View Single Post
Old 31st December 2019
Lives for gear

Originally Posted by owly View Post
DAW channels and mix processed internally by "transparent math".
32-bit float in Live. However, if you're recording internally and the Bit Depth settings of Live are at 24-bit, the recording will be saved as 24-bit integer and when processed again the audio stream will be automatically improved to 32-bit float, as that is what Live runs at.

This works so well, that an unprocessed 24-bit integer recording of a 32-bit float source, will null completely with each other (if played back in perfect sync and at the same level) even as the bit streams are maintained in each track.

A frozen/flattened file is 32-bit float on disk.

At the audio interface of course you're likely truncating to 24-bit for the monitoring.