• Welcome to the General Discussion forum for UAD users!

    Please note that this forum is user-run, although we're thrilled to have so much contribution from Drew, Will, and other UA folks!

    Feel free to discuss both UAD and non-UAD related subjects!

    1) Please do not post technical issues here. Please use our UAD Support Forums instead.

    2) Please do not post complaints here. Use the Unrest Forum instead. They have no place in the the General Discussion forum.

    Threads posted in the wrong forum will be moved, so if you don't see your thread here anymore, please look in the correct forum.

    Lastly, please be respectful.

24 bits vs 32 bits

Sonic

Active Member
Is there any advantage of recording at 24bits and mixdown to 32 bits?
 

giles117

Active Member
Besides Greater Definition. I cant think of any.... ;)
 

geekeye

Member
24 bit has 144 db of dynamic range, and 32 bit has 192 db. since no known d/a or speaker system can reproduce the full range either, the difference is pretty much moot.
 
Sonic said:
Is there any advantage of recording at 24bits and mixdown to 32 bits?
- there is no "0dB limit" => floating point
- there *can* be an advantage for audio processing

I usually record tracks in 32bit float (cubase SX) with 24bit resolution on my ADC. During all the steps (mixing / mastering), I work with 32bit files and cut it down to 16bit at the very last step. I didn't do any tests or comparision with using 24bit files during mixing or mastering. My PC can handle 32bit files (project size and PC performance), so I that works for me very well.
If you have problems with your DAW performance you might work with 24bit....
 

Sonic

Active Member
But is there any advantage of recording some tracks at 32bits and then change the project to 24bits and mixdown to 24bits?
 

cAPSLOCK

Active Member
Beofre this question can really be considered you have to make a distiction and understand the difference between fixed and floating point methods. Just throwing out numbers can be a bit (haha) confusing.

Also... as far as I know, converters are pretty much always fixed point devices.

Upping the bitdepth to floating point is useful when doing processing without running into headroom problems. The idea manefests in the DSP world in the preservation of the low level detail even more than clipping as long as proper recording/mixing methods are used.

In the end a properly recorded and carefuly mixed 16 bit recording can sound better than a badly engineered high resolution recording of any type.

You have to consider what you are adding (or changing) and what you are throwing away when you do conversions.

http://www.uaudio.com/webzine/2003/july ... tent2.html

http://www.seneschal.net/papers/bitstre ... eam035.htm

cAPS
 

giles117

Active Member
Now that is the nail on the head.

Ultimately it is always how well you do your job.

Remember there are recordings done WAY back in the day that STILL have more fidelity than some stuff out today :)
 

jcat

Active Member
32bit float is 24bit resolution with a sliding scale. So it's the same quality as 24bit (with some minor exceptions), but it can slide this scale to accomdate whatever level the material is. I.E. 0dB plus...

The beauty of a 32bit float mixer is that you can't overload it anywhere in the mixer signal chain, as long as you pull the master out down at the end to below 0 you'll be happy as larry...

8)


Cheers,

jcat
 

Michael

Active Member
If you apply any processing to any of the tracks . This is where 32 bits is needed and has the advantage over 24 or 16 . If you intend on doin no processing then it wont make a difference at all other that overloading the levels going out
 

klimt

New Member
There's no point in using 32bit float files since the ADDA converters are 24bit at both ends. For Cubase/Nuendo - its always running in 32bit float internally which is great, but you really need to dither the output back to 24bit before it hits the DAC otherwise all that internal precision just gets truncated to 24bit.
 

JamesR

Active Member
if you have the patience....

http://www.jamminpower.com/PDF/48-bit%20Audio.htm

talks about 24bit fixed and 32 bit foat.

32bit float allows for some headroom and processing with less rounding errors, especially in the multiply and add functions commonly used in DSP.

record in 24 bit because that's all your AD will use, process in 32 (you have no choice here if you use cubase, logic etc) (PT is 48bit fixed point)

summing in your digital mixer, the output will be 24 bit.

Logic freeze files, like the rest of it;s internal processing is 32bit.

\"pulling the master fader down\" to compensate for overs is exactly the wrong thing to do. this will greatly reduce the accuracy of the system as the headroom will be used up, with less bits for signal.

ideally you would mix your channel leverls so as to be close to zero as possible, without overing. although with floting point mixers you will not hear horrible digital distortion, going over 0db reduces the resolution.

J.
 

Rankus

Member
JamesR said:
"pulling the master fader down" to compensate for overs is exactly the wrong thing to do. this will greatly reduce the accuracy of the system as the headroom will be used up, with less bits for signal.


J.

Pretty sure that this statement is incorrect.....
 

Rankus

Member
JamesR said:
ideally you would mix your channel leverls so as to be close to zero as possible, without overing. although with floting point mixers you will not hear horrible digital distortion, going over 0db reduces the resolution.

J.
You are far better off with -3 to -6db as an upper target..... You need to leave headroom for processing...

Even though 32 bit mixers a will not clip, many pluggins will.
 

JamesR

Active Member
hey there,

when i mentioned \"without overing,\" maybe it would clearer if i said \"without overing anywhere in the chain.\"

i was actually referring to the level at the main o/p, but maybe that wasn't clear.

clearly, a plug might well add gain that would produce an over within the channel itself.

More than one channel without o/p signal levels of 0db will obviously create over in the summing / master buss.

so, yeah, mix so that the master signal level is approaching full scale, but not over.

J.

ps: i take it you didn't enjoy the article..?
 

JamesR

Active Member
ps: i think the other statement is cool, but if you have some time, sure, look into it...

J.
 

Rankus

Member
Hi James, Sorry if my earlier post seemed grumpy... I had not had morning coffee yet. Also, this has been discussed to death on ALL the forums. Do a search on this forum in particular looking for input from John at Massive Mastering....

But, anyway,.... The common industry thinking is that The 2 buss should be peaking at around -3 to -6db as well in order to leave headroom for the mastering engineer. In fact these levels should be maintained at all points in the path including input (while tracking).

Aiming for 0db is a holdover from the 16 bit days.... It really is not an issue anymore.

By leaving more headroom throughout the entire recording you will notice a gain in clarity and imaging etc.

I havent read the article you posted due to time constraints.... But there are some realy good books out there by Nika Aldrige (sp?) and Bob Katz that you may want to have a look at, before posting this type of information.

Best regards,
Rick
 

JamesR

Active Member
all good rankus,

i had actually forgotten about the mastering engineer!

although in all of the places i go to in london, the first thing they do is play it back from a regular CD through a Prism DA and then work with analogue gear....

(i'm not one for discussing \"which is better\" (number of bits, mac and pc, whatever) i normally just get on with making music the best i can, i just saw some posts that i thought i'd respond to...)

J.
 
UAD Bundle Month
Top