Instead of erroring with no audio server, restore server-side --disable_audio CLI option or automatically detect no PulseAudio server

[ehfd]

Before v1.5.x, before #96, there was an option to disable audio both from the client and server.

This was needed because that was a time when we did not figure out the root cause of latency which arose from #7.

Even now, it isn't necessary to restore the client-side disable audio button since all web browsers now have the option to mute from the tab, but disabling server-side audio still does make sense.

This is because without an active PulseAudio/PipeWire-Pulse server, the whole connection crashes. Thus, this option is required to run Selkies without PulseAudio.

It would be even better to detect when pulsesrc is None instead of having the --disable_audio option and simply not to initialize the audio WebRTC stack if so instead of crashing.

Not immediately necessary within containers but the need may eventually arise in specialized environments such as bare metal or clusters (especially HPC or certain types of containers).

This would involve restoring the code in #96 to be fit for the current state if going the first way, and handling pulsesrc is None gracefully in gstwebrtc_app.py if going the second way. I prefer the second way but the first way also works.