DAW Stress Test: Logic/ProTools/StudioOne/Cubase

[Dewdman42]

another aspect that would be interesting to include a test comparison would be to actually measure the latency on both systems to make sure that comparative results are based on getting the same latency. This is because DAW's can sometimes add extra safety buffers...which increases the work they can do, but also the latency. A true comparison will be like for like. Everything as much the same as possible. Same buffer size (or rather, same settings that achieve the same RTL), same plugins, same tracks, etc.. with same resulting output, which would be clean audio no drop outs, same latency, same sample rate, etc..

[machinesworking]

Open VEP in DP10, open the Effects Performance Window and look at the results. I have a few VSTis and VEP MAS plug ins open right now, no tracks record enabled. The regular VSTi's are reading as PG or PreGen, the VEP tacks are reading as RT or Real Time.

you are conflating three seperate concepts into two. Just because something cannot be pre-gen'd does not mean its handled like an external instrument. Just because DP calls it real time on the window for musicians does absolutely NOT mean that its running in true real time like an external instrument. Sorry but you are just plain wrong about VEP here.

Right, it's just an efficient MAS plug in running in Real Time.

No, nothing runs in actual realtime, that is just how MOTU labels it for musicians to distinguish between Pregenned and not-pre-genned, but not-pre-genned is still not real time in the same sense that an external instrument is.

You're mincing words here. External instruments even pick of milliseconds of delay, everything that passes through a computer does. Like I've mentioned a couple times, there is another method besides pre rendering that's being used pretty universally by all DAW manufacturers, and that's to ad in extra buffering or milliseconds of delay to tracks that are not armed for recording. VEP is probably capable of benefiting from this other method, which is essentially the same method that DAWs use for plug in delay compensation. This method is better used by plug ins that are not multis, routed to aux tracks, armed for recording, external instruments etc. etc. They are not the same method, a plug can benefit from not being armed and getting a larger buffer size than your settings in Preferences, and still not be able to be pre rendered. We can argue about whether or not a heavily buffered plug in is in Real Time or not, but apparently MOTU believe it is running in "real time", and that, is why I would call it real time.

Yes, because VEP runs in real time

No absolutely not. Real time according to the DP Gui, but its not actually real time. You are misunderstanding core concepts here. its not pre-genned, and that only applies to DP.

Some of this, is coming from the fact that all DAW manufacturers are using their own terminology and explaining it in their own words. You're arguing circles here, and that's not entirely your fault.
In the end though I tend to side with reputable manufacturers of software and hardware we use over people claiming things on the internet. You're arguing semantics, I suspect because underneath all the dismissiveness you know I'm right, stress tests not done to failure are not accurate and serve no real world purpose.

It's pretty dammed simple really, plug ins eat CPU, if Reaper can handle more plug ins at the absolute limit of your computer set up, then it is a more efficient DAW, period.

Now with the fact that there are multiple CPU saving schemes used by DAWs it's I think important to force "live" (or real time or what ever you want to call record arming a track) as another test of a DAW, because in that case, from my limited tests for instance, you would be better off using DP or Logic. With typical compositional strategies of one track armed at a time etc. Reaper is going to absolutely kill the other DAWs, in every test I've done. This plays out in real world performance as well.

Again, doing a test of CPU where Logic comes in at say 70% and Reaper 78, then saying Logic is more efficient is misleading. Reaper can handle more plug ins period. Allyou need to do to prove this is actually stress your system. Also again, it's not my favorite DAW, I have no chauvinism here, it's just a fact that Reaper out performs Logic in stress tests, and that's that.

[Dewdman42]

You're mincing words here. External instruments even pick of milliseconds of delay, everything that passes through a computer does.

No I'm not mincing words. When you use an external instrument, the actual midi has to be sent to something else, which returns the audio back...the thing making the audio is not operating on the actual audio buffer of the DAW..it is in fact moved completely into the real time world as midi events leaving the DAW and then coming back to it as audio into the playing system. That is completely real time.

With buffers, the DAW can work as early as it feels like working on the buffer...could be pre-rendered, might not be, but its ahead of time, as the DAW feels like doing. External instruments do not work that way. VEP does. VEP is NOT an external instrument.

Its no different then any other Instrument plugin that has a lot of latency. When a midi region feeds an instrument with a lot of latency, the midi is sent from the track to the plugin early so that the resulting audio will be right on time. Live midi can't do that of course. So Live midi will be very latent.

What VEP does is it attempts to respect the latency you have your system configured for when it detects that it is receiving live midi. When is that? that is when you have the track record enabled. If the track is not record enabled, then VEP detects the midi can be received early through PluginDelayCompensation, and so VEP cranks up the latency of itself...which the DAW then handles through PDC to get the midi to it early. This is not real time. This is not how an external instrument works at all. VEP works just like any other instrument plugin.

