End of Cubase? Performance Meter Average picking to 99% with very few Fx plugins

[flugel45]

Did I miss where OP ran Latencymon? (The fact that he's still groping in the dark says no).

This simple step will identify the task(s) that's using most resources. Honestly, this should have been troubleshoot step #1.

[dionenoid]

One thing worth mentioning : Don't use Windows to update your graphics card drivers. Manually download the latest drivers from the graphics card manufacturer, and install these.

I did that way

Than i misunderstand you. You said you were going to uninstall everything and use only those from Windows.

Anyway, despite all of us wanting to helpful, i think it's best to contact Steinberg and wait for a reply (yes, this might take long) or contact one or more of the devs from the plug-ins that use these high cpu amounts. Because what you describe is simply not normal behaviour.

[auron]

Another thing i noticed is that Reaper can handle Instrument-Plugins with long release settings much better than Cubase.

This is plugin-dependent and has nothing to do with DAW's. What likely happened is that you used a plugin with like, 64 or more polyphony with long releases versus a plugin with with 8~ voices with long releases. The voices get reassigned once they reach their maximum polyphony.

You sure? I used Halion Sonic 3 with the same patch in Cubase and Reaper (both newest versions..) but reaper is the clear winner. By the way its the same with other demanding Plugin Instruments, sometimes it is Avenger or Falcon (Granular Patches f.e.)

[jancivil]

Another thing i noticed is that Reaper can handle Instrument-Plugins with long release settings much better than Cubase.

entirely off-topic thread hijacking

[Functional]

Did I miss where OP ran Latencymon? (The fact that he's still groping in the dark says no).

This simple step will identify the task(s) that's using most resources. Honestly, this should have been troubleshoot step #1.

Because it's an useless tool for this. OP isn't confused what running task is causing issues; it's Cubase.

[flugel45]

Did I miss where OP ran Latencymon? (The fact that he's still groping in the dark says no).

This simple step will identify the task(s) that's using most resources. Honestly, this should have been troubleshoot step #1.

Because it's an useless tool for this. OP isn't confused what running task is causing issues; it's Cubase.

If that were 100% true, then the graphics drivers/software wouldn't have been suggested as a possible culprit. Bottom line, it's foolish to rule out something when you don't have a clue what the basis of the conflict is: Is this a resources conflict, or just a case of a project using up too much of his available resources? If it's the former, then it makes no sense NOT to run the utility. It hurts NOTHING to try it.

[Functional]

If that were 100% true, then the graphics drivers/software wouldn't have been suggested as a possible culprit. Bottom line, it's foolish to rule out something when you don't have a clue what the basis of the conflict is: Is this a resources conflict, or just a case of a project using up too much of his available resources? If it's the former, then it makes no sense NOT to run the utility. It hurts NOTHING to try it.

Maybe you have missed what OP has written in the very first post, so let me quote:

One project has about 27 plugins and it takes almost 90% of the “Performance Meter Average” with 17% real CPU.

So Cubase performance meter is reporting 90% figure and, out of total available CPU resources, only 17% is being utilized in total. In laymans language this is very simple: either something prevents Cubase from allocating properly in multicore setup the various plugins or alternatively something is, very specifically, affecting Cubase in a way that makes the software itself perform like shit, all the while Windows itself sees no errors. And what's more, it's mentioned that Reaper does not give out these issues.

OP is clearly tired of doing thousand diagnostic steps and I will guarantee that LatencyMon, based on the very first post in this thread, is not going to give any insightful results. This problem is related to Cubase, either caused by Cubase itself or through some interaction with Cubase.

So, all said and done, I'd consider running LatencyMon as useful as running HijackThis; the odds of it giving any useful input on this are not zero, just like taking a break, getting some coffee and giving some thought about this issue, might as well result into a useful thought.

[auron]

Another thing i noticed is that Reaper can handle Instrument-Plugins with long release settings much better than Cubase.

entirely off-topic thread hijacking

Nope, that is a comparison . I could compare also with Reason 11. Now Steinberg says, every time you put a Mastering Plugin on the master bus, the CPU can't do multicore properly. Thats a lie!

Why does it work with a mastering plugin in Reaper and Reason 11 with 80-90% multicore utilization?

Right answer: because Cubase is not optimized for multicore. They won't admit that , they keep the secret that they have some sort of "average" algorithm , just works good more or less on some system but no optimization. You can't even do anymore in Cubase settings than hitting multicore or boost and thats it. No sandboxing , no thread priority , no different settings for aggressive thread utilization.

There are a lot of people saying hey my threadripper doesn't work with 80-90% CPU processing.

