Solved Studio One and Apple Silicon

[Crossinger]

Hi everyone,

I‘m about to replace my old Mac with a new one, and I stumbled across this Video (released today):

Uh, is it really that bad? I mean, I don't see myself reaching the performance limits of my computer in the foreseeable future. But the issues shown in the video are significant.

Any real world experience on this?

 

[chakko]

I haven't watched the whole video, but, it seems like he's doing proper tests, so, I can only say that benchmarks don't lie. I read many times that performance of Studio One on Mac isn't great.

You said the most relevant thing (for you) yourself though: "I don't see myself reaching the performance limits of my computer in the foreseeable future."

 

[Lukas]

I haven't watched the whole video, but, it seems like he's doing proper tests, so, I can only say that benchmarks don't lie. I read many times that performance of Studio One on Mac isn't great.

I appreciate his work and his tests, but they give the impression that you are comparing the performance of different DAWs directly with each other and from that point of view the video is massively flawed.

In of his previous efficiency core/benchmark/DAW comparison videos, he compared DAWs with different settings under completely different conditions, so these comparisions were absolute nonsense. These are complex topics and unfortunately, most people who compare and discuss CPU and multi-core topics have no idea what they're talking about.

 

[Crossinger]

He compared DAWs with different settings under completely different conditions,

Ok, conditions can be altered, but what options do I have for changing settings in a DAW (generally speaking) so that it affects the CPU core utilization? I‘m not aware of such settings.

 

[Lukas]

Ok, conditions can be altered,

Only to a limited extent because different DAWs use different techniques. For example...

"Reaper and Pro Tools have massive buffering: Reaper AFX is not a buffer, but its prerendering works as if it has a few seconds of buffer. Basically, it is offline prerendering."

 

[Jemusic]

Be careful. Do the research carefully and find out what the ratio of performance and efficiency cores actually is for the model you are interested. Its does not matter how any tests were done either. Studio One will not use the efficiency cores in your Mac. Only Cubase and Reaper will use all the cores. eg both performance and efficiency. I have the M2 Max Mac Studio which has 12 cores but 8 are only useable by Studio One. Its still fast and amazing but its a pity not all of them can be used. I would get a performance jump of another 4 cores which would be great if I switched to another DAW. But even on 8 performance cores the results are quite amazing. And will leave most Windows machines in the dust.

Reaper actually out performs Cubase and the reason is the amount of bloat Cubase has slows it down a little. Reaper is the most efficiently coded DAW out there. I am hoping Studio One will eventually take advantage of the efficiency cores. Because if Reaper and Cubase can then it can be done, right?

The new Mac Mini M4 (standard) for example has 10 cores but 6 of them are efficiency which is not great for Studio One. But the Pro model has 14 cores and 10 of them are performance and only 4 are efficiency so that would be far better option. My advice is to wait until the new Mac Studio M4 is released next year.

 

[TDF1981]

Guys, you do realize in that comparison how Logic Pro 11 does not use the efficiency cores, right? Don’t you think Apple would know better and use them if they were intended to be used for audio? The efficiency cores are not built for real time audio and that’s probably the reason why Apple will not touch them and Studio One isn’t. This might change in the future but that’s how it‘s now.
I have been using Studio One on Apple computers since v4 and I have never had any major problems.

 

[Crossinger]

The new Mac Mini M4 (standard) for example has 10 cores but 6 of them are efficiency which is not great for Studio One. But the Pro model has 14 cores and 10 of them are performance and only 4 are efficiency so that would be far better option. My advice is to wait until the new Mac Studio M4 is released next year.

Thanks, Jemusic. I already did my homework First of all: no desktop! For several (not music related) reasons I need to be mobile. And my CPU of choice is the 12 core M4 Pro - best tradeoff between power and cost.

 

[Lukas]

The efficiency cores are not built for real time audio and that’s probably the reason why Apple will not touch them and Studio One isn’t.

Yes. Using efficiency cores in large buffer scenarios could be an option. But there is a high risk of dropouts when used in a lower buffering scenario. This is the obvious reason why many DAWs don't use efficiency cores at all.

 

[Jemusic]

Except Cubase and Reaper can use the efficiency cores without issue and they get the performance boost as a result. Especially Reaper. Not sure efficiency cores are not good for real time. The app might have something to do with it. The video editors are getting the full use of all cores and that is about as real time as you can get. They are moving a lot more data around in real time than audio.

 

[chakko]

With the introduction of Intel's 12th gen CPUs, Steinberg described workarounds for performance issues with CPUs with hybrid acrhitecture (i.e. the ones with efficiency cores), but stated that since Cubase 13, such CPUs are officially supported.

I would be surprised if the same wouldn't hold true for any other major DAW as well. They surely all found solutions to work with these kind of CPUs now.

