Quote:
Originally Posted by JFreak
32bit float has advantages in ONLY two scenarios:
- Your converters record 32bit float
- You use virtual instruments and want to freeze/commit
|
I think you might be describing 32-bit integer, like the converters included in the Avid Carbon? They extend the ‘resolution’ deeper into the noise floor. Most people are a bit sceptical of it. Bigger numbers make for great marketing opportunities.
32-bit floating point deals with the business end at 0dBFS. It’s about the same 24 bit precision as 24 bit, but with an additional 8 bit exponent that expands the dynamic range. You can push a 16 or 24 bit audio file over 0dBFS and render it as 32-bit floating point and not lose a single bit of audio in the process. It gets used in a bunch of production sound recorders by Sound Devices and Zoom and they allow you to capture even the most unexpectedly dynamic performances or sound effects without the fear of clipping of levels when over 0dBFS.
If a 32-bit floating point file has audio above 0dBFS, it will clip as soon as you export it as 16 or 24 bit.