Desktop Environment (if applicable):
GNOME, KDE Plasma
Problem Description:
I have an ASUS ROG G18 2023 laptop with an Intel i9, and RTX 4080. Whenever I try running Linux Distros I get really bad audio crackling on internal audio/USB audio interface (HDMI works fine), and the external display seems laggy every now and again when I’m browsing the internet/doing normal day-to-day things on it (mouse lag is the most visible, but different apps also seem to be stuttering).
The distro I’ve had the most luck until now is Ubuntu though (I’ve tried Fedora, and CachyOS as well), but they are much worse. The basic problem is that when I run multiple apps and the system is under load (for example I play YouTube on one screen, and I open a Steam game) this horrible audio crackling starts happening.
What I’ve Tried:
Some fixes I have tried are PipeWire fixes that have been suggested by different users across different distros, disabling USB Suspend, different profiles (including switching entirely to dGPU), and a whole bunch of GRUB changes such as C_STATES. I had no luck, and there’s never any errors when looking into pw-top.
I am really trying to ditch Windows except for some rare cases, so any help is really appreciated, thank you!!
Relevant System Information:
ASUS ROG G18 2023, i9-13980HX, NVIDIA GeForce RTX 4080, 64GB DDR5, 1TB SSD, and 2TB SSD
I think you should start by checking if the “Pro Audio” PipeWire profile is involved in this. It shouldn’t be used for plain desktop purposes and can cause crackling, when overdoing the latency targets. Some people seem to think of “Pro” as an upgrade of sorts, but it’s really not; it’s what studio professionals need for serious audio processing in realtime with tight feedback (monitor) loops.
I looked into the “Pro Audio” but it wasn’t set to that. I’ll try running those commands when I have some time to stress test the laptop and attach those here. Thanks!
Actually, attaching files isn’t really possible on (this) Discourse, so the next best thing is to copy/paste the file contents into a pre-formatted text / code block to preserve alignment and thus readability. Make sure you are on an empty line, then click </> button and paste the content of the respective file.
Optional helper script
If you are so inclined, I took it upon myself to provide a little helper script for that.
Hi, I’m back, so I tested it some more, and when crackling I ran the commands (I was playing Aska, while watching YouTube on Firefox), here’s the output:
journal.txt
Started pipewire.service - PipeWire Multimedia Service.
ALSA lib parser.c:3264:(parse_master_file) [error.ucm] unknown master file field omment
ALSA lib parser.c:3264:(parse_master_file) [error.ucm] unknown master file field omment
ALSA lib parser.c:3264:(parse_master_file) [error.ucm] unknown master file field omment
ALSA lib parser.c:3264:(parse_master_file) [error.ucm] unknown master file field omment
ALSA lib parser.c:3264:(parse_master_file) [error.ucm] unknown master file field omment
nvidia_smi.txt
Thu Jun 11 09:31:45 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 595.71.05 Driver Version: 595.71.05 CUDA Version: 13.2 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 4080 ... Off | 00000000:01:00.0 On | N/A |
| N/A 86C P0 140W / 160W | 7619MiB / 12282MiB | 100% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 8605 G /usr/bin/kwin_wayland 3MiB |
| 0 N/A N/A 12038 C+G ...am/ubuntu12_64/steamwebhelper 7MiB |
| 0 N/A N/A 59520 C+G S:\common\ASKA\Aska.exe 7380MiB |
+-----------------------------------------------------------------------------------------+
pw_top.txt
S ID QUANT RATE WAIT BUSY W/Q B/Q ERR FORMAT NAME
C 30 0 0 --- --- --- --- 0 Dummy-Driver
C 31 0 0 --- --- --- --- 0 Freewheel-Driver
C 49 0 0 --- --- --- --- 0 Midi-Bridge
C 112 0 0 --- --- --- --- 0 bluez_midi.server
C 55 0 0 --- --- --- --- 0 alsa_output.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__sink
C 62 0 0 --- --- --- --- 0 alsa_input.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__source
C 58 0 0 --- --- --- --- 0 alsa_output.pci-0000_00_1f.3.analog-stereo
