Brian McFee

Giving a nudge on this thread - there are now two viable PRs to resolve the issue, and getting this patched and released would unblock python 3.13 support downstream.

Nudging this again @sampsyo I've managed to unstick my downstream releases from this issue for the time being, but it is going to be a problem long-term. I did just...

> Can I suggest adding python 3.12 and 3.13 to the test matrix in `.github/workflows/main.yml`? I don't think that would do anything at present, since the workflow doesn't actually run...

Thanks for this - it is a dupe of #1859 though. I'll comment on the PR #1886 separately.

Sorry I won't have time to review this for a few weeks.

Sorry for the delay, I'm looking at this now. It's not the easiest PR to review because there's a mixture of cosmetic changes (ie variable names) and refactoring to sift...

A few thoughts: 1. We did make a conscious decision to promote intervals as the fundamental unit of segmentation, not boundaries. I'm not sure if this is documented in the...

Re-upping this after thinking a bit, and yes, I think the way to do it is by exposing an axes parameter in the right places. The call site should probably...

Summary of places where this pops up: - feature.zero_crossing_rate - feature.mfcc - feature.spectral_contrast - core.zero_crossings - onset.onset_strength_multi - effects._signal_to_frame_nonsilent (effects.trim, effects.split) I think in all but the last case, it...

Adding this to the 1.0 docket. It would constitute a very small behavior change (only for multichannel, really), and while we could try to sneak it into 0.11, I'd rather...