[pulseaudio-discuss] Delayed muting of studio speakers

[Arun Raghavan]

On Sat, 19 Dec 2020, at 2:43 PM, Chris Mayes wrote:
> Hi, everybody!
> 
> It takes me back to have subscribed via Mailman to an email 
> distribution list.  I'm generally able to solve my issues via Googling, 
> but this one's proven tricky.

Welcome back to the past! :)

> I recently bought a pair of KRK Classic 5 
> <https://www.krkmusic.com/Classic> speakers to replace a pair of 
> M-Audio Studiophile AV 40 
> <https://m-audio.com/products/view/studiophile-av-40> speakers that had 
> developed a rapid popping noise in the internal amplifier.  The new 
> speakers sound fantastic, but they have an odd spontaneous muting 
> issue, at least as configured.
> 
> The issue is that the speakers seem to become muted (first one, then 
> the other, usually L->R) after some time of playing without any issues. 
>  My provisional solution is to crank the volume past 100%, which 
> un-mutes them (though at an unpleasantly loud volume).  This clears up 
> the issue, though it usually happens again a few minutes later once 
> I've brought the volume back to a reasonable level.
> 
> Based on my Googling, I tried modifying analog-output.conf.common to 
> "ignore" volume.  Here's the PCM block:
> 
> [Element PCM]
> switch = on
> volume = ignore
> volume-limit = 2.0
> override-map.1 = all
> override-map.2 = all-left,all-right
> 
> Sadly, this didn't seem to make any difference.  What else might I try?
> 
> I have a passing familiarity with audio concepts, and one difference 
> that I noted is that the new speakers have about half of the impedance 
> of the old pair (which never had this problem).  Do sound cards use 
> impedance to detect the presence of a device?  It's plugged into 
> line-out (lime green) on an Asus Xonar SE 
> <https://www.asus.com/us/Sound-Cards/Xonar-SE/> (the motherboard (Asus 
> PRIME Z390-A) <https://www.asus.com/us/Motherboards/PRIME-Z390-A/> had 
> the same issue).
> 
> Also, the audio is fed to each speaker separately rather than being fed 
> to a single speaker and bridged to the left via speaker cable.  Maybe 
> that's a factor?
> 
> The brute-force thing would be to just mark the line-out port as 
> "always on" and to skip attempts to detect whether there's a device 
> connected.  Can PulseAudio do this?  More elegant solutions are also 
> warmly welcomed.

This is pretty odd. Could you run the pa-info script (hopefully it's packaged in your distribution) when the stream is playing fine vs. when it's not playing fine? The idea is to see if any mixer controls or any of the stream/sink volumes have actually changed when this happens, or if something else is going on lower down in the hardware.

-- Arun