Bouncing audio for other tracks - 16 or 24 bit

[rainstick]

hi - I'm just pilfering various bit's and bob's from some track's I've recently made and was wondering whether I'd be best off bouncing them as 16 or 24 bit audio files, ? There for use in future projects, probably put them into sampler instruments...

just wondered what peeps thought was best / usually did

cheers

[David]

24 bit files have more dynamic range and take about 1.5x the space/streaming resources vs the 16 bit file. It all depends on the desired dynamic range and the desired efficiency on HD resources.

In many cases 16 bits should be enough.

[Rev. Juda$ Sleaze]

I'd go with 24bit, it will give you more options in future projects.

If HDD space is a serious concern, and dynamic range isn't, then maybe go for 16bit.

[rainstick]

cheers guy's - i didn't realize the dynamic range particularly suffered as a result of this, ? Ill have to do a few test's perhaps!

I guess my main concern was loss in quality - a lot of the file's were probably 16 bit anyway but i only recently realized how much it effects the file to bounce to 320mp3, then bring back into logic, then bounce back out again.... and this felt like a (slightly) similar process.

and i wanted to double check there'd be no compatibility issues in using one versus the other....

cheers

[Eriksimon]

16 bit for simple (drum) sounds, 24 bit for any tonal stuff. If you discover that 24 bits somehow taxes your CPU too much (when used in an EXS24 instrument), you can always copy/convert them to 16 bit. Or if you find it takes too much disk space. Otherwise, no reason to downsample.

[rainstick]

cheers all - I'm probably going to go with 16 bit.... Maybe 24 for anything particularly complex or (like you say erik) tonal....

Just wanted to make sure there were no particular horror story's or rule's of thumb as I'm going to be doing this for a few track's

thanks

[Eriksimon]

Just make sure to have healthy levels (peaks at ± -3dB) when bouncing to 16, and you should be fine.

[rainstick]

yeh it's pretty low as the gain i usually have over the stereo out (which is on +13) is muted..... Ill bounce them all down, then bring them into a fresh project, normalize them and convert them to exs instruments....

* edit - i should probably just normalize when i bounce shouldn't I, ? I'm s#!+ !

[triplets]

yeh it's pretty low as the gain i usually have over the stereo out (which is on +13) is muted..... Ill bounce them all down, then bring them into a fresh project, normalize them and convert them to exs instruments....

* edit - i should probably just normalize when i bounce shouldn't I, ? I'm s#!+ !

What is on +13??

[David]

16 bit for simple (drum) sounds, 24 bit for any tonal stuff.

I'm sorry but that statement does not make any sense. Bit depth has no relation whatsoever to the type of sound (drums vs tonal). It only affects the dynamic range, which is the dB range between the noise threshold and the clipping level (0 dB FS). 16 bit approximately gives you 96 dB of dynamic range, while 24 bit gives you about 144 dB.

96 dB of dynamic range should be enough for most music production applications where you can liberally get close to the clipping level - I somehow doubt that any samples, tonal or not, have more than 96 dB of effective dynamic (dB range between the softest musical sound and loudest musical sound).

However maybe your samples need overhead. For example an orchestra sample library, where the pianissimo samples are actually recorded much lower than the fortissimo ones. Let's say, for the sake of argumentation, that they were recorded MUCH, MUCH lower. Let's say 70 dB lower. Now if you have a 16 bit sample library then that means the pianissimo samples are having an effective dynamic range of 96 - 70 = 26 dB. Not much. If the listener raises their volume in the pianissimo section, they'll raise the noise floor. In that case it can make sense to use 24 bit samples.

[Eriksimon]

16 bit for simple (drum) sounds, 24 bit for any tonal stuff.

I'm sorry but that statement does not make any sense. Bit depth has no relation whatsoever to the type of sound (drums vs tonal). It only affects the dynamic range, which is the dB range between the noise threshold and the clipping level (0 dB FS). 16 bit approximately gives you 96 dB of dynamic range, while 24 bit gives you about 144 dB.

96 dB of dynamic range should be enough for most music production applications where you can liberally get close to the clipping level - I somehow doubt that any samples, tonal or not, have more than 96 dB of effective dynamic (dB range between the softest musical sound and loudest musical sound).

However maybe your samples need overhead. For example an orchestra sample library, where the pianissimo samples are actually recorded much lower than the fortissimo ones. Let's say, for the sake of argumentation, that they were recorded MUCH, MUCH lower. Let's say 70 dB lower. Now if you have a 16 bit sample library then that means the pianissimo samples are having an effective dynamic range of 96 - 70 = 26 dB. Not much. If the listener raises their volume in the pianissimo section, they'll raise the noise floor. In that case it can make sense to use 24 bit samples.

Well, I am not going to argue with you on anything; but that's only because you are right.

"Does not make any sense" almost sounds like you are trying to pick a fight though. I don't claim that it makes sense, it is just what I do.

The OP asked what we "usually did".

What I was trying to say, was that my personal habit is to save anything at 24 bit, but especially complex and/or very delicate sounds (harp, piano, acoustic guitar, any recorded performance from somebody else etc). just as it has always been my personal habit to save most single drum sounds (i have always had harddiskfilling amounts of these - I'm a drumsamples hoarder) as 16 bit. having said that, if levels are healthy I personally find it hard/impossible to hear a difference between certain 16 and 24 bit samples. But if I can't hear it, doesn't mean it's not there...

[David]

"Does not make any sense" almost sounds like you are trying to pick a fight though.

Sorry Erik, that certainly wasn't my goal. I just wanted to point out that the only difference between 16 and 24 bit is the dynamic range. Use them accordingly: look at the dynamic range you need - not whether or not the sounds are "delicate" or "complex" - which should not enter the equation here.

[Matt Mayfield]

I thought this would be relevant to the discussion - a video about the dynamic range of various formats:

http://www.youtube.com/watch?v=GJlovF1AEZg

[rainstick]

nice one matt! - was interesting....

Theoretically now (not really a problem for me) but what are digital synths producing sound's at, ? It's 32 bit's isn't it... ? Or is that relating to something separate, ?