Sample Rate Questions

[plaamook]

Ok, here we go...

I record and process field recordings at 96k. At some point I changed Reasons sample rate to 96k as well to accommodate this (which I think is the correct procedure) but when I really start stacking it up the DSP meter goes bananas and I get drop outs and so on because I'm only on a little 13" MBP 2.9GHz i7 with 16 GB RAM and a HD.

So, can I work in 44k then change the settings and bounce in 96k?

Thanks in advance.

[Gaja]

If all the files you import into Reason are 96kHz that should be possible. Of course if you're creating derivates of your original recordings, you'll have to switch to 96k for it to properly work in the end.
But also really important: why do you need to have files @96k?
There's no audible difference in the outcome, if you're not stretching or pitching a lot (well I don't hear a difference, and don't know anyone who can).
So what is the reason for you to work with 96k files?

[plaamook]

Yeah, I'm stretching and pitching a lot. That's why I started recording in 96k to begin with.  And yeah, I'd thought of that, changing it for any kind of in project rendering as well. Hmmm... what a pain in the ass.

Anyway, cheers dude.

[Benedict]

If output is to be CD or mp3 then work at 16/44 so you hear the right things as well as save a lot of CPU.

[plaamook]

Well that just it. If I'm processing 96k files to all hell, pitching them way down, etc, I'll want it in 96k. But CD output is 44k as you said. So its bounce everything to 96k for mastering, then out at 16/44 for production.

Man, I just know somewhere in there I'll forget to change the sample rate...

[Benedict]

I thought Reason treated all (acceptable) sample rates as equal so to speak so you really don't need to worry. If your core file is 24/96 then the samples are still all there. I'd still work at 16/44 as if you are messing with the samples, quality isn't exactly paramount and the renders will be correct either way.

[plaamook]

Right, well, I think that solves it.

Thanks for your very speedy help as usual guys. Till next time. Over and out.

[ScuzzyEye]

So, can I work in 44k then change the settings and bounce in 96k?

Reason will create proxy files at the working sampling rate, but will always refer to the in-project data when exporting.

So if you record at 96k, but set the Reason session to 44.1k, the 96k data is retained, but temporary files are created for the 44.1 session playback. If you then export the song or tracks at 96k, they will reference the original recordings. If you export at 48k, new temporary files are made at that rate.

There is a but, and it is pretty major, if you're trying to keep 96k audio in the project. If you use the "bounce to new recordings" command on an audio track, it will store the audio in the project file at the current working rate. So if you need to do that, and want to keep 96k, you'll need to change your settings before you make the new recordings, and you can change it back after (there will be processing time as those new recording are proxied into temp files).

[plaamook]

Thanks guys. I'm on it.

[selig]

Well that just it. If I'm processing 96k files to all hell, pitching them way down, etc, I'll want it in 96k. But CD output is 44k as you said. So its bounce everything to 96k for mastering, then out at 16/44 for production.

Man, I just know somewhere in there I'll forget to change the sample rate...

Are you sure your A/D captures the additional frequency range, or does it just use a more gentle filter but still starts rolling off at around 20 kHz? Most higher sample rate A/Ds do the latter as far as I've heard. Folks who record bats etc. are a good source of information on how to TRUELY capture ultrasonic frequencies for later down-pitching.

[plaamook]

Are you sure your A/D captures the additional frequency range, or does it just use a more gentle filter but still starts rolling off at around 20 kHz? Most higher sample rate A/Ds do the latter as far as I've heard. Folks who record bats etc. are a good source of information on how to TRUELY capture ultrasonic frequencies for later down-pitching.

Its more a case of what happens when you pitch it down 2-3 octaves.

[selig]

Are you sure your A/D captures the additional frequency range, or does it just use a more gentle filter but still starts rolling off at around 20 kHz? Most higher sample rate A/Ds do the latter as far as I've heard. Folks who record bats etc. are a good source of information on how to TRUELY capture ultrasonic frequencies for later down-pitching.

Its more a case of what happens when you pitch it down 2-3 octaves.

Not if there's nothing "up top" to make any difference. I'm asking about the 96 kHz sample rate choice - are you saying even if you didn't capture data above 20 kHz, that the higher sample rate makes a difference when pitching down 2-3 octaves (sincere question, btw!)?

[plaamook]

Well as far as I can tell when you pitch something down you're loosing loads of top end. The whole thing shifts down and you have a big space at the top. So I'm assuming that if you record at 96k you're preserving some detail up there. And if you can't, why do people bother?

I'm totally prepared for the possibility that I've been wasting my time and energy on this by the way...

[plaamook]

I mean the top of what you can hear.

[selig]

Well as far as I can tell when you pitch something down you're loosing loads of top end. The whole thing shifts down and you have a big space at the top. So I'm assuming that if you record at 96k you're preserving some detail up there. And if you can't, why do people bother?

I'm totally prepared for the possibility that I've been wasting my time and energy on this by the way...

That's been my whole point all along, to point out that a sample rate of 96kHz doesn't define the frequency response in ANY way. It doesn't mean that the A/D filter is place ANY higher. It's often still left at 20 kHz, and a more gentle slope is used (gentle slopes affect the sound less the steep slopes). 

Again, you're talking ultrasonic frequencies here - check out the gear folks use to record bats. They can't get the same response simply using higher sample rates alone - you also need to raise the A/D filter frequency. Not saying your setup doesn't do this (don't know what you're using), just that it's NOT at all a given that a higher sample rate means you will capture a flat frequency response to a higher frequency than with a lower sample rate.

[plaamook]

Hmmmm....

This is interesting. I switched to 96k because, though mostly I use my own recordings, I've occasionally bought some from other people and most prof recordings are done at 96k or even 192k. Their samples cut off the same as mine when pitched down. So, this may mean that we're all wasting our storage on nothing because the recording gear isn't actually doing the job!

Cheers Selig, I'm gonna go test the hell out of this and get to the bottom of it.

[selig]

I found articles like this interesting a few years back when I was researching this very thing. 
Here's a good discussion on your specific subject:
http://sound.stackexchange.com/question ... requencies

[plaamook]

Actually they aren't cutting off. Its just that there isn't a lot of interesting stuff going on up there once you drop it down 3+ octaves and jack the eq to all hell. This may depend on what you're recording, bats for example, but in a diving bell its just a garbled mess up there.

I'm thinking that it probably pays to consider what I'm recording and what I'm recording with much more than I have or will. I don't have the money to get any deeper into what I do than I am at the moment. Which is ok really. I can polish a turd just fine. I'm just not going to bother trying to polish em at 96k!