[Bug 1207129] New: A few problems in Microsoft Modern Wireless headset
https://bugzilla.suse.com/show_bug.cgi?id=1207129 Bug ID: 1207129 Summary: A few problems in Microsoft Modern Wireless headset Classification: openSUSE Product: openSUSE Distribution Version: Leap 15.4 Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Sound Assignee: tiwai@suse.com Reporter: tiwai@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- We received a new headset "Microsoft Modern Wireless Headset", and this doesn't seem working properly in a few ways on Linux desktops. The device provides two interfaces: USB-audio (with a dongle) and Bluetooth-audio, and the interface can be chosen by a switch on the headset. The connections of both interfaces work out-of-the-box. The USB audio has the following issues: - The playback volume (appearing as "PCM") shows the incorrect dB range information, which confuses pipewire / PulseAudio. - The capture volume and switch have no effect. But the presence confuses PA/PW as if the hardware supports the volume/switch control. - There are occasional (self-)disconnections when the application doesn't use the device. - The volume up/down via dial on the the right ear pad doesn't work properly; it jumps up/down. The BT connection shows fewer problem, as it seems. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |aosthof@suse.com -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 Alexander Osthof <aosthof@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |calmeidadeoliveira@suse.com -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c1 --- Comment #1 from Christian Almeida de Oliveira <calmeidadeoliveira@suse.com> --- Hi Takashi, As a user of this new Headset, I can give you some input: - it worked without the need to install any additional driver as far as I remember. - It's connection to the computer, via bluetooth, it quite fast, takes about 10 seconds. - Sometimes just after connecting to the computer, it disconnects itself. Then it is just a matter to turn off/on again and it connects and stay connected. - I have no audio issues, or seldom, when it is connected with MS-Teams. - Using it for a Huddle in Slack is a nightmare, it mutes my audio (voice) about every 30 seconds. I did not measured it, but is it for sure at least once every minute. - Similar behaviour when using it with Jitsi, it keeps muting, but timing is not so frequent as for Slack. - I got feedback from many people that my voice sounds like in "mono" mode, not stereo. - It has a reasonable noise canceling capability, much better than my previous headset. I'm using openSUSE Leap 15.2, Gnome 3.34.7. In case you need further information, please let me know. Cheers, Christian -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c2 --- Comment #2 from Takashi Iwai <tiwai@suse.com> --- As it's (really old?) Leap 15.2, I suppose you're using PulseAudio, right? And the all behavior are about BT connection, not the USB? The problems I've described are mostly about USB connections. As far as I tested, the BT connection over (the updated) pipewire on Leap 15.4 worked better, no problems described there appeared. Alas, I've been tested only shortly, and just for some application to playback and record shortly. Alexander mentioned in Slack channel that he faces less (but still existing) problem after switching from PulseAudio to pipewire on his Leap 15.4 system. So, the problem is likely dependent on the sound backend, too. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c3 --- Comment #3 from Takashi Iwai <tiwai@suse.com> --- Now I'm building a test KMP with a test patch for this device (in USB mode). It's built in OBS home:tiwai:bsc1207129 repo, for TW and Leap kernels. Alexander, could give it a try later once after the build finishes? Just grab usb-audio-kmp-default.rpm from the repo, install it, reboot and retest. On my machine, the volume up/down works a bit better, but not perfectly (e.g. the volume bar changes but no actual volume changes unless I manually "set" the volume in pavucontrol). As Leap 15.2 was already out of maintenance, I can't provide a fix, unfortunately. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c4 --- Comment #4 from Christian Almeida de Oliveira <calmeidadeoliveira@suse.com> --- All the behaviour I listed is using BT. PulseAudio is installed. As I see it is time to upgrade to the latest Leap version. Do you know if there is a way to avoid using "mono"? Not sure, but checking the PulseAudio cmd pactl list , I came to this part where the channel Map is set to mono. Sink #30 State: RUNNING Name: bluez_sink.A0_4A_5E_F5_65_D7.headset_head_unit Description: Microsoft Modern Wireless Headset Driver: module-bluez5-device.c Sample Specification: s16le 1ch 8000Hz Channel Map: mono Owner Module: 55 Mute: no Volume: mono: 65536 / 100% balance 0.00 Base Volume: 65536 / 100% Monitor Source: bluez_sink.A0_4A_5E_F5_65_D7.headset_head_unit.monitor Latency: 34116 usec, configured 28000 usec Flags: HARDWARE HW_VOLUME_CTRL LATENCY Properties: bluetooth.protocol = "headset_head_unit" device.intended_roles = "phone" device.description = "Microsoft Modern Wireless Headset" device.string = "A0:4A:5E:F5:65:D7" device.api = "bluez" device.class = "sound" device.bus = "bluetooth" device.form_factor = "headphone" bluez.path = "/org/bluez/hci0/dev_A0_4A_5E_F5_65_D7" bluez.class = "0x240418" bluez.alias = "Microsoft Modern Wireless Headset" device.icon_name = "audio-headphones-bluetooth" Ports: headphone-output: Headphone (priority: 0, available) Active Port: headphone-output Formats: pcm -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c5 --- Comment #5 from Takashi Iwai <tiwai@suse.com> --- That's probably because you're using the old version :) The recent version of PA or pipewire allows to select Headphone profile (HFP) instead of Headset (HSP). The former is with a (slightly) better quality of sound as full-duplex. But the input sound itself seems always mono, even with USB connection. Only the output is stereo. