Hi : )
I use Cubase Pro 9.5 and the steinberg UR22 soundcard.
I would like to run Cubase on a very low audio rate (24 or 12 Khz) to save cpu recources.
I know that I can change the audio rate at: Studio > Studio Setup>VST-Audiosystem.
But there I can’t change it below 44.1 Khz.
Is there any possibility to run Cubase at an audio rate below 44.1 Khz?
Thanks in advance
Matthias
I think it depends on the audio interface you’re using, not really Cubase.
How much resources will you save?
I just searched the web for a way to get the Steinberg UR22 audio interface to sample rates lower than 44.1 Khz but it seems not possibble.
I think it could have quite an effect on cpu recources. Just from changing the sample rate from 48 Khz to 44.1 Khz I noticed a drop of about 5-10 percent cpu load.
I could imagine reducing the sample rate to something like 24 Khz (or even lower) while writing a song could save one third to one half of cpu load, which would be huge.
If you find an interface that does allow lower-than-44.1k SR, You should observe how the buffer size and latency time changes when you do that. The savings might be illusory, if you expect the same latency on your VST instruments.
I guess if you only use outboard midi gear it doesn’t make a difference. Is that your case?
Are you getting close to maxing out your computer’s resources? If not, I wouldn’t put much effort into conserving resources. What a lot of folks don’t appreciate is a computer always consumes 100% of its CPU. It uses the capacity by either doing some work (what Task Manager shows you) or by throwing the remaining capacity into the trash. If you make a bunch of changes to drop you CPU use from 50% to 40%, that just means you are now throwing away 60% of your capacity. If the changes you made reduced the quality of your audio, that’s a pretty bad deal - fwiw, I’m an ex (escaped) IT guy.
@steve:
Thank you!
Latency is not an issue for me, since I produce mainly in the box without live midi performance.
@raino:
Thanks for the help!
Yes I’m maxing out my computers recources, at least from what I can tell.
In a recent case I had five instances of the Generate-VST running in real time.
At 48 Khz sample rate the cpu-load-meter in Cubase got to the red and the typical overload crackling noises appeared.
As I dropped the Samplerate to 44.1 Khz the cpu-load-meter dropped and no crackling appeared.
I might be mistaken, but to me this indicates a connection between samplerate and cpu load.
Ofcourse dropping the samplerate will create aliasing artifacts, but having a couple more Synths I could tweak in realtime would make this worth for me.
Also for everything above 16Khz the aliasing is barely noticable.
First, that isn’t a CPU load meter, it’s the System Performance Meter which looks at the performance of the entire audio system & measures a variety of things of which CPU is only one. So while a CPU problem would show up here, other things can also trigger this meter. I assume if you look in Task Manager you see the CPU maxing out.
If you think about it it makes sense reducing the sample rate saves CPU. Every point along an audio waveform needs to have its value calculated. When you lower the sample rate there are fewer points along the waveform that need to be calculated.
What I’m wondering is if something else in your configuration is causing your CPU to start maxing out. And then when you lower the sample rate you regain just enough capacity to not be maxing out. Can you give us some details about your computer and also what a typical Project looks like - number/type of Tracks, main VSTi’s & plug-ins, etc.
First, that isn’t a CPU load meter, it’s the System Performance Meter which looks at the performance of the entire audio system & measures a variety of things of which CPU is only one. So while a CPU problem would show up here, other things can also trigger this meter. I assume if you look in Task Manager you see the CPU maxing out.
Ah thanks for the info, didn’t know that, interesting!
If you think about it it makes sense reducing the sample rate saves CPU. Every point along an audio waveform needs to have its value calculated. When you lower the sample rate there are fewer points along the waveform that need to be calculated.
That was my theory too. In practice I would work barely at where the Nyquist level becomes audible for the ears eg. 16 Khz, save alot of cpu (half the points calculatet in the waveform should mean half the cpu load?) and when bouncing the project I still could go to 192 Khz.
What I’m wondering is if something else in your configuration is causing your CPU to start maxing out. And then when you lower the sample rate you regain just enough capacity to not be maxing out. Can you give us some details about your computer and also what a typical Project looks like - number/type of Tracks, main VSTi’s & plug-ins, etc.
That would be good to know too, and I appreciate the help, but besides of me believing my PC runs quite decent, I’m more interested at the moment in finding the Answer to the sample rate problem.
I was really surprised that I could find NOTHING on the Internet about it, thats why i started this post in the first place.
It seems to me like a great solution for saving computing power. Just Imagine it: twice as much of effects, instruments, busses… anything!
It really is a miracle to me why this isn’t a common technique.
have you tried making the buffer larger?
Yes I tried, but it only works to a certain extend.