[pulseaudio-discuss] How do different sound cards dynamically use different configurations?

[Sean Greenslade]

On Mon, Jul 18, 2022 at 02:42:07PM +0800, zhaochengyi at uniontech.com wrote:
> On 2022/7/17 上午3:18, Sean Greenslade wrote:
> > On Fri, Jul 15, 2022 at 06:04:49PM +0800, zhaochengyi at uniontech.com wrote:
> > > Hi guys,
> > > 
> > > On a special PC, it is required to start the pulseaudio daemon
> > > with some special parameters in the file /etc/pulse/daemon.conf,
> > > as follows:
> > > 
> > > -------
> > > default-sample-rate = 48000
> > > default-fragments = 2
> > > default-fragment-size-msec = 20
> > > -------
> > > 
> > > When plugging in a USB headset with a separate sound card to this special
> > > PC,
> > > I found that there was a noise in the playback audio. After investigation,
> > > the sound card of the USB headset needs to use the following default
> > > parameters.
> > > 
> > > -------
> > > default-sample-rate = 44100
> > > default-fragments = 4
> > > default-fragment-size-msec = 25
> > > -------
> 
> After my further investigation, I found that noise is only generated
> when changing the sink-input volume (pa_context_set_sink_input_volume),
> but not when changing the sink volume (pa_context_set_sink_volume_by_index).

Can you describe this noise more specifically? Does it go away with
larger fragment size / counts? Have you enabled or disabled tsched? Have
you run with Pulse debug logging enabled and looked for any hints there
(e.g. underruns)?

--Sean