02:50 <steev> rc8 is out, i'm assuming because of the holidays?

02:55 <broonie> steev: Yes, that was the preannounced plan - the idea is to do the release next week.

03:01 <steev> ah

03:01 <steev> i miss those :D

03:01 <steev> i'm not on the actual list

05:15 <clover[m]> Happy new year

05:23 <steev> happy new year sir, or ma'am

05:56 <HdkR> /8

13:57 <Bioxvirizm-x13s[m]> HNY!=)

15:09 <Jasper[m]> craftyguy, at first glance, iwd does seem to be somewhat better at switching AP's. Not perfect yet, but NetworkManager is now reporting more than a single bar of signal and it's connected to something other than the one wifi6 ap in my house

15:33 <Jasper[m]> @jhovold I recopied the configs for good measure, stopped alsa-state.service. removed /var/lib/alsa/asound.state and rebooted

15:33 <Jasper[m]> speakers are now crackling but properly amped, headphone jack still blasts too loud

15:40 <Jasper[m]> Opening Chromium froze the laptop, had to reboot and now sound is entirely broken: https://paste.centos.org/view/33c9db6c

15:40 <Jasper[m]> Safe to say it's not having fun

15:59 <Jasper[m]> <Jasper[m]> "speakers are now crackling but..." <- the former is a problem with firefox and pipewire, chromium works fine

16:58 <exeat> This seems to eliminate the speaker crackling on pipewire for me: "pw-metadata -n settings 0 clock.max-quantum 1024". It clamps the quantum range to 32-1024 (from the default 32-8192 IIRC).

16:58 <exeat> To unclamp use a 0 value; to persist see end of https://wiki.debian.org/PipeWire

16:59 <exeat> I haven't investigated enough to understand why _higher_ quantum sizes (more relaxed latency) give choppy audio...

16:59 <exeat> Maybe something to do with pipewire's timing approach, the alsa pcm's period size and batch property (and pipewire's treatment of batch devices), some bug in the driver...

16:59 <exeat> (I say this because it kind of sounds like unplayed samples are overwritten.)

17:00 <exeat> Anyway, I suspect that the reason why you hear crackling with firefox and not chrome is that chrome uses a small enough audio buffer to force pipewire to quantum <= 1024, while firefox uses a larger buffer and allows a higher quantum (should be visible in pw-top).

17:00 <exeat> If so, you should be able to get non-choppy audio from firefox while playing silence with chrome (because pipewire would then use a smaller quantum to satisfy the lower-latency client).

17:51 <jglathe> any idea why sound appears to be the most complicated thing in programming?

18:15 <steev> because it is complicated

18:46 <Bioxvirizm-x13s[m]> <steev> "because it is complicated" <- Why? :)

18:46 <steev> i'm not smart enough to know why, but someone like broonie could possibly explain it (not meaning to actually ping him)

18:48 <jglathe> I'm just baffled. Anyway, Happy New Year everybody.

18:49 <exeat> Low frequency signal processed by ubiquitous commodity hardware, driven by software...

18:49 <exeat> Which can further process it in a myriad ways to amuse the monkey brain :)

18:50 <broonie> You’re doing analog/digital conversion and usually trying to do things with minimal latency which means you can start stressing scheduler stuff if you push hard, or run on a system vastly different to the one you tuned for.

18:50 <broonie> Also there’s a bunch of different ways of building the hardware which sometimes surprises people.

18:52 <broonie> Plus there’s UX stuff where people expect things like headphone jacks to behave in one way or another but that one way can be different in different situations

18:53 <jglathe> okay... just noticed that the sound driver stack was way deeper than any other stuff. That has changed a bit (everything is complex now), but it was striking.

18:56 <jglathe> and config of it is opaque.

20:13 <steev> clover[m]: pushed out lenovo-x13s-v6.7.0-rc8, it's the rc7 stuff, on top of rc8

20:14 <broonie> jglathe: graphics is similarly complex really (for similarish reasons)

20:17 <jglathe> I would expect it for graphics

20:17 <robclark> we just had a big head start on gfx.. 10yrs ago it would have been equally broken ;-)

20:18 <Bioxvirizm-x13s[m]> robclark: So-so perspective :)

21:40 <steev> clover[m]: also pushed 6.6.9