David Nicholson

See also https://github.com/detly/gammatone

Gammatone gives you features, like an MFCC. So this should go in `vocalpy.feature`

Python implementation here: https://github.com/MatsumotoJ/usvseg_python

Edit to update we also want decorators to wrap segmenting algorithms, `returns_seq`, `returns_box`

Write these as class decorators that use [functions.update_wrapper](https://docs.python.org/3/library/functools.html#functools.update_wrapper) ```python import functools class takes_audio def __init__(self, func, data_arg: str, samplerate_arg: str | None = None): self.func = functools update_wrapper(func) self.data_arg =...

Tutorials / vignettes should show how to get some defaults, convert to a dict, over-ride certain parameters by just assigning a new value to a certain key, etc.

Note in that example that I had an idea to make a `PathArray` class that would work with Pandas but that turned out to be more painful than I thought,...

Thinking about this again now that I'm back to development. I think we want something like this, at the highest level: ```python import vocalpy as voc dataset = voc.Dataset.from_dir('./data/bird1/2022-12-04', 'wav')...

We don't want to. Just document that we can't so it's clear. Explainer from Matlab is here: https://www.mathworks.com/help/matlab/import_export/mat-file-versions.html

Thank you @ralphpeterson, really appreciate the input. I can think of plenty of more than a few other papers that use the warbleR features too -- being able to access...