Maybe they have it Intel-optimized only, but Steinis won't admit it either.

[jancivil]

I really don't know what windows shows with that regarding CPU but 17% seems highly weird to me, CF., single core and it not being particularly stressed.
Here, during playback of a project in Cubase, Activity Monitor is going to read, eg., "524%", which indicates a sixth core has been called.

I don't frequently consult Cubase peformance meter, but it's never shown 90%, 50% is orange alert to me. I'm surprised it wasn't (not entirely confident of the methodology there tho) the graphics driver SNAFU.

[mduke123]

Just to add that Steinberg engineer called me and he was trying to fix the issue for almost 3 hours with no sucess. The result was that my moderboard is too simple and cannot be handle to enhance a little bit the performance, just to be at the same height of Studio One or FL Studio. With all this I am telling you Cubase as the worst performance ever, bellow any major Daw. I made my tests with FL Studio, Bitwig, Studio One and of course Reaper.

[Freaqpeak]

I have more than 60 instances of U-he Hive instruments in a project + vst fx, cubase shows 85-90% CPU usage but I have no problems at all. No crackles etc. Cubase CPU meter is "broken" and does not show real CPU usage. Use Windows
256 sample latency. Asio guard on, default settings.
AMD Ryzen 3950x. For Windows Scheduling I use 1smus. Steinberg scheduling works good as well though.

[WasteLand]

I have more than 60 instances of U-he Hive instruments in a project + vst fx, cubase shows 85-90% CPU usage but I have no problems at all. No crackles etc. Cubase CPU meter is "broken" and does not show real CPU usage. Use Windows
256 sample latency. Asio guard on, default settings.
AMD Ryzen 3950x. For Windows Scheduling I use 1smus. Steinberg scheduling works good as well though.

there seem to be a lot of new messages, about the performance of Cubase.

but this last one i can correct; in a DAW it is never a CPU meter, it is the how much the buffer is filled. that is why when it is on half, you can add more instance of several VSTi's for example, and the meter does not go up. it is nót a CPU meter.
it only shows the room you have left (and taken) of the buffer.

it is a short of an avarage (well it is, there is also peak... with asio guard on, and no record arm; mostly you see the peak go down...), a VST can take a big chunk of CPU, an a smaller chunk of buffer, but if CPU core is overloaded, the buffer will overload too... of course, but even when a CPU core isn't overloaded, and other cores are very busy, but not overloaded, the buffer can...

and a general remark; but i have to read the latest posts, and perhaps the others;

Cubase is not known for it is performance, and can be very system (and user) dependent.

in my case it performs quite excellent. even next to Reaper (6.64), for instance.

but short story shorter: it is nót a CPU meter, no DAW uses it, except Reaper shows the load overall, and per core, per instance of a VST. and then you see, heavy plugins, tax the CPU, even in Reaper. sometimes it does not go down, or in a lot of case, my arsenal of plugins.
but with no tracks armed (record armed) the Buffer of Reaper, will show, a load of, at 256 samples/48Khz, 5ms in my case i believe or 6ms, output latency, mostly under 1ms.

but Reaper uses a very big buffer, pre-rendering buffer.

and i see -i didn't do a scientific test- mostly the same cpu usage under all the DAW's i have, and all cores are use (12/24).

[mduke123]

I use average plugins mostly from Plugin Alliance for Mixing. Every time I have tried to mix anything in Cubase it ran out of performance with clicks and pops. The same when I tried to produce electronic trance. No way to freeze busses. For me I have finisehd with Cubase forever. I would prefer to learn Protools before I ever consider using a brass oldy like this.

[zvenx]

I generally don't have cpu issues with cubendo except with plugin alliance stuff which I have stopped using.
Their copy protection has issues with low buffers in cubendo. No i can't say how i know this. But yeah I would avoid pa stuff if you are running at low latencies. ( don't know if you are).
Instead i use fabfilter, Softube, uad, fuse audio et al to mix, and have never had that issue whilst mixing again.

And yes I reported it several times to PA over the last probably five or more years.
rsp

[wuworld]

I generally don't have cpu issues with cubendo except with plugin alliance stuff which I have stopped using.
Their copy protection has issues with low buffers in cubendo. No i can't say how i know this. But yeah I would avoid pa stuff if you are running at low latencies. ( don't know if you are).
Instead i use fabfilter, Softube, uad, fuse audio et al to mix, and have never had that issue whilst mixing again.

And yes I reported it several times to PA over the last probably five or more years.
rsp

Have the issues improved for you with PA's stuff? I have their stuff. I bought CP12 Pro recently. If the PA plugins perform bad, is it PA's responsibility to fix it?