Jump to content

Logic now supports 32bit audio files - so what happens when...


GuyBorlander
Go to solution Solved by Holger Lagerfeldt,

Recommended Posts

I've recently become aware that logic supports 32bit audio files (few updates ago) so what happens if you have a stem consisting of 32bit drum samples is exported at 24bits (downscaled?)

most engineers don't like to work with 32bit float exported stems so what is my best way around this? does this downscaling affect quality? I can't hear a change but I want to know how this works...

thanks guys!

Link to comment
Share on other sites

  • Solution

The short answer is: (pratically speaking) nothing.

32 bit floating point quantized/truncated or even dithered to 24 bit integer doesn't change sound quality in any audible way.

Floating point is practical because it allows for internal processing or summing headroom, and you can scale the signal up and down without loss, before committing to integer.

Logic uses up to 64 bit float internally, both for summing (if enabled in preferences) and with some plug-ins, the Channel EQ for instance. The output generally seems to be truncated or dithered to 32 bit float again, though.

The only and expected exception is the Bitcrusher plug-in, although it's either bugged or they didn't care that the nominal 24 bit integer output is actually 32 bit float. Only once you go to 23 bits or below does it truncate as stated. But I digress.

Very simplified, the additional 8 bits in 32 bit float are primarily used for processing or summing headroom. The sound doesn't have more resolution per se. So approximately 24 bits of actual audible signal is still coming out at the other end regardless of coming from 32 or 64 bit float.

I seem to remember from a test I did some years ago, that Logic automatically TPDF dithers (i.e. flat) rather than truncates when converting from float to integer.

It's debatable whether dithering from i.e. 32 bit float to 24 bit integer has any real life value, but you can measure a tiny difference from 32 bit float to double precision (64 bits) with any near-infinite processing like a high pass filter.

I always dither, but avoid noise shaping on stems that will be re-processed. The closest you get to TPDF in Logic Pro is POW-r#1.

In any case, this is all well below the noisefloor of what any AD converter can record and what any DA converter can play, let alone what any person can hear in a listening environment.

Of course quantization noise can accumulate in a session, but that's not an issue with well recorded or digitally created 24 bit audio to begin with.

  • Like 3
  • Love 2
Link to comment
Share on other sites

Wow thank you for this incredibly in depth response! I’ve taken this into account…

I ran some of my own tests

I put one of my 32bit samples into sampler with “original” setting and drew out a midi note playing the sound back at original pitch - I then did a duplicate track and exported new version at 24bits . also made sure both sounds are identical volume dBTP (reading of 0.00dBTP)

I played them both back together at the same time but with the duplicate (24 bit exported audio file version) phase inverted ; the result is that they cancel each other out completely so that tells me they’re “practically identical” like you’ve said.

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...