Hello,
I've developed a PulseAudio backend [1] for Blender which is going to be
released with Blender 3.0 and should be the default audio backend on Linux.
I've developed the backend based on the documentation [2] and looking at
some code of other implementations [3,4], but I'm not sure if I did
everything correctly. It works on my end, but some users are reporting
that it now distorts audio playback on their system [5].
Basically I create a stream that I cork while there is no playback. I'm
using a pa_mainloop object that is iterated in an own thread, to avoid
deadlocks with the locking mechanism of the underlying audio
device/backend code.
I hoped someone could have a quick look at my implementation [1] and
check if I made a mistake using the API. Or maybe there is another
reason for the issues that some people are experiencing? Thanks for your
help already in advance!
Best regards,
Joerg
[1]
https://github.com/neXyon/audaspace/blob/master/plugins/pulseaudio/PulseAudioDevice.cpp
[2] https://freedesktop.org/software/pulseaudio/doxygen/
[3]
https://github.com/kcat/openal-soft/blob/master/alc/backends/pulseaudio.cpp
[4]
https://source.chromium.org/chromium/chromium/src/+/main:media/audio/pulse/pulse_output.cc
[5] https://developer.blender.org/T88089
_______________________________________________
pulseaudio-discuss mailing list
pulseaudio-discuss@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
