Sample rate matters

[Rawer]

Hi,


I am working on a mix, with my mbox mini, that can goes 24 bits/48khz

I read that it is somehow better to mix in 32/96 and 24/96

I also have the maudio firewire 410 that can go very high on the sample rate but I really prefer my mbox for some reason, so my question is it really best to mix in 32/96?

Thanks in advance

[Bill Denton]

Please don't take offense, but if you're asking that question 24/48 is probably adequate for your needs.

Do some research on Nyquist sampling theory and read some of Dan Lavry's white papers (http://www.lavryengineering.com/lavry-white-papers/).

It is pretty much agreed that 32-bits is generally better as it allows more headroom than does 24-bit. But as of a month or so there were no 32-bit AD/DA converters widely available, so the 32-bits only comes into play inside the DAW software. And if you're using proper gain staging you won't really gain anything from 32-bits.

As far as sampling rate goes, there are claims that even 44.1 is overkill. Some claim that higher sampling rates somehow sound better.

There are some plausible claims that 96 kHz is a sound "sweet spot" due to filter designs and (insert scientific gibberish here).

But in the end, the only thing that matters is what comes out of the speakers (or earbuds, as the case may be).

Do the research I mentioned, then test some different settings for yourself and make your own decision...

[dr_daw]

+1 on Bill's post

Also remember that if the original recording is at 24/48k using 32 float in the mix may be beneficial but converting the sample rate will not do anything but cause grief. Once it's recorded digitally you don't gain anything by re-sampling it to 96. If you just go and change the sample rate on your hardware, everything will just play back faster.

Good luck and don't over analyze things, just have fun and use your ears.

[Rawer]

Originally Posted by Bill Denton

Please don't take offense, but if you're asking that question 24/48 is probably adequate for your needs.

Do some research on Nyquist sampling theory and read some of Dan Lavry's white papers (http://www.lavryengineering.com/lavry-white-papers/).

It is pretty much agreed that 32-bits is generally better as it allows more headroom than does 24-bit. But as of a month or so there were no 32-bit AD/DA converters widely available, so the 32-bits only comes into play inside the DAW software. And if you're using proper gain staging you won't really gain anything from 32-bits.

As far as sampling rate goes, there are claims that even 44.1 is overkill. Some claim that higher sampling rates somehow sound better.

There are some plausible claims that 96 kHz is a sound "sweet spot" due to filter designs and (insert scientific gibberish here).

But in the end, the only thing that matters is what comes out of the speakers (or earbuds, as the case may be).

Do the research I mentioned, then test some different settings for yourself and make your own decision...

I understand, my situation here is not about sampling sounds, I have a all session to mix, just wondering if I should change the session to another sample rate to have a better sound at the end, I hear you on the fact that it is better for sampling in 32/96 tho

[Bill Denton]

Originally Posted by Rawer

I understand, my situation here is not about sampling sounds, I have a all session to mix, just wondering if I should change the session to another sample rate to have a better sound at the end, I hear you on the fact that it is better for sampling in 32/96 tho

I fear you really don't have much understanding of how all of this works...

I am not talking about "sampling sounds"...that's a totally different ballgame.

When you record someone singing or playing the guitar, it creates an analog signal going in to your MBox. From your MBox, the signal goes to the computer. But before the MBox sends the signal to the computer it must go through an A/D (analog-to-digital) converter because the computer only understands a digital signal.

It is here, in the A/D converter, that sample rate really comes into play. Once the signal has been changed from analog to digital, that's it. That determines how much actual information is recorded.

Say you make a recording at 24/48. If you then change your session to 24/96, you don't magically have any additional data recorded. The computer just does SRC (sample rate conversion)...it doesn't really add any more data to the digital signal.

You really need to read the information I noted earlier, and you need to read about the fundamentals of A/D and D/A conversion, and how a DAW actually works.

In the meantime, just use 24/48 and you'll be fine...

[acelo]

It has been said that it's the actual plugins that sound better at higher sample rates.

[YYR123]

Originally Posted by Bill Denton

There are some plausible claims that 96 kHz is a sound "sweet spot" due to filter designs and (insert scientific gibberish here).

Love it

[YYR123]

Originally Posted by Rawer

I have a all session to mix

Mix it as is - 24/48

Originally Posted by Rawer

just wondering if I should change the session to another sample rate to have a better sound at the end

No bill answered that - don't change anything after the fact (recorded)

Mix for what u have now

[CME]

Yep and just in case this point eluded you, always record at 24-bit fixed regardless of sample rate. There are very few converters that can do 32-bit float. Once you're mixing you "could" change the session to 32-bit and gain some extra headroom in the mixer. But proper gain structuring basically negates that also.

And finally to reiterate. There's no reason to SRC your session to a higher resolution. Nothing to be gained and more hard disc space will be used. It will also tax your CPU more. If you want to use 88.1/96k or higher, do that in the recording stage. :)

[JFreak]

Originally Posted by Bill Denton

There are some plausible claims that 96 kHz is a sound "sweet spot" due to filter designs and (insert scientific gibberish here).

Yes and no.

You are right in that 44.1/48k filters are not that good, and they in most cases introduce some problems to the audible <20kHz area. So the freq needs to be upped.

But no, the industry got this all wrong when they decided that the sampling frequency needs to be doubled. That is wrong, it is all marketing. Double this and you get more and we get to charge a premium. That's what it's all about.

Anyone remember a 32kHz sampling rate? Yes, there was that too. Should the industry have any more technical understanding than a monkey with a banana, they would have made 64kHz sampling a standard. There the filter imperfections are not affecting what a human ear can hear, but the "cost" of storing and processing such audio would have been only 67% compared to 96k.

Just my view about this. And that said; good converters do a very good job at 48 as well. No need for higher sampling rates really, only for higher quality conversion (and filters). I've seen results from not so good converters (aka 002rack) at 96k being pretty much equal to better converters (aka digi96) at 48k so the filter design and analog stage is really very important. Sample rate in itself is not.