As to why DP is able to pre-render certain instruments and not others, you will have to ask MOTU about that, but it does not mean that all instruments that can't be pre-rendered are the same as external instruments. They are not.

As I said before you are conflating three things into 2.

Like I've mentioned a couple times, there is another method besides pre rendering that's being used pretty universally by all DAW manufacturers, and that's to ad in extra buffering or milliseconds of delay to tracks that are not armed for recording. VEP is probably capable of benefiting from this other method, which is essentially the same method that DAWs use for plug in delay compensation.

PDC only ramps up the latency on AUX channels. Instrument and Audio channels generally have the data from the region fed to it early in order to compensate for the latency. In essence, PDC is acting more like Pre-rendering. The buffer size is not being increased...the pre-feeding of the source data is being sent in early.

This method is better used by plug ins that are not multis, routed to aux tracks, armed for recording, external instruments etc. etc.

AUX channels generally can't be fed early, so they will be as latent as they are. What happens is that the rest of the project will be made more latent to compensate. That is what you consider to be PDC, but that is only how latency on AUX channels is handled, not how latency on audio and instrument channels are handled.

We can argue about whether or not a heavily buffered plug in is in Real Time or not, but apparently MOTU believe it is running in "real time", and that, is why I would call it real time.

Nothing is ever in real time. There is always a buffer and the computer operates on that buffer not in real time. When MOTU says "realtime" they mean its not pre-rendered prior to playback.

Yes, because VEP runs in real time

No.

It's pretty dammed simple really, plug ins eat CPU, if Reaper can handle more plug ins at the absolute limit of your computer set up, then it is a more efficient DAW, period.

Not true. You are making a generalized statement based on an isolated situation. I actually don't believe your data for a variety of reasons including you appear to be very biased and argumentative in general. But what you can say, is that your tests show that Reaper can handle more instances of Diva then LogicPro on the same machine. Your isolated test does not mean that categorically Reaper is more efficent then LogicPro, in fact numerous tests from numerous people have shown that in more normal real world situations, its quite the opposite.

Now with the fact that there are multiple CPU saving schemes used by DAWs it's I think important to force "live" (or real time or what ever you want to call record arming a track) as another test of a DAW, because in that case, from my limited tests for instance, you would be better off using DP or Logic. With typical compositional strategies of one track armed at a time etc. Reaper is going to absolutely kill the other DAWs, in every test I've done. This plays out in real world performance as well.

Show us the data. You have not.

[machinesworking]

Not true. You are making a generalized statement based on an isolated situation. I actually don't believe your data for a variety of reasons including you appear to be very biased and argumentative in general. But what you can say, is that your tests show that Reaper can handle more instances of Diva then LogicPro on the same machine. Your isolated test does not mean that categorically Reaper is more efficent then LogicPro, in fact numerous tests from numerous people have shown that in more normal real world situations, its quite the opposite.
 

I'm boiling this down because this has gotten completely out of hand, and your statement right there, is the basic argument, the rest is fluff.

Again, this isn't an isolated situation. You on your computer can do the exact same tests, or even just similar, and get the same percentage differences. I 100% encourage you to do tests to failure. That's been my point all along, tests to failure eliminate the computer, the plug in and everything else besides the overhead that the DAW has.

There are essentially three tiers to DAWs at this point.

- Bitwig, Live, MPC2, and probably Reason coming in at the rock bottom. They use the absolute most CPU.

- Logic, DP etc. coming in the middle, skunking the previous.

-Reaper, always pulling out a few more plug ins than the others.

There's nothing isolated about this situation, This has been the conclusion on multiple computer setups forever. A stress test not to failure does nothing, it only shows what the computer and measuring devices can read, not what the DAW can do.

For instance DP uses all the CPU before it crackles, as read in Activity Monitor, Logic a bit less, and Reaper literally only reads about 65% CPU use before it can not longer load plug ins. This is even with Reaper loading 25% more plug ins than either Logic or DP. This is why tests not to failure are worthless, DAWs grab CPU in a tragically non linear fashion, but the results when pushed are always the same. Like I said, DP is by far my favorite, but I do no one including myself any favors by ignoring the fact that Reaper is as lean as they come.

[Iva]

this test is crap. cpu usage does nothing to DAW performance. Make this test instead - insert Roland system-8 vsti with default patch and play two notes sequence. On my 4-core cpu Logic Pro can handle only one track, studio one - 3 tracks, Cubase - 4 !!! (MBP late 2016- 2,6 Ghz)
I also made test on the same system with waves CLA Effects - Logic - 480 instances, Cubase - 562. Reaper sucks too comparing to Cubase on Mac. So I don't know how you tested, but your test isn't showing the reality

[Dewdman42]

Iva,

In order to do a fair comparison between Logic, S1 and Cubase, you need to make sure that none of them have record-enabled tracks with the Roland plugin on them. In the case of LogicPro, if you have one of the Roland Tracks "selected", then its record-enabled. So create another empty track and select that one before playback. You should get more efficiency and a fair comparison with S1 and Cubase which were probably not using record-enabled tracks, so thus it was not a fair comparison.

