Repository navigation
Dell XPS 13 Plus 9320 (SKU 0AF3): no speaker tuning ships — internal speakers sound thin and mid-forward
#15727
nazariitaran
started this conversation in
Suggestions
Replies: 1 comment 4 replies
|
Hey guys - opening this issue because my agent suggested it's worth flagging. I installed Omarchy on my 4-year-old Dell, and the audio sounds different compared to Windows on the same machine. Fixed with the agent |
4 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Summary. Omarchy ships no speaker tuning that matches the Dell XPS 13 Plus 9320
(SKU
0AF3), so the internal speakers play the raw Linux path, which on this unit isstrongly mid-forward and sounds clearly worse than the same machine's Windows
install.
omarchy audio tuning statusreportsMatches: nothing ships for this laptop. The only shipped tunings aredell-xps-2026(XPS 14/16, SKUs0DB9/0DBA)and
asus-proart-p14-n1x.Hardware.
0CW9KM, SKU0AF34.0.4-1sofsoundwire(Intel Alder Lake-P SOF SoundWire), topologysof-adl-rt1316-l12-rt714-l0.tplgrt1316(2 channels each); micrt715-sdcaWhat it sounds like. Thin, quiet and heavily mid-focused — the usual description
is "telephone". This is a voicing gap, not the silent-sidecar-amp class of bug:
every amp/DAC/mixer switch is on and the HiFi profile is active, so audio does flow.
Linux simply applies none of the vendor voicing Windows gets.
Measurements (built-in mic, relative, not calibrated). Flat multitone through
the raw physical speaker sink, recorded with the built-in mic, each band normalised
to 700 Hz:
The response peaks around 700 Hz–1.5 kHz and collapses either side — a telephone
band. The built-in mic has its own rolloff, so treat this as shape, not absolute
level.
Local workaround (works, but tuned by ear). A filter-chain dropped into
~/.config/pipewire/omarchy-speaker-tuning.conf.d/90-tuning.conf, using Omarchy'sown paths and names (
omarchy_speaker_tuningsink,omarchy-speaker-tuning.service),so
omarchy audio tuning statusrecognises it. Chain:loud_comp_stereoequal-loudness contour (ISO 226:2023)sqrt→ high-pass, mixed back in withmixer), which is what finally stopped it sounding like a phone, since the drivers can't move much air down lowlimiter_stereoMeasured effect of the enhancer: playing a pure 100 Hz tone, 2nd-harmonic content
rose ~19 dB (dry vs tuned, via the mic) — the ear then reads the low note that the
driver cannot physically produce.
Why not a PR. It's a reasoned, ear-shaped voicing, not a measured correction. I
don't have a calibrated measurement mic, so I can't produce the
magnitude_rms_db/bass_group_delay_swing_ms/limiter_headroom_db/dynamic_range_delta_lufieldsdocs/audio-tuning.mdasks a tuning to report.Precedent. #8621 is the closest filed item, but it is a different machine and a
different cause (DX13260 / SKU
0E53, CS35L56 sidecar amps never enabled).Ask. Please ship a tuning for SKU
0AF3, or say which target curve to fit so thelocal one can become a proper PR.
All reactions