C 74 0 0 --- --- --- --- 0 alsa_input.pci-0000_00_1f.3.analog-stereo
C 61 0 0 --- --- --- --- 0 alsa_output.pci-0000_01_00.1.hdmi-stereo
C 89 0 0 --- --- --- --- 0 v4l2_input.pci-0000_00_14.0-usb-0_8_1.0
C 102 0 0 --- --- --- --- 0 Firefox
C 146 0 0 --- --- --- --- 0 Firefox
C 113 0 0 --- --- --- --- 0 Firefox
C 110 0 0 --- --- --- --- 0 Aska.exe
S ID QUANT RATE WAIT BUSY W/Q B/Q ERR FORMAT NAME
S 30 0 0 --- --- --- --- 0 Dummy-Driver
S 31 0 0 --- --- --- --- 0 Freewheel-Driver
S 49 0 0 --- --- --- --- 0 Midi-Bridge
S 112 0 0 --- --- --- --- 0 bluez_midi.server
R 55 256 48000 56.0us 6.0us 0.01 0.00 0 S32LE 2 48000 alsa_output.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__sink
R 102 900 48000 37.8us 7.2us 0.01 0.00 0 F32LE 2 48000 + Firefox
R 113 3600 48000 45.1us 3.4us 0.01 0.00 0 F32LE 2 48000 + Firefox
R 110 256 48000 48.6us 3.0us 0.01 0.00 0 S16LE 2 48000 + Aska.exe
S 62 0 0 --- --- --- --- 0 alsa_input.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__source
S 58 0 0 --- --- --- --- 0 alsa_output.pci-0000_00_1f.3.analog-stereo
S 74 0 0 --- --- --- --- 0 alsa_input.pci-0000_00_1f.3.analog-stereo
S 61 0 0 --- --- --- --- 0 alsa_output.pci-0000_01_00.1.hdmi-stereo
S 89 0 0 --- --- --- --- 0 v4l2_input.pci-0000_00_14.0-usb-0_8_1.0
I 146 0 0 0.0us 0.0us ??? ??? 0 F32LE 2 48000 Firefox
S ID QUANT RATE WAIT BUSY W/Q B/Q ERR FORMAT NAME
S 30 0 0 --- --- --- --- 0 Dummy-Driver
S 31 0 0 --- --- --- --- 0 Freewheel-Driver
S 49 0 0 --- --- --- --- 0 Midi-Bridge
S 112 0 0 --- --- --- --- 0 bluez_midi.server
R 55 256 48000 36.0us 4.1us 0.01 0.00 0 S32LE 2 48000 alsa_output.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__sink
R 102 900 48000 19.5us 6.3us 0.00 0.00 0 F32LE 2 48000 + Firefox
R 113 3600 48000 25.9us 2.5us 0.00 0.00 0 F32LE 2 48000 + Firefox
R 110 256 48000 28.5us 2.3us 0.01 0.00 0 S16LE 2 48000 + Aska.exe
S 62 0 0 --- --- --- --- 0 alsa_input.usb-Universal_Audio_Volt_2_24092037089927-00.Direct__Direct__source
S 58 0 0 --- --- --- --- 0 alsa_output.pci-0000_00_1f.3.analog-stereo
S 74 0 0 --- --- --- --- 0 alsa_input.pci-0000_00_1f.3.analog-stereo
S 61 0 0 --- --- --- --- 0 alsa_output.pci-0000_01_00.1.hdmi-stereo
S 89 0 0 --- --- --- --- 0 v4l2_input.pci-0000_00_14.0-usb-0_8_1.0
I 146 0 0 0.0us 0.0us ??? ??? 0 F32LE 2 48000 Firefox
That “[error.ucm] unknown master file field omment” looks interesting. Apparently there is some master file — whatever that is; no PW expert — which contains a malformed “Comment” field, that’s missing the first character. I don’t know if it’s related to your issue, though.
But maybe that causes PW to fallback on some defaults that don’t quite fit your device?
Be aware, though, that it’s going to be decommissioned at the end of the month:
That’s good to know, I’ll avoid using pastebin. As for the comment, good catch, I actually had the same issue when I tried Fedora/CachyOS, and I fixed the comment. I’ll try doing the same here and I’ll see if it fixes it.
One thing I noticed is that with Ubuntu it is much more stable for a longer time than other distros I tried, so the crackling only starts after ~15-20 mins of being under load, while on other distros it was mostly always present.
As for the last thing I observed, when doing things on my external display (such as browsing Firefox, or anything) the display seems to become choppy/micro-stutter, but if I move the window to my main laptop display it works fine (don’t know if it’s related, just hope someone stumbles onto this thread that maybe had the same issue and know what causes it).
That looks rather suspicious to me. If I am reading it (alsa_output...Direct__Direct__sink) correctly your apps are talking ALSA, as opposed to PW’s native protocol, directly to your USB audio device, or PW’s representation of which. Plus the latency settings seem quite low as well.
If my math is correct, a quantum of 256/48000 equates to ~5 milliseconds, which seems rather ambitious, if my past experiments with a Soundblaster Pro on JACK are anything to go by — those are way in the past, though.
