On Fri Jun 14 18:55:45 2024 +0000, Paul Gofman wrote:
Good catch, we do as there is no way to communicate to the application
that we chose a period other than 10 ms (which is why `IAudioClient3` exists in the first place). At least for older interfaces across sound APIs there were a few games that firmly assume 10ms (and overflow buffers with larger) periods even if there is a way to communicate a different period.
On that note, I was thinking: should we perhaps force 10 ms even if the UNIX driver reports a higher period?