Macbook Air M2 and DP 11 Disappointing Performance

[James Steele]

Well maybe it's specific to the MacBook Air M2 I guess. I wouldn't know. As far as users who have reported this problem here, there is exactly one.

[HCMarkus]

Sigs only get outdated if one doesn't update them! I posted this because I've seen way too many posts where folks don't provide any system-specific info at all. But if one is diligent in posting important details, I'm fine with that, although it can get inconvenient to always have to refer back to the first post in a lengthy thread...

I run Blackhole on a bus in every DP project (It's in my template) without issue. But I am on a M1 Studio Ultra.

That's the thing, people that don't provide information in their OP are the same folks that are unlikely to update their sigs, and then you are running on misinformation. Sigs are really only a measuring contest, and a waste of bandwidth.

How many instances of Blackhole can you run in DP on your Ultra?

It would be nice if you moderated your tone a tad here. "Sigs are only a measuring contest" is the kind of statement that can cause unnecessary hard feelings. In a sense, this sentiment may reveal the emotional context underpinning your aggressive verbiage, which I personally find off-putting. Please consider a slightly gentler approach here. Or not. Your call.

Back to the subject at hand...

Unless I have a specific patch that I want to alter using Blackhole as an Insert Effect, I will only have a single instance running, the aforementioned Aux Bus Reverb. I have definitely run more than a single instance in some projects, but not more than two or three.

The fact that you are seeing different results with different DAWs on the same Mac does suggest a potential issue with M2 and DP. As has been noted, there are many variables here. You may wish to reach out to MOTU directly at MOTU.com with a Techlink detailing your issues. MOTU will accept a file, so you can send a problematic file directly to them for research.

Let me mention that, as you are probably aware, this is an independent site, not sponsored by or commercially affiliated with MOTU.

[cuttime]

May I respectfully ask that this topic be moved to the "Gripes..." section?

[James Steele]

May I respectfully ask that this topic be moved to the "Gripes..." section?

I'll leave it here, but I would like to see the *tone* moderated. We're doing our best to help, and IIRC, the OP returned the Mac, right? I'll try to reach out directly to a contact at MOTU with a link to this thread. Obviously, if DP has a problem with M2 MacBook Airs... (from what I understand M2 Mac Studios are fine)... they need to look into it.

[James Steele]

I've sent a link to this thread to a contact at MOTU so they can look at it. Obviously, if as alleged, DP11 (and no other DAW) has a problem with M2 MacBook Airs, then they need to verify it and address it. I have a certain responsibility to verify this information as MOTU is kind of getting a black eye here. If it's deserved and they do indeed need to fix it, then let the chips fall where they may. But this really bears verification and scrutiny.

[chokobo99]

It would be nice if you moderated your tone a tad here. "Sigs are only a measuring contest" is the kind of statement that can cause unnecessary hard feelings. In a sense, this sentiment may reveal the emotional context underpinning your aggressive verbiage, which I personally find off-putting. Please consider a slightly gentler approach here. Or not. Your call.

Markus, the tone of this thread, and the board really, was set well before I arrived. It's actually why I signed up. Peter was being pummeled for making what were totally innocuous statements about his situation. If you really feel the need to chastise someone maybe start with the early posts before I arrived.

As for the comment I don't find it out of line. In cases like these pushing back against the messenger tends to happen when the message beats close to home. If you are put off by it then maybe it's because of the weight of truth carried. But it's not only here, most foros employ sigs this way. So you don't have to feel alone in this matter. If you really want diagnostic information it is much better placed on the side with the avatar, and kept to a very strict minimum.

The reason for this is people get busy, and they just don't have time to garden their information on the various foros, and what not. I've looked at the sigs in these posts, and while a few people keep them up to date, plenty do not, and that's all it takes for them to be effectively useless, or as I said previously, misdirective—leading people to malformed conclusions. It's much better to put this information in the OP. OP's are very easy to find, update, and refer to if necessary.

Unless I have a specific patch that I want to alter using Blackhole as an Insert Effect, I will only have a single instance running, the aforementioned Aux Bus Reverb. I have definitely run more than a single instance in some projects, but not more than two or three.

Right, that's under normal conditions, I mean as a test how many can you get going before your system says Uncle!

The fact that you are seeing different results with different DAWs on the same Mac does suggest a potential issue with M2 and DP. As has been noted, there are many variables here. You may wish to reach out to MOTU directly at MOTU.com with a Techlink detailing your issues. MOTU will accept a file, so you can send a problematic file directly to them for research.

