[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Switch default from PulseAudio to PipeWire (and WirePlumber) for audio



On 23 February 2022, I wrote this in https://bugs.debian.org/999363#35:

On 23 Feb 2022 17:21:29 +0100 Diederik de Haas <didi.debian@cknow.org> wrote:
> On 10 Nov 2021 15:40:21 +0100 Sebastien Bacher <seb128@ubuntu.com> wrote:
> > Switching the default sound service is an option but should probably be 
> > a project discussion and not a consequence of a dependency tweak.
> 
> Is that discussion taking place somewhere? If so, where?
> If not, then shouldn't that discussion start ASAP?
> 
> The discussion itself will likely take some time and if the decision is made 
> to do the switch, then it could very well be that various upstream changes 
> should take place (where it hasn't already) and then also Debian packages
> need to be updated.
> 
> I don't know exactly when the Freeze for Bookworm is scheduled, but assuming
> a 2 year release cycle, we now have ~ a year to make that switch happen and 
> suitable for a Stable release.
> If we wait another 6 months or so, it'll become really problematic and we'd 
> have to carry the technical debt likely into Trixie.

I thought ~ a YEAR to make the transition _should_ be doable.
But now that 2/3 of that have passed, I think it would be a mistake to switch 
the default from PA to PW+WP.

On Thu, 26 Aug 2021 11:56:05 +0200 Dylan Aïssi <daissi@debian.org> wrote:
> The best way to deal with that issue is to update all packages that
> Depends, Recommends or Suggests pulseaudio to depend either on
> pulseaudio or pipewire-pulse (i.e. pulseaudio | pipewire-pulse).

This is from more then a YEAR ago and AFAIK very little work on that has been 
done.

Here are some more 'tidbits' from me:
- pipewire-media-session (as session manager) was already deprecated upstream 
since the beginning of the year. It was already then strongly urged to switch 
to WirePlumber, but that has still not been done in Debian
- I saw several references to Wayland and Gnome. Keep in mind there are other 
Desktop Environments. Some 'daredevils' are (sometimes) trying Wayland with 
KDE and IIUC, it'll be a while before anyone can safely switch to using it.
- IIUC/IIRC, even the RTKit maintainer doesn't (really) want to touch the code 
in fear of breaking things. And there were quite a number of issues reported 
wrt RTKit and PW+WP, last time I looked. IIUC that was one of the reasons for 
the (new) recommendation to using /etc/limits.conf and the pipewire group
- When I told/asked upstream how to use PW+WP in an unattended/headless 
manner, I got treated like I was insane and/or this was an insane use case. My 
primary use case was a 'soundserver' device connected to my AVR and one of the 
services it would run was mpd. In the end it was suggested that a systemd 
system service may be a way to get that to work (bug #1014927)
- Tuning the 'quantum' parameter seems to be considered a normal operation. 
Guess what. Not everyone is a (professional) sound engineer. I'm pretty sure 
95+% of Debian users have no idea what that is. I certainly don't. And I've 
spend quite a bit of time to figure out how to make PW+WP work and eventually 
just bailed out. I'm sure things have improved since, but at the beginning of 
the year I wasn't impressed.

I think there currently aren't that many Debian users which do use PW and I 
think that switching the default audio provider in all of Debian (not just 
Gnome) is FAR more involved then the impression I got from this ML thread.

I expect PW+WP to be the future and in several cases it is already superior 
then the alternatives. And I think a small minority of Debian users need that.

I think this is a switch that should be done at the beginning of the Trixie 
development cycle and then there is plenty of time to work out all the kinks 
which undoubtedly will surface when making the switch.
I thought ~ a year was tight, but 4 months is surely too short a timeframe.

My 0.02

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: