Nick Porcino

... I've pushed the patches to `main` as it makes things cleaner overall, and if you have a chance to test, could you LMK if you still get the memory...

Thanks, much appreciated.

I notice that UpSampler has this code ``` for (unsigned i = 0; i < sourceFramesToProcess; ++i) destP[i * 2 + 1] = oddSamplesP[i]; ``` and the destination buffers are...

Is the crash at the end of the demo? I just got repro due to deleting the ShaperNode while it was still processing in the audio thread. Disconnecting the node...

When you say "audio is garbled" do you mean, all the time for all demos? Or sporadically, with regards to the WaveShaperNode demo?

I pushed a fix for m_channelCount. The full TSAN report you posted shows the creation of two pulse audio threads, and that they are frequently racing. RtAudio creates two pulse...

We are using RtAudio 5.1.0. I notice the following in the 5.2.0 release notes. Perhaps we are bumping against the issues they have fixed in 5.2.0, as my current theory...

Pushed an update to 5.2.0, but not 6, as 6 has many API changes.

I'm now able to reproduce the data race locally! This should help things along :)

i don't hear the buzzing. The -1 I introduced in the "odd samples" write could be the cause... I'm working on the race first, the new info is also interesting,...