This is what it looks like for me:
S ID QUANT RATE WAIT BUSY W/Q B/Q ERR FORMAT NAME
R 68 2048 48000 798,8us 18,7us 0,02 0,00 0 S32LE 2 48000 alsa_output.pci-0000_05_00.6.analog-stereo
R 34 0 0 46,7us 30,5us 0,00 0,00 0 F32P 8 0 + effect_input.virtual-surround-7.1-hesuvi-dtshx
R 35 0 0 15,0us 590,9us 0,00 0,01 0 F32P 2 0 + effect_output.virtual-surround-7.1-hesuvi-dtshx
R 109 3600 48000 63,0us 50,6us 0,00 0,00 0 F32LE 2 48000 + Firefox
The shortcomings of ALSA are why PipeWire exists. What version of Firefox are you using and have you configured it to use its legacy ALSA interface?
Plus, I see Aska.exe, so maybe Wine plays a part in this as well?
That “[error.ucm] unknown master file field omment” looks interesting.
That’s harmless, just a typo.
This isn’t a Pipewire issue, pw-top shows no ERRs and a low busy times (3-6us vs 256/48kHz quantum)
It looks like your GPU is getting throttled, it’s pinned at 100%, which would explain the stuttering on the display, and generally disrupt system scheduling, causing delays in the audio threads, which would also explain the crackling.
I don’t have ASUS hardware to test, but you might try experimenting with asusctl. It looks like you can enable a performance mode that might give you some more headroom.
You’re right; when I circumvent it the quantum drops to 1024. I still feel like 5 ms is tickling the dragon. I remember it getting really hairy when aiming for <10ms with said Soundblaster Pro, which had the benefit of a real hardware mixer, I might add. And when I overdid it with latency the crackling (xruns) started coming thick and fast and CPU load rose expenentially as it seemed with the reciprocal of latency. But then again, looong time, so just take it FWIW.
Since I don’t seem to be able to render any actual help, I think I’ll just sit back and see if I can learn something form here on out.
Thank you both, I’ll try with a higher quantum just in case (can’t hurt, plus I’m not doing anything that requires that low of a latency), as for asusctl I’ll retry that, I’ve tried it in the past but it honestly just felt like it made it worst, but I’ll post more findings here.
To answer the Firefox question, I am running 151.0.4, and I have not changed the configuration to use its ALSA interface, but I can try if you think it would be a valid test?
I’m also wondering why Firefox/KDE randomly gets laggy when on the external display (that maybe would be somewhat related), and Chrome was waaay worse. I’ll also retry fixes I’ve tried on other distros to see if somehow one of them fixes it.
I think you shouldn’t set any quantum manually. That’s just asking for trouble and burns CPU cycles for no noticeable gain. Apps which really require low latency will request what they need, usually.
Oh no, disregard my comment about that; I was misreading the log. I actually suspected that using the ALSA API was contributing to the problem and would have suggested to switch to the PulseAudio interface. You can double-check, if it’s enabled by opening about:support#media, which should say "Audio backend: pulse-rust`.
Maybe that is something to do with HDMI. Maybe this thread can help, if that’s the case:
OK, as promised I’m back, I’m just going to document everything I find in case any other poor soul that has to go through the same issues with an ROG laptop might stumble onto this thread.
So for the desktop issues, I found that enabling KDE with X11 was actually a really good step in the right direction (though scaling is a bit messy if you use different refresh rate and size/resolution monitors). This seems to have fixed most of it, though I know the same thing failed on different distros, it seems that Ubuntu plays much nicer with it, so that’s great news!
Still seeing a bit of crackling even with this, but it seems to be much more manageable (it goes away on it’s own at least for now).
I’ll keep updating this post for the next week or so since I wouldn’t say this is fixed just yet, but if it is I’ll post and mark as solved.
Thanks again for for the support charles05, and peterwhite23!
What about changing said quantum? I’ve been doing some reading of the docs, and it looks like Aska.exe — or Wine more likely — is requesting that:
When there is a value in the QUANT column it shows what the client explicitly requested. PW, by design, then uses the minimum quantum of all requests, disregarding their wishes and, if a client can’t fill the buffer in time xruns (crackling) happen.
I also changed the quantum, what I’m about to do is run a fresh install of Kubuntu since I broke something while trying to set up X11 after running Wayland manually. And I’ll re-run the tests to see which fix helped, or if both helped (switching to X11 and changing Quantum) and write back here for anyone else encountering the issue to see.
Sorry for the delay, but I just took more time to test things. So it seems like the quantum change did not impact anything. I got to a point where I just had to reinstall things again, and it seems like the most stable version I found is KDE w/ X11. Everything seems ok and crackling is at a minimum. One thing I did mention is the crackling gets worse while multitasking and IO operations are running. I’ve noticed really intense crackling if I’m watching a video and I download something on Steam. So I’ll look more into that and update here as I go.
