Which 32-bit format is better?

In the Save As... options menu there is a list of 32-bit file formats.  
I'm wondering which format is best to use?  

Are floats better than ints?  Are packed ints worse than unpacked ints?  

Which format is better?

In the Save As... options menu there is a list of 32-bit file formats.  
I'm wondering which format is best to use?  

By and large there's no difference in the 32-bit formats from a 'best' point of view, except that Audition has least to do when it opens its default one.

Are floats better than ints?  Are packed ints worse than unpacked ints?  

Floating point formats are going to be better than any int format below some ridiculous figure for the number of bits - around 256 bits of int roughly equals 32-bit FP. Packed or unpacked makes no difference - all this represents is a way of not wasting so much space by needlessly filling it with zeros, and using it for a byte of the following 24-bit word instead. But there is no inherent quality difference between the formats - except in their ability to store more detailed audio information.

Just stick with Audition's 32-bit FP format - you won't go wrong.

But you should note that the foregoing only applies to storage - for manipulation, a different set of rules applies.

[is]

Which format is better for manipulating, processing the audio?

Already told you...

I did some research and experimentation and it seems like 24 bit int is the most compatible for the programs that I use and for other programs out there that I've heard of.  It's not as good as 32 bit, but there are still lots of programs that can't open up 32 bit files, but can open up 24 bit files.  Rather than spend all that time converting files, I'm going with the greatest common denominator.  Besides, soundcards are still 24 bit too. 

I did some research and experimentation and it seems like 24 bit int is the most compatible for the programs that I use and for other programs out there that I've heard of.  It's not as good as 32 bit, but there are still lots of programs that can't open up 32 bit files, but can open up 24 bit files.  Rather than spend all that time converting files, I'm going with the greatest common denominator.  Besides, soundcards are still 24 bit too. 

That's fine.  Just remember that if you use AA, it will take much longer for you to open your 24-bit int files.

Besides, soundcards are still 24 bit too. 

... and they will remain so.  Nobody is ever going to produce 32 bit hardware.

If you are going to be doing a lot of processing, then 32 bit is the way to go in the software.

Graeme's absolutely right. And when you must have 24 bit recordings Audition will easily convert from 32 bit.

[to]

Graeme's absolutely right. And when you must have 24 bit recordings Audition will easily convert from 32 bit.

Alan, Just to clarify and relieve me of some of my confusion, did you mean to say "Audition will convert to 32"?


I know that a sample rate discrepancy between a sound card setting and an audition session can create synchronization problems during recording and playback.  I've experienced this first hand and read about it on this forum.  However, does the bit rate factor into this at all?


Out of general curiosity, why doesn't Audition have an option of recording at 24 bit? Forgive me if this is a completely ignorant question.

[to]

Graeme's absolutely right. And when you must have 24 bit recordings Audition will easily convert from 32 bit.

Alan, Just to clarify and relieve me of some of my confusion, did you mean to say "Audition will convert to 32"?

I think that what Alan said was quite clear - you process your material in 32-bit in Audition, and if you must have a copy of it for use in other software that can't cope with that, then Audition will convert the material easily from its 32-bit form into a 24-bit file.

I know that a sample rate discrepancy between a sound card setting and an audition session can create synchronization problems during recording and playback.  I've experienced this first hand and read about it on this forum.  However, does the bit rate factor into this at all?

No - this is simply a synchronisation issue. The bit rates have to be initially identical to synchronise, and the act of synchronisation brings the two clocks into phase. In reality this isn't what happens with audio systems though - an external clock source is used to trigger an internal source to generate sync pulses at the same rate as it is running. This isn't really synchronisation at all, but regeneration. For the sake of simplicity it's referred to as 'synchronising', but as anybody who's ever worked with genlock systems and live TV studios will tell you, it's not the same thing at all!

Out of general curiosity, why doesn't Audition have an option of recording at 24 bit? Forgive me if this is a completely ignorant question.

It does. 32-bit FP when it is recorded is effectively  24-bit audio, with all of the scaling bits set to zero. There is a bit more to it than this, because it's translated to a different scaling system from a conventional integer one, but fundamentally it's still 24-bit audio that's being recorded - which is why it's not difficult to store it in one of the conventional 24-bit integer formats if required.

I'm clearly lacking some fundamental knowledge in this area. I tried doing some research on this but could not find anything useful. Does anyone know of an online resource that deals with these issues?

I understand that it is possible to process material in audition at 32 bits and then convert it to a lower bit-rate for the purposes of other software.  What I don't quite understand is what happens if you record at 16...can you then process at 32 without negative consequences? Conversely, what happens if you record at 32 and process at 16?  Audition has modifiable options pertaining to these very scenarios (settings -> multitrack, etc.) and I don't know which options to choose.

See this long Adobe Audition FAQ thread for more discussion about 24 versus 32 bit.

http://www.adobeforums.com/cgi-bin/webx?128@@.2ccfa51e

Thanks a lot for the link. I've read it and it has answered many of my questions.

[to]

Graeme's absolutely right. And when you must have 24 bit recordings Audition will easily convert from 32 bit.

Alan, Just to clarify and relieve me of some of my confusion, did you mean to say "Audition will convert to 32"?

I think that what Alan said was quite clear - you process your material in 32-bit in Audition, and if you must have a copy of it for use in other software that can't cope with that, then Audition will convert the material easily from its 32-bit form into a 24-bit file.

I imagine Steve's answer has cleared up that point.