Which is exactly why I tested with different apps. By eliminating a variable you can hone it on the root cause.

Let me mention that, as you are probably aware, this is an independent site, not sponsored by or commercially affiliated with MOTU.

Indeed, I do realize this is the James Steele Show.

[chokobo99]

Well maybe it's specific to the MacBook Air M2 I guess. I wouldn't know. As far as users who have reported this problem here, there is exactly one.

James, it would seem to be portables. I don't have an Air, as I stated previously I'm on an M2 MBP. But there's also this thread with an M1 MBP suffering rather similar issues.

I have a 2021 MacBook Pro with M1 Max and 64 GB RAM, a maxed-out system. Before macOS 13.3, I was able to record 17 channels of input at 44.1khz and a single stereo bus in real-time mode with 1 to 3 plugins per channel, and a 64-sample buffer and plugin latency compensation ON, and get no audio glitches.

However after updating to macOS 13.3, and then the subsequent versions, now I get really poor performance with frequent CPU clipping causing crackles during recording. If I set the input buffer to 128 samples, then it's fine, but I shouldn't have to do that.

This machine should be more than capable of handling this task as even my prior Intel-based system could do it.

Does anyone have any similar experience or idea of why this might be happening? I tried updating DP to 11.21 and 11.22 to address the issue, but no dice. This is really frustrating, overall the payoff of Apple Silicon just doesn't seem to be worth it, at least for DP... thanks for any tips.

[chokobo99]

I've sent a link to this thread to a contact at MOTU so they can look at it. Obviously, if as alleged, DP11 (and no other DAW) has a problem with M2 MacBook Airs, then they need to verify it and address it. I have a certain responsibility to verify this information as MOTU is kind of getting a black eye here. If it's deserved and they do indeed need to fix it, then let the chips fall where they may. But this really bears verification and scrutiny.

As I posted previously, I have been in contact with Motu about this issue. For FIVE weeks they only would test it on an Intel machine running the other OS. And of course they didn't find any issues—go figure. ¯\_(ツ)_/¯

They are aware of the issue, but don't seem to be overly concerned about it so far.

And it would seem that more evidence has popped up. But this time on an M2 mini.

[James Steele]

Yeah... there's another thread on here covering this as you mention. Also I just noticed that this topic is in the wrong forum. Should be in the troubleshooting forum, so I'm moving it there.

[James Steele]

I wrote a friend about this issue and this is what he said:

I've seen this happening in two scenarios, one with an external monitor and another without. The culprit on both was screen scaling which makes the whole SoC run at a different pace with lots of load on the limited RAM. Add to this that the M2 Air has a single-channel SSD and with the limited RAM they get a lot of swapping, which again limits the processing speed and the lifetime of the SSD which is not replaceable. Having only two ports to handle all this doesn't make it any easier either and might add stress to the bottleneck.

Someone on the thread mentioned the single-core performance problem, and on those M2 you only get to use 4 cores in total anyway. If you hit a core-intensive plugin the system will try to speed up the single-core performance but multi-core runs at a slower clock speed. Then if you're taxing the GPU with scaling, you'll be hitting a resource conundrum with clocking running wild. Yes, they talk about 8-12 cores but only 4 are performance; the rest are "efficiency" cores that won't be used for any DSP at all.

[Michael Canavan]

I wrote a friend about this issue and this is what he said:

I've seen this happening in two scenarios, one with an external monitor and another without. The culprit on both was screen scaling which makes the whole SoC run at a different pace with lots of load on the limited RAM. Add to this that the M2 Air has a single-channel SSD and with the limited RAM they get a lot of swapping, which again limits the processing speed and the lifetime of the SSD which is not replaceable. Having only two ports to handle all this doesn't make it any easier either and might add stress to the bottleneck.

Someone on the thread mentioned the single-core performance problem, and on those M2 you only get to use 4 cores in total anyway. If you hit a core-intensive plugin the system will try to speed up the single-core performance but multi-core runs at a slower clock speed. Then if you're taxing the GPU with scaling, you'll be hitting a resource conundrum with clocking running wild. Yes, they talk about 8-12 cores but only 4 are performance; the rest are "efficiency" cores that won't be used for any DSP at all.

Interesting, but to give some credit here if the problem is related to the machines themselves, I've only had very minor issues using two monitors with the Macbook Air here and pluggable which uses USB3 and CPU for the second monitor, (when you reach the limits of your CPU, windows make a zipper crackle noise when moved on the Pluggable monitor), Plus, DP does in fact use the Efficiency cores, you can watch it in Activity monitor, which is great because that means DP gets higher track counts than Logic etc.

