[Pkg-utopia-maintainers] Bug#1089799: sound drops to mono when i join a videoconference

[Antoine Beaupre]

Package: pipewire
Version: 1.2.7-1
Severity: normal

Hi!

I've been a happy user of Pipewire since bullseye (!). Recently,
however, I noticed something strange: I'm listening to music (whatever
music player, say Youtube in Firefox, or Audacious installed from a
debian package, but I also reproduced with Supersonic, installed from
flatpak) and I jump on a video call (without stopping the music) and
suddenly the quality of the music drops significantly.

At first, I didn't quite understand what was going on, but recently I
*think* I understood what happens, after staying at qpwgraph:
everything drops to mono!

I'm using a [Jabra Evolve 40] headset, that's connected over USB with a
device that looks like this:

Dec 12 09:58:37 angela kernel: usb 3-5.1.3: new full-speed USB device number 49 using xhci_hcd
Dec 12 09:58:38 angela kernel: usb 3-5.1.3: New USB device found, idVendor=0b0e, idProduct=0306, bcdDevice= 4.01
Dec 12 09:58:38 angela kernel: usb 3-5.1.3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Dec 12 09:58:38 angela kernel: usb 3-5.1.3: Product: Jabra EVOLVE LINK
Dec 12 09:58:38 angela kernel: usb 3-5.1.3: Manufacturer: GN Netcom A/S
Dec 12 09:58:38 angela kernel: usb 3-5.1.3: SerialNumber: 00116ADFB53E0A
Dec 12 09:58:38 angela kernel: input: GN Netcom A/S Jabra EVOLVE LINK as /devices/pci0000:00/0000:00:14.0/usb3/3-5/3-5.1/3-5.1.3/3-5.1.3:1.3/0003:0B0E:0306.001D/input/input53
Dec 12 09:58:38 angela kernel: jabra 0003:0B0E:0306.001D: input,hiddev3,hidraw9: USB HID v1.00 Device [GN Netcom A/S Jabra EVOLVE LINK] on usb-0000:00:14.0-5.1.3/input3

[Jabra Evolve 40]: https://www.jabra.ca/business/office-headsets/jabra-evolve/jabra-evolve-40##6393-823-109

According to qpwgraph, the microphone output is mono which makes
*kind* of sense when you think about it: it's a single audio source,
after all. What doesn't make sense is that the headset *inputs* would
drop to Mono, and *only* when we're recording from the device.

I'm filing this in pipewire because, frankly, I don't know where else
to file this. It's quite possible this is a problem that lives
entirely elsewhere. This could be, for example, a kernel driver
bug. There is, indeed, some stuff in the kernel to support this
device. I've been staring at this patch to improve support on the
buttons that thing has, for example:

https://lore.kernel.org/all/1507113082-31678-1-git-send-email-nolsen@jabra.com/

But I figured the pipewire folks could at least give me tools to
confirm "nope, not us, the output *we* are sending to the kernel is
stereo"; right now I don't even know *how* i could confirm
that. Strace? EBPF? A large hammer and butterflies?

Any advice is welcome!

-- System Information:
Debian Release: trixie/sid
  APT prefers testing-debug
  APT policy: (500, 'testing-debug'), (500, 'buildd-unstable'), (500, 'testing'), (1, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 6.11.10-amd64 (SMP w/16 CPU threads; PREEMPT)
Locale: LANG=fr_CA.UTF-8, LC_CTYPE=fr_CA.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages pipewire depends on:
ii  adduser                  3.137
ii  init-system-helpers      1.67
ii  libpipewire-0.3-modules  1.2.7-1
ii  pipewire-bin             1.2.7-1

pipewire recommends no packages.

pipewire suggests no packages.

-- no debconf information