Oversampling vs Underclocking

Studiowaves

Doesn't it make more sense to start out with a high sample rate like 192k and divide the clock for things that don't need a high sample rate. That way high frequencies won't suffer from being underclocked going in or out of the ensemble. It also seems a better option for sound generators that can be interpolating at the final output stage. That way you don't lose the data from interpolation because the sample rate is high. I guess in theory the daw should also run at 192k because nowadays there are other choices for sound storage other than .wav files that rely on bit rates for quality. So it seems ok to set the master sample rate at 192 or 384k and underclock the internal workings of an ensemble around 48k for low cpu. I don't know guys, it just seems a waste to oversample internally and do a nice job of interpolation only to have it ruined with a slow master sample rate like 44.1k or less. Also for convenience sake I wish Reaktor had more choices for the SR.C. like SR.C/2 or /3 so we would not have to use the clock divider. Perhaps even name them with what they are. Like 96k instead of sr/2 with 192k master clock or 48k when the master clock set to 96k. Does this make sense? or am I on the right track just the wrong train as usual. lol

Sunborn

https://community.native-instruments.com/discussion/34088/oversampling-vs-underclocking

I don't think that there is a "yes" or "no" answer on this… It depends on what is your goal on each scenario and how much extra CPU and Latency you are willing to "sacrifice" (and of course, if your devices can handle it)…

I have this link from 2022 in my bookmarks, you might find it helpful… :-)

https://www.linkedin.com/pulse/oversampling-do-we-really-need-matanya-lemmy-keller

Studiowaves

https://community.native-instruments.com/discussion/comment/168501#Comment_168501

Although oversampling within an ensemble sounds like a good idea, it might produce high frequencies that are actually closer to the Nyquist of the true sample rate which can produce aliasing. Higher sample rates also reduce latency and cause sound card buffer overruns. Yet sound cards have buffer settings that increase the latency along with the buffer size so the latency can end up the same. I guess I'll try it and see what happens. Win11 is pretty good about prioritizing interrupts so that's a big help. Disabling the ethernet port really cuts down on crackling too. I think I'll bump up a project rate from 48k to 96k and play around with it.