Brian McFee

Tacking onto this, it might also be a good time to revise and update the documentation glossary. Or remove it - it only got 40 views total across all versions...

Thanks @matthewfeickert ! Nothing to add from me right now, but I'll keep an eye on it.

A couple of notes: 1. We should probably have a mechanism to disable parallelism. (Helpful for debugging, or if folks are implementing parallelism on their own at a higher level,...

> There might be some bad interactions between numba caching and parallelism. We'll have to investigate. This turns out to be a known issue in numba caching https://github.com/numba/numba/issues/6845 - the...

Coming back to this: [asv](https://asv.readthedocs.io/en/stable/) seems to be the tool of choice these days. It makes it easy to compare things like time, memory, peak memory, and arbitrary numbers (eg...

The orient argument is (I believe) a seaborn-ism, which provides a more consistent interface over matplotlib's mixture of doppelganger functions and argument swaps.

I think #1539 is done enough, but it could use a CR for the docs.

Thanks for taking a crack at this! A couple of thoughts: - I think the spectrogram-like display makes the most sense here. You can bypass the interpolation step by passing...

That looks great, thanks!

Implementing #1480 will also require bumping the numba version (to one which does not yet exist) - so we'll need to resolve this issue first to avoid environment conflicts in...