https://bugs.kde.org/show_bug.cgi?id=523050
Bug ID: 523050
Summary: Opening the audio applet flyout causes an audible
transient glitch/"thinning" of playing sound via
PlasmaPA-VolumeMeter monitor stream
Classification: Plasma
Product: plasmashell
Version First 6.6.5
Reported In:
Platform: Kubuntu
OS: Linux
Status: REPORTED
Severity: normal
Priority: NOR
Component: Audio Volume widget
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Target Milestone: 1.0
Summary:
Every time the Audio Volume panel applet flyout is expanded (and a peak-level
monitor stream is created for each active sink/source), currently playing audio
on the built-in speaker sink audibly changes for a moment - it sounds
noticeably "thinner"/"shallower" before settling back to its normal ("fuller")
character a second or so later. This happens reproducibly and specifically
correlates with plasma-pa attaching a new `PlasmaPA-VolumeMeter` monitor stream
to the sink.
Steps to reproduce:
1. Play music through the built-in laptop speakers (PipeWire/pipewire-pulse
backend).
2. Click the Audio Volume icon in the system tray to expand the device list
flyout.
3. Listen closely to the music at the moment the flyout opens.
Observed result:
The sound briefly becomes noticeably thinner ("shallower") for roughly a
second, then returns to full/normal character. This can be triggered repeatedly
by opening/closing the flyout.
Expected result:
No audible change to currently playing audio when the volume flyout / peak
meters are shown.
Investigation notes (may help triage):
- Confirmed via `pactl subscribe` that opening the flyout creates 5 new
`source-output` objects that are removed again when the flyout is
closed/collapses.
- Confirmed via `pw-mon` that one of these streams attaches directly to the
built-in Speaker sink node with these properties:
media.name = "PlasmaPA-VolumeMeter"
application.id = "org.kde.plasma-pa"
application.process.binary = "plasmashell"
stream.monitor = "true"
resample.peaks = "true"
node.rate = "1/25"
- Ruled out as the cause (measured directly, no change observed during
reproduction):
- PipeWire graph quantum/clock.quantum (polled every 200ms during
reproduction, stayed fixed at 1024 the whole time)
- Sink-input (Spotify) stream volume/ducking (polled every 200ms during
reproduction, stayed at 100% / 0.00 dB the whole time)
- ALSA mixer control values (`amixer -c0` before/after diff showed no
changes)
- The remaining correlation is the `PlasmaPA-VolumeMeter` stream attaching
to/detaching from the sink node, which appears to force a brief reconfiguration
of the underlying ALSA node (`alsa:acp:sofhdadsp:3:playback`,
device.profile.name = "HiFi: Speaker: sink").
System information:
- OS: Ubuntu 26.04 LTS
- Kernel: 7.0.0-27-generic
- Plasma (plasmashell): 6.6.5
- WirePlumber: 0.5.13
- PipeWire (pipewire-pulse server version): 15.0.0
- Audio hardware: Intel Meteor Lake-P HD Audio Controller, Realtek ALC274
codec, sof-audio-pci-intel-mtl / snd_soc_skl_hda_dsp driver (SOF)
- firmware-sof-signed: 2025.12.2-1
- Laptop model: MSI Prestige 13 AI Evo A1MG
Additional context:
Multiple prior plasma-pa bugs have involved the PlasmaPA-VolumeMeter
monitor-stream mechanism interacting poorly with PipeWire (e.g. bug 439347 -
streams not freed/CPU usage, bug 449532 - crash with headset profile, bug
457023 - leaked sinks with WEBRTC). This report may be a related/adjacent
symptom of the same underlying mechanism (peak-meter monitor streams attaching
to real sink/source nodes) rather than an isolated new bug - please dedupe as
appropriate.
--
You are receiving this mail because:
You are watching all bug changes.
