Cool VL Viewer forum

Not a surprise for either of these two: WebRTC is using ALSA, not PulseAudio.

You must find, in the list of devices shown in the viewer voice audio device settings, the one that corresponds to your speakers, but it won't show in the PulseAudio mixer: even if the latter will likely have it listed under another name, you won't be able to control voice volume with it.

You might also want to try and disable PulseAudio altogether (you do not need any "sound daemon" to get audio working under Linux, unless you are using networked audio devices or complex audio filters). For a GUI-based ALSA mixer, I'd recommend QasMixer (and QasHctl, for a complete control over all the ALSA devices and inputs/outputs).

LL can do little about it.

While the WebRTC SDK could be (and has been, for a short while) compiled to use PulseAudio, the resulting WebRTC SDK library would instantly crash on systems not using PulseAudio (or having it disabled), so it would actually be worst and one of the reasons why I always compiled my own WebRTC SDK for Linux with ALSA support enabled and PulseAudio (and PipeWire) support disabled.
ALSA is the one and only common denominator under Linux, and since WebRTC is incapable of proper fallback when compiled with PulseAudio support (and likely PipeWire support: did not try that one, but likely the same crash issue), then we must just accept that ALSA is the only solution for WebRTC voice. But feel free to complain to the WebRTC SDK developers about it...

This is likely a bug... Maybe the auto-tune workaround could help you there (set its debug setting to 1.5s and see how it fares)... But it might as well make things go worst in some cases, so...
Stopping and restarting voice might prove simpler in such cases...