craffel / mir_eval

Evaluation functions for music/audio information retrieval/signal processing algorithms.
MIT License
593 stars 110 forks source link

Better docstrings for separation #220

Open craffel opened 7 years ago

craffel commented 7 years ago

Currently, the docstrings for the metrics in mir_eval.separation are not super helpful - they basically say "Computes xyz function from bss_eval". Specifically, this makes it unclear what the difference between bss_eval_sources and bss_eval_images is. @ErikJohnsonCU @faroit @carlthome when one of you have a minute, can you make a PR to make them a bit more informative? Thanks!

carlthome commented 7 years ago

@craffel, I don't feel I have a solid enough grasp of the single-channel and multi-channel difference aside from bss_eval_sources pretty much only being a fair estimate for point sources and instantaneous mixes (let's say dry vocals recorded in a dampened booth, or line-in recorded Moogs), so I don't dare to add anything to the docstring. @faroit is the guru here. :hamster:

carlthome commented 7 years ago

Ping @aliutkus