16/24/32 floating/fixed can someone explain?

[milesholmwood]

Can someone explain all the differences?

is this 32 bit PROCESSING or 32 bit RECORDING depth. and really whats the difference with that?

Does it mean you record at 32 bit? and how does this effect bouncing the session?

32 bit at the input stage or output stage...

I just keep getting the run around from people at music stores... They have no clue.


Thanks

[jhamrick]

I've been researching as well... And still continue. I've read arguments in the Summing wars ITB vs OTB where with 32-bit float you don't need to sum outside the box. I'm not going to get into a summing discussion as it is subjective. Here is something I read which makes sense when using higher track counts...
The 32-bit float format is a 1-bit sign and a 23-bit mantissa, which adds up to 24 bits of dynamic range from smallest to largest representable values (for a given range). But it also carries a handy 8-bit exponent, which means you can turn the gain up or down by 100dB and print it, and then somebody else can come along and turn the gain down or up by 100dB and you get back what you started with, no loss. If you record in 24-bit format alone, when you print at -100dB or +100dB, you'll get a lot of noise or a lot of distortion if you try to undo that.

In the film world, where hundreds of tracks are the norm and when it is quite possible that one wants to make lots of changes to levels across entirely different parts of the process, floating point makes a lot of sense. So much so that products like the Harrison Xdubber will record in 32-bit floating point (and feed their consoles which can process 64-bit floating point at 96kHz).