Audio stutters with USB Dacs on Macbook M1 Pro

[nerdstrike]

I'll add a progress report also. So my M2 MBP 13" 24GB had the stutters whenever Safari was loaded up with a gazillion pages/tabs. After that, opening a new page/tab or scrolling while playing music—via Music app, not just web players—would do it.

It's certainly not for want of parallel processing capacity! Something pretty weak must be going on in the USB chipset/drivers on this new chipset. Or perhaps a cataclysmic scheduling of certain work to a single favoured core? At the very least it's well out of reach for the poor user to fix.

 

[DWPress]

Keeping browser tabs open will cripple most machines intel or silicon. I keep a 2018 mini with 32GB RAM in my stack as well which runs Mojave for my aging Adobe CS6 installation, a Plex server and a couple VMs for the most part. By comparison the silicon machines do run smoother for daily activities though I have had some bugs but I'm not sure if they are hardware or OS related.

The basic M1 machine I had with 8GB did suffer from lack of real memory though and 8 channels of convolution filters combined with a work load would exhibit the above issues. As I said above - increasing the audio buffer size in a plugin host eliminated the problem but increased latency which is a problem if you use your system for video.

 

[leolozes]

Macbook Pro 14" with M1 Pro and 32Gb RAM user here. It has nothing to do with the cables IMHO.

The issue is widely discussed here, it's been a problem for years and it's seemingly not going to be fixed anytime soon, which is very puzzling to me since these machines are heavily marketed to creators and music in general.

My experience has been:
- If I wire my Sundaras directly to the 3.5mm jack of the macbook, there's no sound cuts.
- If I use my Sony H9 wirelessly (which dongle is connected to a KVM switch, then to a Thunderbolt dock), there's no sound cuts.
- With both FiiO K3 and K7, and with my Kanto TUKs, which have DACs connected through USB, there's sound cuts anytime I do something remotely taxing. IMHO it's not any hardware limitation or incompatibility, it's just very poor implementation of sound through USB.

As I said, I have everything wired through a KVM switch, which is also connected to my Windows desktop, and that machine has never had any sound issues.
It's not a hardware problem, it's a software problem. And yes, I have tried without the KVM switch. I can't try without the dock, because Apple laptops have no USB A ports .

It's VERY frustrating and infuriating that a machine that cost more than 3.500€ has these kind of problems. You can try to contact the Apple Support, but the amount of system logs and use cases they ask is really baffling (I have provided more than 4 system logs so far).

 

[danielgbaena]

I just purchased a pair of ELAC DEBUT ConneX DCB41 as desktop speakers but my MacBook Pro 14" (2021) cannot reach them using an UDG GEAR U96001BL cable USB 2.0 (C-B) nor a Cable Matters USB-C to USB-B printer cable. I know this is not exactly the same problem but I suppose it could be related.

Of course, the speakers work on a Windows PC with the same cables and all my MacBook Pro ports are perfectly functional.

 

[badspeakerdesigner]

Macbook Pro 14" with M1 Pro and 32Gb RAM user here. It has nothing to do with the cables IMHO.

The issue is widely discussed here, it's been a problem for years and it's seemingly not going to be fixed anytime soon, which is very puzzling to me since these machines are heavily marketed to creators and music in general.

My experience has been:
- If I wire my Sundaras directly to the 3.5mm jack of the macbook, there's no sound cuts.
- If I use my Sony H9 wirelessly (which dongle is connected to a KVM switch, then to a Thunderbolt dock), there's no sound cuts.
- With both FiiO K3 and K7, and with my Kanto TUKs, which have DACs connected through USB, there's sound cuts anytime I do something remotely taxing. IMHO it's not any hardware limitation or incompatibility, it's just very poor implementation of sound through USB.

As I said, I have everything wired through a KVM switch, which is also connected to my Windows desktop, and that machine has never had any sound issues.
It's not a hardware problem, it's a software problem. And yes, I have tried without the KVM switch. I can't try without the dock, because Apple laptops have no USB A ports .

It's VERY frustrating and infuriating that a machine that cost more than 3.500€ has these kind of problems. You can try to contact the Apple Support, but the amount of system logs and use cases they ask is really baffling (I have provided more than 4 system logs so far).

And here I am looking to buy a macbook to get away from DPC latency issues with windows laptops. Welp.

 

[jayadubya]

FWIW some other users with the exact same issue have been highlighting this issue over at Fiio's forums. Fiio does not seem to be helping or discussing the issue, often just asking people to email in. I've done that before and had no useful exchange except 'send us a video showing the issue' and then that's about it. No useful resolution.

Today I'm using the K7 and I haven't had an issue or stutter since fully restarting my computer. The inconsistency is what really makes it infuriating. I'm considering getting a different DAC just to see if it's still an issue, but this is getting expensive. I feel like that might be a fool's errand if it happens again. I sometimes just use my phone for music, but the idea is to use this desktop dac to drive all of the sound from my MacBook - particularly for calls and meetings etc.