All that said, it's worth noting that GPU might be an issue. I recall some weird hack in the past with third party monitors, Mac OS and DP, where it could be sluggish if you didn't do a set of things. Hard to think that GPU might be it when the Ultra Studio here has more than 9 times the GPU of the Air.

[chokobo99]

I wrote a friend about this issue and this is what he said:

After reading this post I have done pretty extensive testing with video resolutions from native to default, and many in-between. It didn't appear to have a noticeable difference in performance. It was still easy to max out the machine with just one synth plugin, and a reverb. Didn't matter what synth plugin, nor what reverb, and didn't have to be played via MIDI, could be done with the synth plugin's keyboard using the mouse.

Tonight I was able to get it to max out with ZERO plugins simply by sending MIDI to an external device, and bringing that back into DP. It didn't happy with the same frequency, but it did still max out.

Motu suggested that connected devices could be the issue. I have tested with a variety of devices connected, and no devices connected. Doesn't seem to have an influence.

As was reported early on the computer isn't running out of CPU power. Employing Activity Monitor's CPU meters they hardly ever go past four cores, and usually stay within the bottom 1/3 of the meter. Even though DP will max out.

I'm not using anything intensive like Omnisphere, and not pulling samples off of a disk. Not even reading any tracks off of a disk. Just playing a plugin synth through an effect is enough to hit the limit, and as of today receiving single MIDI notes, sending them out, and monitoring an audio source is enough to max it out.

Logic and Reaper don't seem to be having these same sorts of issues on the same machine, with the same plugins.

It's very easy to understand why the OP was frustrated enough to return his new machine. DP on the new portables, and potentially M1 portables running Ventura, is, from what I've seen, not a viable system today.

[Michael Canavan]

It's very easy to understand why the OP was frustrated enough to return his new machine. DP on the new portables, and potentially M1 portables running Ventura, is, from what I've seen, not a viable system today.

M1 Air here, 16GB RAM, 1TB SSD, no issues at all with spiking plugins in DP. Far far better performance than any Intel macbook, period. Dozens of instances of Diva etc. Again, you're not detailing your setup, just making outrageous claims. I wish you would give specifics but until you do, I'm going to point out how little evidence you're willing to give.

[Michael Canavan]

I wrote a friend about this issue and this is what he said:

After reading this post I have done pretty extensive testing with video resolutions from native to default, and many in-between. It didn't appear to have a noticeable difference in performance. It was still easy to max out the machine with just one synth plugin, and a reverb. Didn't matter what synth plugin, nor what reverb, and didn't have to be played via MIDI, could be done with the synth plugin's keyboard using the mouse.

Tonight I was able to get it to max out with ZERO plugins simply by sending MIDI to an external device, and bringing that back into DP. It didn't happy with the same frequency, but it did still max out.

Motu suggested that connected devices could be the issue. I have tested with a variety of devices connected, and no devices connected. Doesn't seem to have an influence.

As was reported early on the computer isn't running out of CPU power. Employing Activity Monitor's CPU meters they hardly ever go past four cores, and usually stay within the bottom 1/3 of the meter. Even though DP will max out.

I'm not using anything intensive like Omnisphere, and not pulling samples off of a disk. Not even reading any tracks off of a disk. Just playing a plugin synth through an effect is enough to hit the limit, and as of today receiving single MIDI notes, sending them out, and monitoring an audio source is enough to max it out.

Logic and Reaper don't seem to be having these same sorts of issues on the same machine, with the same plugins.

It's very easy to understand why the OP was frustrated enough to return his new machine. DP on the new portables, and potentially M1 portables running Ventura, is, from what I've seen, not a viable system today.

OK so a bit of crow to eat here. M1 portables do not seem to be affected, but I bought an M2 Studio that has had some serious issues with various plugins spiking and crashing DP, essentially NI and Akai, but who do not support Sonoma yet. It's not every plugin, but there's a good amount of weird spiking specific to the M2, and DP11. In my case I'm on Sonoma, but the issue from what I've seen so far is more DP specific to the M2 rather than Ventura or Sonoma. I have too many DAWs and none of the others are experiencing this, it's an IMO DP, plugin, and m2 specific issue. Basically a DP issue. Maybe solved when NI officially support Sonoma but my guess is MOTU will need to come up with a bugfix update.

[mikehalloran]

I have a big DP/Komplete project that I was going to revisit this week but I’m in no rush. My deadline is in December. Perhaps NI will get its act together by then. I’m not putting any iMac Pro back in service for it, however.