logic pro x 32 bit floating

A few things to add to octopi's explanations:

• THe ability to import 32-float files without converting them to 24b is on the top of my Logic wishlist (long overdue)
• Logic can already bounce your mix as 32-float
• Logic can already export files as 32-float
• When Logic creates freeze files, they are 32-bit (for many years)
• Although Logic is processing audio internally with 32-float (and 64-float), you have to be careful with Plugin that emulate legacy gear that could technically produce clipping.

Hope that helps,

Edgar Rothermich

Graphically Enhanced Manuals

www.LogicProGEM.com

It's never been a problem for me when people send me songs to mix.

The bit depth is basically your noise floor. The higher the bit depth the greater the dynamic range, and the lower your noise floor. That's the key point.

There are several parts in getting to grips with this.

Logic works at 32 bit float internally. When you make a recording your AD would clip anything over 0dbfs. On playback you can push the tracks way way louder, actually into the red but here's the catch. It has to, at some point come out into the real world and pass through your DA to get to the real world. If you're at the point where you're being sent or are recording so hot that you are clipping your outputs then you need to grasp your gain structure.

Your bit depth gives you greater dynamic range and you need that extended dynamic range as it is a multitrack recordiing. Anywhere between 2 and 100+ tracks. The gain gets summed into your 2 bus. The greater dynamic range enables you to keep the tracking gains way lower and not suffer any noise. Internally in logic, gains can be passed between plug-ins much safer as it's very easy to clip a plug-in.

The thing with bit depth is that it will never ever be perfect no matter how much computer CPU cycles you throw at it unless the bit depth became infinite. The good news is that it doesn't need to be perfect. The quietest gear these days is still noisier than a 24 bit noise floor. Way noisier.

If the recorded tracks are not clipping at 32 bit float, they won't clip at 24 bit fixed. I.e. the recordings are either clipped or they are not clipped. If they are clipped then there's little chance of me doing a decent mix. If the signals are hot (but not clipped) I pull all the faders down. Last song I mixed I had to start with all faders (aside from the master) at -5dbfs.

There's little point in having an extended dynamic range if it's not used. 32 bit float comes into its own whilst still inside the DAW. Yes, it would be handy if Logic could import the 32 bit float files and not convert them but it's never been a problem for me as the final output song will always come down to a fixed bit depth.

I suppose my bottom line is: I don't worry about things I can't control unless the actual recordings are clipped. Then there's no way out of it.

Hope some of this makes sense.

Mr. Rothermich,

Nice to see you back! I can just join others to say welcome back!

While I understand your request for 32 bits I/O audio encoding file exchanges. I don't see the commercial value for it. There are no workable use cases for 32 bits transductors in the physical world. I'm sure you understand that.

WD

Wilddrums wrote:

While I understand your request for 32 bits I/O audio encoding file exchanges. I don't see the commercial value for it. There are no workable use cases for 32 bits transductors in the physical world. I'm sure you understand that.

It is not about the transducers. 24bit fix-point has a theoretical dynamic range of 144dB which is greater than the noise floor of the electrical components of an audio interface. Although there are now 32bit fixed-point AD converter (I don't know why yet, besides bragging rights).

There is a commercial use for 32bit float audio files, and that is file exchange. When you export your mix and send it to the mastering engineer, there is not ned to export it 24bit fixed because your audio interface plays back in 24bit. You stay in the digital domain (audio file) deliver that file, and the mastering engineer imports it without any clipping no matter how hot it was mixed. Therefore, the delivery format for mastering engineers is 32-bit float.

32-bit float has a theoretical dynamic range of 1,500dB. Only at the last stage (i.e. mastering engineer or musician before uploading to SoundCloud), that delivers the file that is played back by the end user (through an audio interface "D/A"), only that stage needs to save the audio in the fixed point format (uncompressed audio data or compressed audio data).

The Dithering will then take car of the Quantization Distortion of the lower Bit Depth audio format and everything is fine.

Mr Rothermich,

I really - I meant really - didn't want to start an argument on this.

I think we just agree.

32 bits encoded files have some use for file exchange and no use in the physical world.

I just want you understand all the respect I can have to your opinion.

Respectfully,

WD

No. When you import a 32 bit float audio file into Logic it will convert it to fixed point 24 bit.

Nice to see you back on the forum Edgar. It's been a little while. We've missed you.

Yes, I'm trying to be a bit more active again (late New Years resolution)