I'd like to escalate it with apple (I'm still well within my warranty). I wonder if 10 people pushed the issue and tried to open tickets if it would hit their radar?

 

[leolozes]

FWIW some other users with the exact same issue have been highlighting this issue over at Fiio's forums. Fiio does not seem to be helping or discussing the issue, often just asking people to email in. I've done that before and had no useful exchange except 'send us a video showing the issue' and then that's about it. No useful resolution.

Today I'm using the K7 and I haven't had an issue or stutter since fully restarting my computer. The inconsistency is what really makes it infuriating. I'm considering getting a different DAC just to see if it's still an issue, but this is getting expensive. I feel like that might be a fool's errand if it happens again. I sometimes just use my phone for music, but the idea is to use this desktop dac to drive all of the sound from my MacBook - particularly for calls and meetings etc.

I'd like to escalate it with apple (I'm still well within my warranty). I wonder if 10 people pushed the issue and tried to open tickets if it would hit their radar?

I'm not sure it's a FiiO problem, as the issue happens also with my powered speakers, and in the link I provided you can see plenty of people having issues with other DACs and equipment. It's true that it happened a lot more and sooner with the K3 than with the K7 or the speakers, but I still think it's mainly an Apple issue.

AFAIK the problem has been reported dozens of times and it hasn't helped so far, but try to open yours, personally I just got tired of sending system diagnostics and logs. The issue is probably getting lost in the communication between the support and the engineering team...

Personally I've tested multiple times and the issue happens a lot sooner when (but it's not exclusive to) I'm developing with Java, compiling and running multiple applications, or in meetings with Google Meet. That's what makes me believe that it's probably a software related issue with some bad management of the CPU priority scheduling. Windows sound service does use fairly more CPU than Apple's core audio, but at least it makes sure that the sound is never cut.

 

[Axo1989]

... That's what makes me believe that it's probably a software related issue with some bad management of the CPU priority scheduling. ...

This I think is the likely problem area too. There's almost nothing I can think of that I do that couldn't wait a millisecond to allow continuous audio (it's more complex than that obviously, but still). Quoting myself:

My guess is it's an issue with macOS and Core Audio scheduling of application-level services that arises in my particular configuration (and some others) and one or more of the audio services isn't prioritised as I'd prefer.

 

[palm]

There are few things worse than debugging an intermittent issue.

You could try the Apple usb-c to usb-a adapter if not done yet.

And also when the laptop isn’t charging. Saying this because I had several issues with a monitor delivering power to the Mac : it would work better with the right ports than left ports, but still generating interferences for bluetooth devices and other weird stuff that *really* all disappeared with the studio display. So try on battery power and if it helps use the charger and cable shipped with the Mac if you were charging via a dock.

 

[jayadubya]

Battery power! That's an interesting idea. I'll give it a go!

 

[palm]

I don’t like suggesting this kind of experiments like the resets one can do on a Mac and hope for a result.
But I really experienced weird usb-c charging issues, like the increased kernel_task load when using ports on the left.
So it’s not about using battery power to reduce supply noise but more eliminating one source of complications.
I also had once to make an SMC reset to recover a dead usb-c port after plugging a HP dongle supposedly supporting power delivery.

 

[leolozes]

Tried the battery option, no luck. I still have to try to bypass the dock completely, but that's not really a viable option in the long run for me .

Anyway, just for fun, I noticed that if you force coreaudiod to have the maximum process priority, the sound cuts stop for a while (or at least drastically diminsh). I guess with time another process "regains" priority and it starts again, but it somewhat validates my theory that it's mainly a software problem.

If anyone wants to try:
In your terminal, identify the coreaudiod process with ps -ef | grep coreaudiod. This should return something like:
202 650 1 0 Mon07AM ?? 14:14.00 /usr/sbin/coreaudiod

The important number is the PID, in this case 650.

Then you can change the priority to the max, which is -20 with:
sudo renice -n "-20" -p 650

Where 650 is the PID that was returned in the previous command.

It would be interesting to know if it also fixes temporarily your issue or if there are different use cases.

 

[diffidentDude]

Tried the battery option, no luck. I still have to try to bypass the dock completely, but that's not really a viable option in the long run for me .

Anyway, just for fun, I noticed that if you force coreaudiod to have the maximum process priority, the sound cuts stop for a while (or at least drastically diminsh). I guess with time another process "regains" priority and it starts again, but it somewhat validates my theory that it's mainly a software problem.

If anyone wants to try:
In your terminal, identify the coreaudiod process with ps -ef | grep coreaudiod. This should return something like:
202 650 1 0 Mon07AM ?? 14:14.00 /usr/sbin/coreaudiod

The important number is the PID, in this case 650.

