notes

Goal: Create a program which automatically slices vinyl-ripped wav files.

Input: audio file, [optional] output dir, [[optional] time between tracks, [optional] number of tracks]
Output: audio file sliced into pieces in output dir

Using: python2, numpy (analysis), audiolab (io/processing)
Algorithm:
1. Determine silence threshold
  - How: take as arg? parse out of first possible gap?
  - Should it be amplitude or frequencies or both?
2. Given a start point, locate start of next track/gap before next track
  - Find transition points (using amplitude or frequency domain or both?)
  - Select the earliest transition point that has some sec of "silence" prior
3. Select from start point to start of second track, add to export list.
4. Go to step 2 until file end.
5. Export all files in export list.

Possible future features: non-greedy, allowing for number of tracks and fewer false positives