Arch Linux

wand_am

Member

Registered: 2026-09-04

Posts: 2

wireplumber 0.5.17-1 - broken audio with Scarlett Solo 4th Gen

i first ran into these issues with wireplumber's 0.5.16-1 update; after reading up about the node duplication bug i figured that they were related to it, so downgraded to 0.5.15-1 and waited for a fix. the list of sinks does not grow when plugging the interface in and out, but all of the other issues i first experienced are the exact same. im not sure how important this is for the issue, but i am on KDE Plasma and the 7.2.3-arch1-2 kernel
in the output of aplay -l i can see that the interface is recognized:

**** List of PLAYBACK Hardware Devices ****
card 0: Gen [Scarlett Solo 4th Gen], device 0: USB Audio [USB Audio]
  Subdevices: 1/1

switching outputs/inputs is very buggy - the radio buttons stay on when clicked and dont switch off when another button from the same category is selected. plugging the interface out and in again causes them to reappear, all unselected. sometimes none appeared (im unsure why), with the interface marked as an 'inactive card' in the settings. manually switching it from 'Off' makes the output and inputs reappear in the GUI
attempting to play audio from a source produces the following message in journalctl:

pipewire-pulse[4505]: mod.protocol-pulse: [] timeout on stream 0x564424417420 channel:6
pipewire-pulse[4505]: mod.protocol-pulse: [] timeout on stream 0x564424f446e0 channel:7
pipewire-pulse[4505]: mod.protocol-pulse: [] timeout on stream 0x564424deea10 channel:8
pipewire-pulse[4505]: mod.protocol-pulse: [] timeout on stream 0x564425022780 channel:9
pipewire-pulse[4505]: mod.protocol-pulse: [] timeout on stream 0x5644248bc8b0 channel:10

all of the audio outputs and inputs appeared correctly in pavucontrol. while attempting to play audio from a youtube video, the output always defaulted to some "Unknown output". trying to switch to a different one, like the interface's headphone output, produced no audio. the selected output would get replaced by "Unknown output" after a few seconds
the node priorities from pw-cli ls Node appear to be correct as far as i understand:

 		node.name = "Dummy-Driver"
	id 30, type PipeWire:Interface:Node/3
--
 		node.name = "Freewheel-Driver"
	id 57, type PipeWire:Interface:Node/3
--
 		priority.session = "100"
 		priority.driver = "1"
 		node.name = "Midi-Bridge"
 		media.class = "Midi/Bridge"
--
 		priority.session = "2100"
 		priority.driver = "2100"
--
 		node.name = "alsa_input.hw_Gen_0"
 		node.nick = "Monitor Input 3"
--
 		priority.session = "1100"
 		priority.driver = "1100"
--
 		node.name = "alsa_output.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Line1__sink"
 		node.nick = "Scarlett Solo 4th Gen"
--
 		priority.session = "2100"
 		node.description = "Scarlett Solo 4th Gen Input 1 Inst/Line"
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Mic1__source"
 		node.nick = "Scarlett Solo 4th Gen"
--
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Line3__source.split"
 		media.class = "Stream/Input/Audio/Internal"
--
 		priority.session = "2100"
 		node.description = "Scarlett Solo 4th Gen Monitor Input 3"
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Line3__source"
 		node.nick = "Monitor Input 3"
--
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Mic2__source.split"
 		media.class = "Stream/Input/Audio/Internal"
--
 		priority.session = "2100"
 		node.description = "Scarlett Solo 4th Gen Input 2 Mic"
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Mic2__source"
 		node.nick = "Scarlett Solo 4th Gen"
--
 		node.name = "alsa_input.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Mic1__source.split"
 		media.class = "Stream/Input/Audio/Internal"
--
 		node.name = "bluez_midi.server"
 		media.class = "Midi/Bridge"

below is also a comparison of output from pactl list short sinks. the first is on the most recent wireplumber version, the 2nd is when everything is functioning properly after a downgrade:

126	alsa_output.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Line1__sink	PipeWire	s32le 2ch 48000Hz	   SUSPENDED

