Page 1 of 2

24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 4:15 pm
by richie43
Since polls are the fashion (and I just figured out how to post them) I am wondering what bit depth people are using and for what tasks. Thanks!

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 5:39 pm
by faun2500
32bit, of course! :P

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 6:32 pm
by richie43
faun2500 wrote:32bit, of course! :P


You track in 32bit FP? I thought that almost no converters actually capture more than 24bits. Any insight on that?

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 8:05 pm
by Mercado_Negro
24 bit while recording, 32 bit while mixing.

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 8:08 pm
by nino
I do track in 32bit also,
downside is that DAW records 24bit + zeros (the rest of bits).
Upside is that my DAW doesnt have to resample every event realtime when I start editing and adding fx.

Cubase auto resample to 32bit when editing even if your project is in 16bit.

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 8:10 pm
by Mercado_Negro
nino wrote:I do track in 32bit also,
downside is that DAW records 24bit + zeros (the rest of bits).
Upside is that my DAW doesnt have to resample every event realtime when I start editing and adding fx.

Cubase auto resample to 32bit when editing even if your project is in 16bit.


Resample in realtime? We're talking about bitdepth, no sample rate :)

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 8:24 pm
by RJHollins
24 bit recording,
32 bit[fp] mixing & mastering.

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 9:59 pm
by nino
Mercado_Negro wrote:Resample in realtime? We're talking about bitdepth, no sample rate :)


Well you may call it ''realtime convert to 32bit'' or ''realtime conform to 32 bit'' my source is mr. Friedemann Tischmeyer from proworkshops.de, he even says that Cubase/Nuendo uses more CPU power if working in 16bit or 24bit because files always get ''converted'' to 32bit..

His way is 44.1kHz 32bit
My way is 96kHz 32bit..just to be shure :)

Here is little something about analog summing, 24bit summing and 32 bit fp:
http://www.youtube.com/watch?v=3cdcffo27fY

Re: 24fixed or 32bit floating point?

PostPosted: Thu Nov 17, 2011 11:34 pm
by faun2500
richie43 wrote:
faun2500 wrote:32bit, of course! :P


You track in 32bit FP? I thought that almost no converters actually capture more than 24bits. Any insight on that?


I don't really 'track' as I make electronic dance music but I do sample vinyl and I sometimes bring tracks out to an analog mixer and back in again.

All in 32bit (I think). lol, I don't pay much attention really.

:lol:

Re: 24fixed or 32bit floating point?

PostPosted: Fri Nov 18, 2011 12:14 am
by mathias
24bit for tracking, as almost* no converter does 32bit, so it is just wasted harddiskspace.

32 bit fp, for all editing/mixing/summing/bouncing, intermediate storage or whatever you call it.


*there are converters that do 32 bit AD and i even own one (stagetec truematch). but at the moment, there is no way to transfer this to a file, as the digital-interface-standard does only support 24bit (adat, madi, aes/ebu,...).
24 bit gives you enough dynamicrange (144 db), to capture even the quietest microphone.