Post
by Calagan » Wed Feb 14, 2024 1:29 pm
Use whatever works. On stage, I would recommend the lightest option and (personally) would never go on stage with Live running at 96kHz : if you want stability (and stability is the most important thing on stage) you need a lot of processing headroom and 96kHz will not give you that...
I used often 44kHz and 48kHz and it's perfectly fine... Especially that on stage, it's not a "pro studio quality" experience, so even if higher sample rates sound theoretically better (and that is not completely the case, see below), nobody will hear...
It may happen that some audio interfaces sound better at some sample rates (I noticed, for exemple, that my Focusrite Clarett was sounding better at 48kHz than any other sample rate, be it 44khz or 88kHz), but it's very subtle and I didn't noticed that with other interfaces (like Motu ones).
By the way, as you mention Dan Worrall, he never said that using high sample rate was better - quite the contrary.
He is a fervent advocate of plugin oversampling instead of using very high samples rates for the whole session : because when running a full session at 96kHz or higher, what happens is the aliasing is just happening above Nyquist where it's impossible to hear. But it's still processed by plugins after that, it polluates the whole session and accumulate, until you start to hear it in the output with intermodulation distortion.
It's counter-intuitive, but higher sample rate can lead to more distortion...
And obviously, you burn a lot of CPU processing for many plugins that don't need high sample rates (like EQs, chorus, reverbs, delays, etc. etc.)
This is the reason why some clever plugins dev did create tools like Ultrasonic, that filter the ultrasonic part after each non linear processor.
Finally, there is Operator : you can't oversample it into Live (or with Metaplugin for exemple), and maybe it sounds better at high sample rate (it actually should sound better, I never tried). This would be the only reason to use high sample rate. But if your Live session is using only audio samples, and if your Operator parts are previously recorded, it doesn't make sense IMHO to use more than 48kHz... Run Operator at 96kHz and export at 44 or 48 Khz.
(It's also true that higher sample rate sounds better for extreme pitchshifting or audio manipulation. But if the processing is done previously, and you only use the exports, again no need to use 96kHz or higher).