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c6 --- Comment #6 from Alexander Osthof <aosthof@suse.com> --- Takashi, I'll bring the headset to the office tomorrow (as it's currently at home where I could test with BT) and will try your KMP then. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c7 --- Comment #7 from Alexander Osthof <aosthof@suse.com> --- OK, I've tried the KMP but unfortunately without good results. Although the devices show up correctly they don't really work: - When turning the volume wheel on the headset itself the display shows jumps from e.g. 100% to 45% to 0% although I'd expect it to be 5% steps - During the one time I managed to get sound output on the headphones to work the volume was very low although it was at 100% - I didn't manage to get the mic to work at all All the above was done using the USB connection with the dongle as here in the office I don't have BT. I've meanwhile switched to pipewire on Leap 15.4. Please let me know if I can/should run any further tests or provide any debugging information. In any case, thanks a lot for investigating this issue. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c8 --- Comment #8 from Takashi Iwai <tiwai@suse.com> --- I checked the volume jog behavior with USB connection. It's a simple HID device, and supposed to send volume up / down events. The generic HID driver can already handle it, so you've seen already the result. The problem is that, this HID device behavior is pretty buggy. It sends the same volume up/down events always twice. This made pipewire reacting too much. (Also I'm afraid that there can be a bug in pipewire for handling this, too.) The second problem of this HID device is that inconsistent volume up/down events. When you turn up the volume largely (rotate many) and turn down again. The device sends volume up events multiple times while turning up -- that's normal. But at next, when you turn down, it sends again the volume up event, not the volume down. This repeats until you reach to the neutral point (as it seems). After that point, the device starts sending the volume down. I can fix the doubled key event; the similar workaround is seen for the Platonorics HID device driver, too. But, the latter crazy volume up/down behavior is hard to fix... Maybe Windows driver switch to a different vendor-specific mode somehow, but I can't trace it, unfortunately. You can uninstall the KMP to restore the original driver. I'll check the driver behavior again on my machine. -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c9 --- Comment #9 from Takashi Iwai <tiwai@suse.com> --- It seems that some of weird behavior of the volume up/down is caused when pvucontrol is opened at the same time. My understanding is that pavucontrol tires to keep the shown volume there, and it drags the actual volume while the apparent volume has been already changed by the up/down key event. Other than the volume job issue, it seemed working on my machine. But my system is with the updated version of pipewire, and that might be difference. (And DE is XFCE.) You can find the update packages in OBS home:tiwai:leap-15.4:update repo. The non-working mic on your machine is a bigger problem, though. If it's reproducible, please make sure that you didn't (mic-)mute it. There is a hardware mic-mute function, and this isn't notified to the OS. Also, you can try to record directly like: % arecord -Dplughw:1 -fdat -vv foo.wav where "1" is the card number of this device (see /proc/asound/cards). If you get a busy error, try to stop pipewire temporarily: % systemctl --user stop wireplumber % arecord ... % systemctl --user start wireplumber -- You are receiving this mail because: You are on the CC list for the bug.
https://bugzilla.suse.com/show_bug.cgi?id=1207129 https://bugzilla.suse.com/show_bug.cgi?id=1207129#c10 --- Comment #10 from Alexander Osthof <aosthof@suse.com> --- I've tested with your packages and got mixed results. Sometimes I was able to record with 'arecord' after stopping wireplumber, sometimes I even had to restart pipewire for a successful recording, sometimes nothing worked. I could really find a pattern to reproduce though. In any case, when trying the headset in conferences (Jitsi, Slack, Teams) it performed poorly - either the volume was low, the volume jog was more or less in binary mode (vol+ = max, vol- = mute). So, I cannot rely on this headset (yet) for conferences. After nearly a day of trying various combinations of settings I gave up. My colleague who's running Gnome on Leap 15.3 (IIRC) had no issues at all with my headset - plugged in, detected, both speakers and mic worked without issues. So it seems there are some parts in KDE (which I'm running on top of Leap 15.4) which make things worse. I'll pause testing/debugging now but I'll try once more when Leap 15.5 is available. In any case, thanks for looking into this issue Takashi, very much appreciated! -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠suse.com