Closed oliver-bendig closed 4 months ago
Thanks for the report, and for figuring out a way to solve it! The check is there to catch an issue, I mostly have seen with VB Cable, where the capture device sometimes would miss an event. The check simply checks that the time between events isn't too long. If it's over twice the expected time, then there was a missed event. I put the limit at 1.5 because it was an easy number that seemed to work well, but it could well be a larger number closer to 2.0. You could try 1.8 or even 1.9 to see if that gets rid of the last few warnings. There is a risk that it doesn't catch a missed event (larger risk the closer you get to 2.0). If that happens, there will be lots of warnings, and the sound becomes really choppy.
Hi, thanks for the very quick reply and the technical background! There was no difference for me when running with VB-Cable and my onboard Soundcard (Realtek Audio) as capture device. Both showed the same error messages and the same intervals. I changed to 1.7 now and no more errors in the log anymore. I am enjoying the music now! :-) CamillaDSP is absolutely great! Thanks again!
Thanks for testing! I will include this fix in the next version.
The fix is included here: https://github.com/HEnquist/camilladsp/pull/319/files
fixed in release 2.0.2
Hi,
first of all a great thanks for this fantastic project! I tried to run camilladsp 2.0.1 on my wIndows desktop machine (3.7GHz Core i5 12th gen) . Unfortunately, there is a lot of crackling/distortion in the sound and there are buffer underrun errors. I tried many different combinations of chunksize and target_level, but didn't get it to work. CPU usage is not measurable < 0.1%. I removed all filters and also tried to set realtime thread priority to the camilladsp process, but nothing worked.
The log looks is full of lines like this:
Actually, I wondered about the long floating numbers and thought that they are pretty close to the mentioned 0.01s - so maybe something with the if condition? I don't know all the technical details, but I simply tried to change the if condition wasapidevice.rs line 358
to
to disarm this check and avoid restarting the audio stream. This fixed the buffer underruns and the distortion for me. At least I don't hear any distortion anymore. From time to time, there are still complains about Interval 0.016xxx now in the log, but only very few and no more crackling/buffer underruns anymore.
Maybe my findings are somehow helpful...
Herre a logfile with chunksize=8192, target_level=16383. These were best values I was able to find. camilla_trace.zip
Thanks, Olli