render and mix to 24-bit?

[jbm]

I've been tossing around overall mix ideas for an album project I'm working on. Today I had a sudden flash of anxiety when I realized that I'm in the time-worn habit of bouncing tracks in 16-bit while working in Logic. Now, it's not too late to go back and re-render those bounces at 24-bit -- as I say, I haven't really settled on the overall sound of the album yet -- but I'm wondering what everybody's general practice is when it comes to bit-rates? Should I be updating all my sessions to 24-bit, re-bouncing everything I've rendered, and starting from there? Do people just start their entire session in 24-bit now? Does anybody here even touch 16-bit anymore?
I feel like a confused old dog... (and I'm far too young to feel that way!)

J.

[KMuzzey]

Hey J,

I actually stopped doing my mixes in 24-bit and went back to 16-bit after reading somewhere that it's better to record natively in 16-bit than to record in 24, then downsample it to 16-bit for the purpose of putting it onto a CD. I suppose if a piece of music was going to live in its digital state forever and be transported by FTP or email, then 24-bit recording might be ideal.

Can anyone else shed some light on this -- is what I've read correct?

Kerry

[Peter Roos]

My habit is to bounce to 24 bit and master in 32 bit FP (and then dither down to 16 bit using Waves L2.
I think I can really hear better results. If your bounce is your final "master", then there may be less need for 24 bit. But maybe for your own listening and archiving pleasure 24 bit (or higher) is a better format IMHO. I did my latest IR-library (recording, processing (EQ's), fades etc) in 32 bit FP, and the IRs are definitely better than 16 or 24 bit versions. I regred I did not do it all in 96 Khz 64 bit...

BTW, DAWs have internal 24 or 32 bit busses, so why record in 16 bit? Most audiophiles agree it is too limited for pristine recording and playback.

[jbm]

You see, this all gets murky to me... My cards are native 32-bit (RME cards). The VIs are 24-bit. But generally, when I bounce, I've been bouncing the files to 16-bit. God knows why. As I said above, it's just a habit, I suppose. I don't feel like I've noticed any degradation, but then, I'm not really looking for it, so there's no attempt made to "A/B" them (and there's really nothing to A/B, unless, I suppose, I listened to the original VI playback against the bounce(??) Which might actually be the best test.).


As I understand it, Logic uses internal 32-bit FP processing throughout. Now, I generally don't choose to dither when I bounce, but maybe that's a mistake as well?

---------- the passage of time... image goes foggy... reverberated voices... then to black...------------------------------------------------------------------------

Okay, I read up on dithering. So I'm guessing that I've been really silly not to dither my 16-bit VI bounces, since this is absolutely a step-down from 24-bit to 16-bit. Yes? Dithering at this stage would presumably be as good as I could get in Logic anyway. No? OR... are effects and so on placed on the bounced VI tracks going to have more depth and transparency in *future processing* (spatialization, etc.) if I bounce to 24-bit (no dither), then use only 24-bit files in my mixing process.

Last question then: Are people generally happy with Logic Pro's dither?

Thanks again (and in advance),

J.

[mathis]

James, not much time right ow, sitting in an internet cafe. But I strongly recommend rebouncing in 24 bit, or better 32 bit if Logic allows that. It does make a difference.
Go to 16 bit at the very final end.

[Dietz]

Like Mathis says. It makes no sense to go to lower bit-resolution when you're not aiming for a final delivery format (i.e. CD, MP3, and the like).

[jbm]

mathis,

I just noticed you were on that other thread, where I posted with all the might of my 37 IQ points...

Thanks for the definitive support of higher bit-rate bounces. I reckon I should try it, at least. Can't hurt, can it?

cheers, and enjoy the coffee!

J.

[jbm]

--- and ditto, Dietz! [;)]

J.

[JWL]

@jbm said:

Thanks for the definitive support of higher bit-rate bounces. I reckon I should try it, at least. Can't hurt, can it?


J.

J--

"Trying it" was my suggestion. Experiment with bouncing at two different bit rates (or 3 as the case may be) and just compare the integrity of the sound. We've heard all the theories and techniques, but at the end of the day it's what sounds the best to your ears that counts.

[jbm]

JWL,

true, true.

Yesterday I re-bounced one of my pieces in 24-bit (the tracks from my slave machines, that is). It sounds good to me. It sounded good to me before, as well! However, the one reason I'd like to stick with the theoretically 'sound' approach is that the various stages in my mixing process are all temporally isolated -- they happen at discrete moments in time, separated by periods as long as a couple of weeks -- and it would be quite awkward to A/B different approaches during each of these stages. Each stage may also involve further bounces or, more likely, the application of reverbs, compressors, and so on, to already-bounced tracks. Over time, and with each new stage in the process, there is the potential for introducing noise, but the process could be so slow that I may not really notice from one stage to the next. However, if I take the wrong approach at the A->B stage, I might genuinely notice it at the B->C stage. Which is to say I could probably tell right away on an A/C listening (air conditioning?...), which would be too late, since B is effectively cast in stone.

How's that for a convoluted (convolved?) explanation!

The other factor is that in my studio, on my system, all may be well, to my ears. But on a different system, in better-designed room, problems may become more obvious. So, if I keep things theoretically intact, I don't necessarily have to depend only on my ears, which are fickle at the best of times! And after all, bouncing is a strictly technical excercise -- it shouldn't require even a moment of thought or contemplation: do it right, do it once, and move on!

cheers,

J.