32-bit float and dither?

Discussion forum for Amadeus users

Moderator: Martin Hairer

Post Reply
cd2go
Posts: 14
Joined: Sat Feb 09, 2013 3:56 pm

32-bit float and dither?

Post by cd2go »

Hi. Recently started recording in 32-bit float and am wondering how to handle saving the file to 24-bit for final delivery. Does Amadeus apply dither automatically upon saving or will it truncate unless you manually dither--and how best to do that?

Thanks in advance,
James

User avatar
Martin Hairer
Site Admin
Posts: 1975
Joined: Wed Nov 08, 2006 11:49 am
Contact:

Re: 32-bit float and dither?

Post by Martin Hairer »

A 32-bit float does only have 24 bit of precision for values between -1 and 1 (the rest is for storing the exponent, which would be 0 in this case), so the answer is that it is truncated, if you will, but that this does not actually result in any meaningful information loss. When saving to 16 bit, then the 8 least significant bits are discarded, no automatic dithering.

cd2go
Posts: 14
Joined: Sat Feb 09, 2013 3:56 pm

Re: 32-bit float and dither?

Post by cd2go »

Thank you. Last question: Is anything labeled as “32-bit” in Amadeus, actually 32-bit float?

For example, after importing a 32-bit float recording and making edits, does clicking “save” keep the file in float?

Post Reply