 

[Jemusic]

With the introduction of Intel's 12th gen CPUs, Steinberg described workarounds for performance issues with CPUs with hybrid acrhitecture (i.e. the ones with efficiency cores), but stated that since Cubase 13, such CPUs are officially supported.

I would be surprised if the same wouldn't hold true for any other major DAW as well. They surely all found solutions to work with these kind of CPUs now.

I sure hope so and that Studio One gets on board with this. There might be more taking up the Mac option as time goes on too.

 

[Jemusic]

Thanks, Jemusic. I already did my homework First of all: no desktop! For several (not music related) reasons I need to be mobile. And my CPU of choice is the 12 core M4 Pro - best tradeoff between power and cost.

Excellent choice. The M4 Pro versions have 10 performance cores and being the M4 chip will be very fast indeed!

 

[ZoeMusic]

I think it is a good thing that the DAW doesn't have priority over the efficiency cores because other programs and background resources can use them and provide better stability and overall performance. I do Post Work with the external Dolby Atmos Renderer which will use the efficiency cores which gives me more CPU performance over all leaving the Performance Cores for my DAW. This has been my experience. See Attachments. Also the computer knows how to allocate all of this even if the DAW does or does not take full advantage of it.

 

[Lukas]

Uh, is it really that bad? I mean, I don't see myself reaching the performance limits of my computer in the foreseeable future. But the issues shown in the video are significant.

Just forget about this video.

Let me (again) quote Leon from the S1 FB group who is one of the true experts in this field:

I cannot believe this guy is doing the same bullshit comparison like last year. He uses different buffer settings for various DAWs. Reaper, Logic and Cubase are set to huge buffers, while S1 is set to "high" instead of max. Even at max, the other DAWs have larger buffers.
I told him last year he is not comparing apples with apples, yet he not only chooses to ignorebthisbfact, but also boldly claims again to set Reaper to AFX and aggressive mode with the comment "if a DAW offers such functions you might as well use them".

 

[uOpt]

Not using efficiency cores is a sign of not understanding how CPU scheduling works in the OS, and what real-time processing really means. That's why DAWs stopped doing this nonsense. I am surprised that Logic of all people is still doing this. Probably not much longer.

 

[Lukas]

Not using efficiency cores is a sign of not understanding how CPU scheduling works in the OS, and what real-time processing really means.

Can you elaborate on what you think real-time processing really means?

 

[uOpt]

Can you elaborate on what you think real-time processing really means?

Real-time doesn't mean "fast", or "maximum throughput". In fact, strict real-time goes against maximum throughput.

It means that there is predictable timing, a guaranteed point in time by which a certain piece of work is done. The actual amount of work is strictly limited.

Real-time systems in the narrow sense (stricter than our music systems) require special OSes and special programming techniques that are incompatible with our (relatively) complex algorithms. For example, you would be unable to use dynamic memory allocation. But music system don't even aim for that. The closest we get is the PREEMPT_RT preemption mode in Linux-6.12 which gives you a real-time OS kernel (at the expense of throughput and general speed), but even then DAWs are very far from real-time programs.

Anyway, E-cores are actually more suitable for real-time programming since they don't clock as high and hence it is more predictable what amount of work they get done in what amount of time.

When it comes to DAWs on heterogenous cores CPUs it makes no sense to exclude the slower cores since the processes will migrate between cores anyway, and should be allowed to migrate to and from the E-cores as well. I know that not all DAWs let the OS scheduler do this properly, but there is light at the end of the tunnel with more and more DAWs seeing the light and just let the OS and the hardware do their jobs without misguided attempts to help.

The big surprise, as mentioned, is the non-E-core policy of Apple's own DAW. I think that must be the result of some misguided corporate order to leave the E-cores free for background tasks. In any case, I expect this to fall soon. It makes no sense.

 

[Logical]

I haven't watched the video and I probably won't. One of his previous efficiency core/benchmark/DAW comparison videos was heavily flawed. He compared DAWs with different settings under completely different conditions, so these comparisions were absolute nonsense. These are complex topics and unfortunately, most people who compare and discuss CPU and multi-core topics have no idea what they're talking about.

If the buffer settings are not the same on the tests then how does this make any sense? I’m surprised that pro tools is scoring higher than studio one.

 

[Lukas]

If the buffer settings are not the same on the tests then how does this make any sense?

It doesn't. That's what nobody commenting under this video seems to realize. The videos shows which cores are used - that's fine. But the setup does not at all allow a fair comparison of actual performance between DAWs.

CPU/multi core/audio engines are complex topics (not my expertise at all) that I don't think users should even have to deal with. I actually find it bizarre how many people on social media are currently acting like they know better then DAW developers just because they can quote two sentences from the (as you say flawed) video.

I'm not saying that there's nothing to improve about Studio One's performance (there is!). But if you're going to compare, you should compare fairly, and not apples (REAPER AFX with a huge render-ahead value) with oranges (Studio One with Dropout Protection = High).