249	alsa_output.usb-Focusrite_Scarlett_Solo_4th_Gen_S1QQKEM360B9CC-00.HiFi__Line1__sink	PipeWire	s32le 2ch 48000Hz	IDLE
257	alsa_output.usb-Generic_USB_Audio-00.HiFi__SPDIF__sink	PipeWire	s16le 2ch 48000Hz	SUSPENDED
258	alsa_output.usb-Generic_USB_Audio-00.HiFi__Speaker__sink	PipeWire	s32le 2ch 48000Hz	SUSPENDED
259	alsa_output.usb-Generic_USB_Audio-00.HiFi__Headphones__sink	PipeWire	s32le 2ch 48000Hz	SUSPENDED
262	alsa_output.pci-0000_12_00.1.hdmi-stereo-extra2	PipeWire	s32le 2ch 48000Hz	SUSPENDED

restarting wireplumber only causes the inputs & outputs to briefly disappear from the GUI, that's the only discrenible effect i can spot. restarting the whole audio stack similarly has no useful effect.
so far ive only managed to avoid this problem by staying on the 0.5.15-1 version of wireplumber. if im on 0.5.17-1 and my audio is broken, downgrading to 0.5.15-1 and restarting wireplumber during the same session instantly makes the audio work properly again. i couldnt find any other posts concerning an issue like this, and im unsure how to diagnose it on my system further. the bug being tied solely to the version of wireplumber i have installed leads me to believe theres something wrong with it specifically, especially with how seamlessly it starts working again after i downgrade and restart it. of course, staying on an old version is a bad idea in the long term, so ill be very grateful for advice on how to address this odd behavior :)

wand_am

Member

Registered: 2026-09-04

Posts: 2

Re: wireplumber 0.5.17-1 - broken audio with Scarlett Solo 4th Gen

small update with some of the things ive tried:

• my laptop has the most recent version of wireplumber, and there its not causing me any problems at all. i can plug the interface in and it functions exactly as expected. so somethings up with it on my pc specifically, but im still not sure what it might be

• i looked at the .local/state/wireplumber folder to see how it behaved while on the most recent version of wireplumber. if im on 0.5.15-1 and i remove the default-nodes and default-routes files, they get replenished pretty much instantly (i understand this happens as soon as audio plays?). this does not happen for me on 0.5.17-1; the folder instead just stays empty. if i update to the newest version with files already in the state folder, the situation does not change, as if wireplumber suddenly becomes oblivious to the directory's presence. as i described before, downgrading to 0.5.15-1 and restarting wireplumber fixes these issues, and the folder gets populated with the usual files it should have

• checking the 'Recording' tab in pavucontrol while opening the audio volume gui in plasma reveals further odd behavior. on 0.5.15-1, each volume meter receives audio from an appropriate monitor or input. on 0.5.17-1, just like with the outputs earlier, every volume meter has some "Unknown input" selected for it. as before, all of the actual inputs do exist in the dropdown and can be selected, but this has no effect, and they always get swapped out for "Unknown input" after a second or two

for some reason these local issues only happen with versions newer than 0.5.15-1, but simultaneously (of the systems i have access to) appear to only happen here on my pc, with my laptop being completely unaffected. it feels like a configuration issue, but i do not have a .config/wireplumber folder, and getting rid of the files in .local/state/wireplumber doesn't seem to do anything useful

for a potentially relevant thing i forgot to include at first, i also checked the output of pactl list sink-inputs when the audio is faulty and when it is working to check the differences in meld:

• the sink number was 905 for the functioning audio and 4294967295 for the faulty audio

• the volume for the broken audio was "front-left: 0 /   0% / -inf dB,   front-right: 0 /   0% / -inf dB"

• the broken audio had 'media.name = "AudioStream"' instead of the correct document title

• the broken sink had no mention of pulse.attr.maxlength, pulse.attr.tlength, pulse.attr.prebuf or pulse.attr.minreq

Last edited by wand_am (Today 11:36:55)