Brian McFee

Thanks for getting this started! I've left a few specific comments to make the text more appropriate for us. Bigger picture, it would probably be good to take a look...

> This PR fixes a TypeError in the `time_stretch` function where `round()` was applied to an ndarray inappropriately. The fix replaces `round()` with `np.round()` to handle the numpy arrays correctly....

Ah I see - I hadn't considered plugging tempo estimates into the time_stretch like you're doing. That makes sense. There are other bits of this function that will not be...

Finally circling back on this after a hectic couple of months. I'm actually unable to reproduce the behavior that motivated the change here. Can you fill in the issue template...

The dB colorbar should be pretty straightforward - it's just a matter of putting the right format string in place and otherwise passing through args. Phase ticking is a little...

Small note: for dB scaling, it would probably be nice to use a step locator in increments of 3 or 6 dB.

Update: the final version of #1938 removed the unwrrapped phase mode, so we don't have to worry so much about unbounded phase values and ticking at multiples of 2π.

Test coverage is up on the new scaler code, but I'm seeing failures on ubuntu-arm and win-x64, likely due to numerical differences in phase handling. This surfaced a bug in...

MPL image comparison failures are basically due to phase disagreements in silent regions of the spectrogram: Baseline (my dev machine, linux x86_64):  Result (arm64):  Diff:  Not much...

I think this is functionally there now. It still needs documentation. I did realize in coding this up that unwrapped phase should not use a cyclic colormap; I opted for...