Benifits of running at 64bit floating point...?

Hi guys,

Ive been using cubase for years and am currently running Cubase Pro 8.5.xx and for the last few years i have been running all my sessions at 24bit/96khz and everything has been smooth sailing…

However now i am thinking of upgrading to 10.5.xx and have learned that it can run sessions at 64bit floating point and i am curious as to what benifits that might give over running at 24bit…?

Thanks for and info


You’re confusing bit depth of recorded audio with precision of the audio engine. Two separate things. You can store your audio as 32 bit float instead of 24 bit int, which does have some advantages in dynamic range when transferring files to be used in another project. The 64 bit you mentioned is the precision of the audio engine. That rarely has any advantage over the other option: 32 bit float (not to be confused with 32 bit audio files). Any processing that would benefit from 64 is already running at that precision, so there’s no point to using the 64 bit option.

Hmm if there is no point in selecting the option, then why is it even an option…?

Marketing. Logic Pro has a 64 bit mode. Even though it’s of little use, Steinberg felt compelled to add it to avoid the perception that it is missing a feature found in Logic.

Oh ok - so then is 32bit floating point pointless to use also…? - Or are there benifits to running in 32bit floating point over standard 24bit…? Thanks for the info btw :slight_smile:

If you record with VST-Fx in the input bus, recording in floating point resolution can prevent clipping that occurs at the output of the VST Fx processors. so values above 0db fs can be stored. This is also valid for offline processing, and exporting files that have an internal level of above 0 dB Fs. So if you set your project to 32 or 64 bit float, the results of all offline processing is stored in 32 bit float, therefore basically not causing clipping with too high levels - all internal in Cubase, not from recording too high levels in the analog world.

Where do we find that setting, please?

Studio -> Studio Setup… -> VST Audio System -> Advanced Options -> Processing Precision

PG (WaveLab developer) gives some interesting info in this thread …

What PG fails to mention in that other thread is that the noise level he’s talking about is low enough that you would need to cool your audio interface in a cryogenic chamber with liquid nitrogen before it would be detectable.

Also, the repeated computations he mentions that could theoretically have a benefit with 64 bit floats only happen within plugins, not at the mixing bus. Any decent plugins where that is an issue already use 64 bit floats for their processing.

Waves does but use 32bit float for I/O to the DAW.

Thanks. I really did try to find it in the Help, but I didn’t search for “processing Precision”.