Glitch at the end of playback.

[DavidBailes]

Describe the bug
A clear and concise description of what the bug is.

To Reproduce
Steps to reproduce the behavior:

1. Generate a few seconds of a tone
2. select some audio within this
3. playback
4. There's a "glitch" at the end

Expected behavior
No glitches!

Screenshots
Audio selected in Audacity:

Playback recorded using loopback in Reaper - note spurious audio after the correct audio has been played:

Using Audacity 3.1.2, same project, recorded using loopback in Reaper - no spurious audio:

Additional information (please complete the following information):

• Audacity 3.1.3 commit: fa4ef2
• Windows 10

[DavidBailes]

Using chirp instead of tone to give more info as to what spurious audio is being played:
Audio selected in Audacity:

Using Audacity 3.1.3, playback recorded using reaper:

Using Audacity 3.1.2, playback recorded using reaper:

[petersampsonaudacity]

Confirmed on W10 with altest 3.2.0 alpha

And I have a sneaking suspicion it's related to: Play-at -Speed with a selection gets stuck in Play mode #2316

[Paul-Licameli]

Confirmed on W10 with altest 3.2.0 alpha

And I have a sneaking suspicion it's related to: Play-at -Speed with a selection gets stuck in Play mode #2316

No, not related

[Paul-Licameli]

The problem bisects to 81f74f7

I can fix it by increasing the numbers in NewDefaultPlaybackPolicy::SuggestedBufferTimes such as to 0.25, 0.25, 0.5 but this will have the consequence of making the response to the changing slider more laggy.

[DavidBailes]

I can fix it

Please fix it. If a user selects some audio for playback, then that audio should be played back - not the audio and bit of what comes afterwards. This could lead to inaccurately positioned edits.

[Paul-Licameli]

I can fix it

Please fix it. If a user selects some audio for playback, then that audio should be played back - not the audio and bit of what comes afterwards. This could lead to inaccurately positioned edits.

I fear endless rounds of grief if I do it that way, because the complaint about lagginess will predicatably follow. I hope to find a different way that can reconcile both demands but it is not obvious yet.

Adjustable play-at-speed really is getting a total reimplementation so it can also accomodate loop region changes on the fly.