Your results don't make any sense to me between Logic and Cubase, my results and the results of some others were directly opposite of that. But. which version of Cubase by the way? Also did you have Cubase's ASIO guard turned on? If so, then what is the reported latency in Cubase? Bump up the buffer in LogicPro and S1 until they have the same latency you are getting with Cubase and then do the test again.

[Iva]

I am on cubase pro 10.0.40
record on tracks turned off, Asio buffer is the same. Asio guard turned on as is logic's dropout protection. other daws than cubase have clearly worse performance. just test it yourself and face it. do a real ASIO stress test, not just simple cpu usage.

p.s.: its very sad to see all this bullshit about Logic Pro x and reaper as the uber cpu daws. its simply not true. I even start to think that posts like this on pro forums are just a part of apple's marketing strategy.

[CrystalWizard]

Totally agree with machinesworking. Coming from an (electronic as well as sound) engineer background a stress test means throw everything at it that it will take ‘til it tips over (too intense too stay under pressure) or some such term. Stress in this case means fault. When it stops working. Sometimes at the bench that meant when you let the smoke out of the chips.
My first day with ableton live i kept loading samples on new channels until it sank in the mire of digital hash (chill out everyone, it’s not that strong) then i fitted it up with as many channels of a vsti until it burped up digits everywhere. Then i armed as many tracks to record as i could (8) and kept adding loops on the rest. And variations of above, always ‘til the snap, crackle, and pop. Meters? Meh. Stress Test.

[Dewdman42]

I am on cubase pro 10.0.40
record on tracks turned off, Asio buffer is the same. Asio guard turned on as is logic's dropout protection. other daws than cubase have clearly worse performance. just test it yourself and face it. do a real ASIO stress test, not just simple cpu usage.

p.s.: its very sad to see all this bullshit about Logic Pro x and reaper as the uber cpu daws. its simply not true. I even start to think that posts like this on pro forums are just a part of apple's marketing strategy.

I did test it myself and cubase performed worst of any of them on my system. Vsl engineer also confirmed to me in dialog that he felt cubase performance sucks FWIW. If Asio guard is on then it creates more latency, make the latency equal on the two daws prior to testing and make sure LPX does not have tracks re odd enabled as i said.

It’s interesting that so many people report opposite results but I think it comes down to configuration and personal preferences. The animosity being expressed by some is also interesting.

[Dewdman42]

I am on cubase pro 10.0.40
record on tracks turned off, Asio buffer is the same. Asio guard turned on as is logic's dropout protection. other daws than cubase have clearly worse performance. just test it yourself and face it. do a real ASIO stress test, not just simple cpu usage.

p.s.: its very sad to see all this bullshit about Logic Pro x and reaper as the uber cpu daws. its simply not true. I even start to think that posts like this on pro forums are just a part of apple's marketing strategy.

I did test it myself and cubase performed worst of any of them on my system. Vsl engineer also confirmed to me in dialog that he felt cubase performance sucks FWIW. If Asio guard is on then it creates more latency, make the latency equal on the two daws prior to testing and make sure LPX does not have tracks record enabled as i said.

It’s interesting that so many people report opposite results but I think it comes down to configuration and personal preferences. The animosity being expressed by some is also interesting.

[Iva]

ASIO guard is supposed to be turned on. You mustn't turn it off. It's present in the most modern daws like Logic, Studio One, Reaper, but called differently like dropout protection in logic. Cubase has Asio latency compensation for midi messages during recording. also if u need super realtime performance, there is a special button in Cubase, which temporarily turns off Asia guard and heavy latency plugins for recording. I can't imagine scenario where u would need less than 256 buffer size - 20 ms latency in studio environment. for recording vocals and other instruments u may also use direct monitoring which almost any modern sound card has.

[Dewdman42]

I didn’t say to turn it off. Turn up the buffer size in logic until the latency is equal

[Iva]

its equal, I must have wrote that already. I was a logic user prior to cubase's Asia guard introduction. since then, I switched and never looked back. u can find on this forum my old thread about cubase 8 I think, where logic outperformed cubase almost twice in cpu load. but since then everything changed and logic is somewhere behind now.

here is the old thread:
viewtopic.php?f=7&t=420359&p=5874138#p5874138

I am glad its all in the past. Steinberg did all I asked for that time.

[Dewdman42]

So you’re saying you have buffer size set to a larger value in Logic Pro compared to cubase? How can you explain the results you are reporting now in 2019 that are directly opposite to the results of myself and others?

[Iva]

looks like we don't understand each other. I am sorry. I don't see point in future discussion.