I would record 24/96 but when I load sonarx3 it causes my interface to click back and forth between 48kHz and 96kHz I have to close the file. And I don’t know what’s causing that. I’m using Windows 8.1 and Focusrite Pro 40 interface.
in Pro Tools no problem

sharke
As an aside, am I right in assuming that if your intention is to slow audio down (time stretch), the end result will sound better for audio recorded at 96kHz than audio recorded at 48kHz? In much the same way as film recorded at a higher frame rate looks smoother when slowed down.

Only if the audio you want to slow down contains frequencies > 24kHz before you slow it down. Unlike frames of film, the entire part of the waveform between the existing samples is already stored in the samples. 
 
EDIT: I should add that if you are using a less precise/lower quality algorithm for transposing then you can indeed get better results starting from a higher sample rate, all things being equal.
 
 
I don't recommend thinking about digital sampling in terms of film, photography, video, printers or anything like that because those sorts of analogies tend to be tenuous at best.

Anderton

sharke
As an aside, am I right in assuming that if your intention is to slow audio down (time stretch), the end result will sound better for audio recorded at 96kHz than audio recorded at 48kHz? In much the same way as film recorded at a higher frame rate looks smoother when slowed down.

That's a really interesting question for which I have no answer.
 

 
 
http://forum.cakewalk.com...sampling-m3121307.aspx
 

What he said ^

Here's an interesting article relating to the question I asked, with some audio examples:
 
http://www.musicofsound.c...-use-high-sample-rates

That makes sense. If you slow down by basically extending a sample to fill a space where no sample existed, if it didn't have to be extended as far it seems that would increase the potential for great fidelity. Ditto if you did sample skipping to transpose, like the old E-Mu gear.

It seems like a time stretch system could use oversampling when it's time to time stretch.
 
Just saying.

I am confused as to why a higher sample rate does not provide higher resolution?  If I take 12 samples per second.  It will sound very grainy and choppy, like looking at a very slow fan.  At one point does that graininess and choppiness go away?  
So it would seem that if computers were powerful enough and the rate of samples was extremely high, that there would be a higher resolution.
Meanwhile, I like 88.2 or 96K.  It does not impact my computer much, and provides lower latency.  But I have to admit, I have not tried 386 or something higher. 
It is hard to believe that humankind will go on for a million more years and that the sample rates will not go up.
 
 

deswind
I am confused as to why a higher sample rate does not provide higher resolution?  If I take 12 samples per second.  It will sound very grainy and choppy, like looking at a very slow fan.  At one point does that graininess and choppiness go away?  
So it would seem that if computers were powerful enough and the rate of samples was extremely high, that there would be a higher resolution.
Meanwhile, I like 88.2 or 96K.  It does not impact my computer much, and provides lower latency.  But I have to admit, I have not tried 386 or something higher. 
It is hard to believe that humankind will go on for a million more years and that the sample rates will not go up.
 
 

The crux of the sampling theorem is that if you filter out all of the frequencies greater than one half the sampling rate, ALL of the remaining data is preserved. This includes everything between the samples. You can't get higher resolution because ALL of the data is already there.
 
The problem is that the way that this works is not intuitive and involves a fair amount of math.
 
Or to put it another way, Nyquist and Shannon wouldn't have gotten a theorem named after them if everyone could just imagine how it works in their head.

mike_mccue
It seems like a time stretch system could use oversampling when it's time to time stretch.
 
Just saying.

Yes indeed.


But if the authors of the time stretching algorithm didn't bother to do this (for whatever reason), it does make sense to either run the process at a higher sampling rate or else use a different time stretching algorithm that already does what it should.