• if file contains chapters metadata, then it's split to chapters - it's probably the best case
• you can supply CSV file with chapters definition (name, start, end). You can create this file yourself, by listening to the original file (good tool is audacity - detect silence (Analyze/Silence Finder) spots and listen around silences to find chapters beginnigs, write table in a spreadsheet - name, start, end and save as CSV)
• split into pieces of approximately same size (defined by --length) on silence moments in audio (for best listening experience). This method is used if none of previous two is available.

You can use --dry argument to see what actual split will be.

Resulting files are stored in directory, with same name as file (just without extension). Also cover image is extracted to that directory, if available in metadata. By default script transcodes parts with Opus codec, which is my favorite audio codec for audio books. Result then can be easily used in my audiobooks server audioserve. If you want to just split, but not to transcode use --split-only argument.

For full usage use --help argument.

And I guess you noticed you need to have ffmpeg and ffprobe installed. And yeah, I tested only on Linux (Win/Mac users your are on your own, although I tried to make it as generic as possible).

Install

It's just one script - so you can just download and run in python3 interpreter. Script is intentionaly using only python core modules. Or you can install locally with sudo pip3 install git+https://github.com/izderadicka/split_audiobook.

License

MIT