16 Bit to 24 bit?

No loss at all. What you're essentially doing is adding 8 bits of all zeros.

I would up the bit level prior to mixing. Any mixing processes/plugins will still benefit from the added "math space".

Interesting--coming from ProTools, I was always under the assumption that a 24-bit project would sound better than a 16-bit project, even if all the files originated as 16-bit, for the reason lwilliam gives: better resolution on the fader/plug/mixing math.

But with Logic, if I understand correctly, there is no project-wide wordlength. You specify whether or not to record 24-bit files, but Logic lets you include both 16 and 24-bit files in the same project,right ? I assumed from this that there would be no sonic advantage to bit-extending audio files before mixing--Logic is going to give you the same 32-bit floating math either way.

Of course, recording at 24-bit gets you a better source audio region, and a 24-bit bounce has more resolution than a 16-bit bounce.

But, does anyone know for sure that converting 16-bit audio regions to 24-bit will result in higher resolution mixing ? Seems "logical" to me that it would not...

James
james@nashvillain.com

Dual 1.25 G4 Mac OS X (10.4.8)

I'm not so sure about that. Logic stores its audio in 32-bit float format, which is converted from the originating audio file. Thus, if a change were made to 24-bit, I couldn't see it just tacking on zeros as it would in fixed format. I could see it as being written back with less resolution between samples within the 24-bit width (greater jumps between samples), seeing as the audio is being converted from 32-bit float to 24-bit fixed.

jord

It's true that you can run a Logic session that includes both 16 and 24-bit files. It's one of the cool things about Logic that it will let you do that.

But I don't think that Logic is so aware of the bit depth of any particular file that it would set up a different summing scheme based on the bit depth of the source audio. If you wanted to be a lil' superstitious about it, you could enable the "24-bit recording" checkbox in your audio prefs.

I don't see the advantage of changing your 16-bit files to 24, though. That extra 1/3 amount of data per file (or per side of a stereo file) is only going to reduce efficiency when it comes to handling your tracks. Maybe not by much, but then again it would all depend on how many tracks there are.

My 2-cents.

I agree mr schwartz. seeing as logic works internally at 32 bit float, then all the extra numbery-precisioney headroom that plug ins benefit from in their calculations is already there. ideally you should just record at 24bit to start off with and be done with it.

Yeah, to reanswer this post, the 16 or 24 bit file is converted ( in realtime ) to 32 float format before entering the audio engine/mixer, or immediately. You will only 'improve' your mix from extending the wordlength ( 16 to 24 ) if you have file based ( destructive ) operations, such as gain change, normalize, or several other offline effects ( though not all, such as DC or Reverse ).

You would bounce as high as possible, given the media. Since it is being mastered outside, bounce at 24.

J

Good point about the internal resolution!

I thought that ALL DAWS except 1. Sonar which has a 64-bit mode and 2. Protools HD which uses 48-bit fixed, were using 32-bit floating math.

If that's the case, then adding 8 zeros to the bit depth wouldn't give you any benefit UNLESS (as was mentioned above), one or more processes re-writes the file back to disk. It would give better resolution in that instance.

I can't see it doing any harm to the sound simply by adding 8 bits of zeros. It will simply make the hard disk work harder.

does anyone else feel that we are splitting hairs a little here? all this esoteric pico-precision we're talking about here all becomes irrelevant in an instant with a bad mix, a rubbish song or converting to mp3 when you're done..

..just to keep some perspective..

tbird, what response(s) are you referring to?

ok this is what I am saying:

i'm trying to put into perspective the value of wording up audio already recorded at 16bit to 24bit in order to give more mathematical precision to your plug ins and so on when you get into mixing, all in the context of logic's 32 bit float engine. this to me is splitting finite nano-hairs, the improvement you'd get, if any, is not exactly going to be life-changing. just get on with your mix and focus on getting your music to sound good, and next time remember to record at 24bit to start off with.

I can see how many would find it splitting hairs. However, some people want to know how their tools work ( surely, you have at some time in your music/audio career ) - it may be of interest, it may be important to their job... I don't see what is wrong with this digression - the OP has long left the conversation. If you're recording an album, no, this probably isn't going to be a life changing event. : ^) 16 to 24 is a quick fix and it can help preserve what is recorded, there is nothing wrong for people who are interested to discuss why it can help, in what cases it won't. Even you recommended using 24 at the end of your post, when 16 bits has been enough IMO to present and capture art, creativity and emotion. I suppose the lack of inflection could lead some to think I am being dead serious or derogatory throughout my reply. Just imagine we are friends chatting at a pub and I am throwing in my poor attempts to amuse you throughout my response. I may have a strange personality, but I don't mean any confrontation. If you don't believe me, just as iS ( one of several references ) just how lighthearted I can be. Thought I'd put that in the open since it has been a while for both of us. Some truth, some bad humor, no harm.

Best,

J

uhh.. ok. I didn't notice that I was being read as confrontational either. and, sorry but you're wrong in saying that I was trying to switch off a discussion about how the tools work, quite the contrary. I was just giving my opinion on the matter like everyone else. and my opinion was just that I don't think it would make a qualitative difference once you're in the 32bit float environment for your plug ins and mixing. you're not adding any more detail to audio already captured at 16, and the wordlength already gets massively increased to give you higher precision and better rounding once you are in the 32bit engine, so plug in and mixing precision is already in a different ballgame.

I just think you're better off getting on with it and focusing on your mix, and to make sure you actually record at 24bit from then on.

uhh.. ok. I didn't notice that I was being read as confrontational either
I wouldn't say so - I was just trying to clarify that I had no bad intents when I wrote that, because people do react differently to my humor.

J

yeah, no kidding.