Then you can change the priority to the max, which is -20 with:
sudo renice -n "-20" -p 650

Where 650 is the PID that was returned in the previous command.

It would be interesting to know if it also fixes temporarily your issue or if there are different use cases.

I tried this as well and found that it didn't make a difference.
The memory pressure seems to be at the root of it

 

[Axo1989]

I tried this as well and found that it didn't make a difference.
The memory pressure seems to be at the root of it

One of the virtual memory/swap or memory compression process may be getting in the way. As this appears to be an Apple Silicon issue (M1, M2 etc) I note they do have somewhat novel memory architecture which generally performs well but maybe not in this specific case. Since I started using Safari like a parsimonious Luddite on my M2 machine I've had no stutters.

 

[diffidentDude]

One of the virtual memory/swap or memory compression process may be getting in the way. As this appears to be an Apple Silicon issue (M1, M2 etc) I note they do have somewhat novel memory architecture which generally performs well but maybe not in this specific case. Since I started using Safari like a parsimonious Luddite on my M2 machine I've had no stutters.

is that because overall the memory pressure is lower or something to do with Safari?

Something weird for me, I don't get any issues using my airpods, I only get issues when listening with my DAC and headphones, and only when the memory pressure is up

 

[Axo1989]

is that because overall the memory pressure is lower or something to do with Safari?

Something weird for me, I don't get any issues using my airpods, I only get issues when listening with my DAC and headphones, and only when the memory pressure is up

Oh yes, per my posts upthread, Safari is just the usual way that memory pressure goes into the yellow/red zones on that machine.

On AirPods vs DAC good catch, I have similar experience, so yes it appears to be happening with USB audio rather than wireless.

 

[jayadubya]

Trying Safari was one of the suggestions from the Genius Bar folks, but they also admitted it's usually pretty memory hungry. I definitely notice more stuttering when I have two browsers open. Right now I have Vivaldi and Chrome (both chromium builds) open and it's a sure-fire path the stutter town.

I also did some work to get all of the Intel-based apps/processes out of my operations and just focus on the Apple/silicon compatible stuff. Stuttering continues, but I also think it may not be possible to entirely eliminate Intel stuff and/or Rosetta without causing pain.

I should try the safari option!

 

[diffidentDude]

I wonder where the problem is, with USB, with the audio streaming, or somewhere else?

The fact that wireless seems to perform without stuttering, seems to indicate that the problem is with USB

 

[SoonToBeRetired]

I need help getting external DACs to work correctly on my 2019 Macbook Pro M1.

I experience audio stutters as soon as I have a high system load. The stutters occur once every few minutes, each lasting only very short. The audio quality itself is excellent, and the DAC itself work fine.

The problem happens with the following DACs:

- Topping E70
- SMSL SU-8
- iFI Zen V2
- My USB-C-based Qudelix 5K Dac works fine, though - no issue at all.

Things I've tried:

- Change the audio format in Audio Midi Setup - no difference
- connect the DACs to a different USB port - either on a passive USB 2.0 Hub that is linked to my Thunderbolt Docking Station or going to the Thunderbolt Docking Station directly
- Restarting and updating the firmware (where possible)
- Disabling "Set time and date automatically" - this seems to improve the stutters minimally
- Try different audio sources (they all stutter)
- Using Bluetooth Audio output - this works fine.

The problem is undoubtedly related to the Macbook because a Windows Laptop has no issues using the DACs with the same setup/docking station.

If I had to guess, my Mac has an issue with the USB 2.0 port, which is the common connector of all DACs (except the Qudelix).

Do you have any other ideas of what I could try? I'm at my wit's end, ordering several DACs from Amazon to see if one device magically works.

I know you said that you tried to change the audio format in MIdi setup, but I think it has something to do with that. I have an ancient late 2010 27" iMac which exhibited stuttering when streaming with the Qobuz app and by changing the format to two channel 24 bit integer 96.0 kHZ solved my problem and the audio does not stutter no matter what else the computer may be doing. Make sure you are changing the format for the correct output.

 

[Axo1989]

Just fyi I realise this can be misread without the context of my previous posts:

... Since I started using Safari like a parsimonious Luddite on my M2 machine I've had no stutters.

Aside from deliberate browser testing, I've only used Safari for browsing, I didn't switch from another browser. What I meant was that I normally/previously left windows and tabs open en masse, leading to memory pressure creeping up as more and more windows and tabs are opened. As memory pressure appears to correlate with audio stutters (which were happening if I simply scrolled a web page or switched to another tab if I had enough of them open) now I'm being parsimonious—unwilling to use resources—and opening and closing web pages and tabs as I go. It feels primitive—hence Luddite, used in the vernacular not the historical/political sense—but it means no stutters lately.

So I think it's a combination of doing something that macOS prioritises while memory pressure is on the high side that I associate with the issue.