Setting gain levels in 16 bit recording.

  • Thread starter Thread starter ClapHands
  • Start date Start date
C

ClapHands

New member
Considering there's less space to deal with in a 16 bit recording, how high should you be able to let the signal peak?
 
well I would have usually tried to get the signal as close to clipping as possible w/o the actually clipping. but it definitely also is important to record the instrument as loud as u want it in the mix, that way you don't need risking it clipping.
 
It's the same general considerations as in 24bit in that you leave some head room on the converter to where you don't have to worry about peaks hitting full scale. In almost all cases you room noise, analog hiss etc is still higher than the conversion error noise.
 
I would respectfully disagree, because w/ 24bit it does not matter how much headroom u leave, but with 16bit you have to make sure to be as hot as possible, so that you don't have a problem when i.e. over-compressing.
 
If you're over compressing' I would maintain you would then have other more troublesome things to worry about.
Now. When you compress a signal where the overriding noise is the original room and or analog noise, does that relationship change? No.
Recording a nats with magnums?
 
aimpark said:
with 16bit you have to make sure to be as hot as possible
Click to expand...
This is a myth. With 16-bit you have 90dB of room to work with. That's a good 20-30dB over the dynamic range of the typical analog signal coming in. There's more than enough room to do anything you need to do without pushing the heat. And compressing narrows the signal's dynamic range, actually making it easier to fit, not harder.

G.
 
SouthSIDE Glen said:
This is a myth. With 16-bit you have 90dB of room to work with. That's a good 20-30dB over the dynamic range of the typical analog signal coming in. There's more than enough room to do anything you need to do without pushing the heat. And compressing narrows the signal's dynamic range, actually making it easier to fit, not harder.

G.
Click to expand...

Learn something new every day. ;)

I was under the same assumption...
Analog reel to reel?...saturate it.
16 bit? Push it but no red.
24 bit? don't push it. Let it breathe.

I dig this place. :)
 
SouthSIDE Glen said:
... And compressing narrows the signal's dynamic range, actually making it easier to fit, not harder.

G.
Click to expand...
I'm presuming Aimpark meant compression after the fact that would bring your noise floor up.
But all those ADAT, DAT years, I just don't recall this being a huge problem. My analog mix and effect paths were a bigger source of noise IIRC. Use reasonable care, move on. :)

There's a thread floating around 'What's holding your recordings back', some such- From what I see, here, this isn't it. :D
 
mixsit said:
I'm presuming Aimpark meant compression after the fact that would bring your noise floor up.
But all those ADAT, DAT years, I just don't recall this being a huge problem. My analog mix and effect paths were a bigger source of noise IIRC. Use reasonable care, move on. :)
Click to expand...
What many folks don't remember is that it's not the compression that brings the noise floor up, it's the cranking of the make-up gain *after* the compression. And you're right, that's going to bring up the analog-generated noise, not the digital noise floor, because the analog noise is always going to be above the digital limit (unless one digitally records their levels way too low; but in my years here I have yet to hear that being a problem ;) :D). And that's going to be equally true at 16bit as it is at 24bit.

Snoop Dogbreath, look at it like this. If you are very, very lucky and the stars are all aligned, and your mother-in-law is in Europe, you might just have an analog signal with a dynamic range of 70dB from peak to noise level.

In digital, you get 6 dB of dynamic range per bit of word length. 16bits x 6 dB/bit = 96dB. Subtract 6dB because the last bit itself will not be completely accurate (i.e. there will be digital noise down there), and that leaves you with 90dB of digital room above the digital noise floor.

With that in mind, even if you recorded your 70dB analog signal to peak at -10dBFS, you still have the analog noise floor coming in 10dB above the digital floor.

Now, granted, there *is* substantially less digital room in 16bit than it 24bit (90dB vs. 138dB) so it's true that one cannot afford to be a slack about recording levels in 16bit, but as I said, when was the last time you seen or heard of anybody having a problem with recording their levels too low? ;)

G.
 
With SSG on this (as if anyone would be surprised otherwise).

Levels aren't decided via bit-depth. While you can't really afford tracking at -40dBFS like you could in 24-bit, you certainly don't need to be anywhere near full-scale. "Somewhere around -10dBFS" is more than high enough while risking less distortion from the front end.

I would respectfully disagree, because w/ 24bit it does not matter how much headroom u leave, but with 16bit you have to make sure to be as hot as possible, so that you don't have a problem when i.e. over-compressing.
Click to expand...
Whuh?
 
So should I be more concerned with the gain setting in the multi-tracker than in what I do on the computer?
 
You should be most concerned with the levels set at the preamp itself. The first gain stage is the one that counts. Once it's digitized, all the rules of analog go out the window. But you only get a chance at that analog headroom once - If you blow it there, you've blown it for every step afterwards.

More here: http://www.massivemastering.com/blog/index_files/Proper_Audio_Recording_Levels.php

While we're on the subject - Is there any particular reason you're recording in 16-bit...?
 
I can only think of one time in my life where 16 bits was not eough for the content I was recording, and that's when I was recording different types of guns and then having to record a conversation without clipping the gun and without adjusting the preamp in between takes. Surprisingly, I could make out the conversation, though it was soaked pretty heavily in quantization noise.

With music, you really should treat 16 bit much if any different than you would treat 24 bit tracking. As Glen mentioned (look glen, I'm not taking an opposite stance) your analog noise is likely going to be higher than the quantization noise whether you are recording in 16 or 24 bit.
 
bozmillar said:
(look glen, I'm not taking an opposite stance)
Click to expand...
LOL, awwww, you're no fun anymore! ;) :D (J/K!!!)

What can I say, boz, you're a running clock and I'm a broken, stopped one. Even we will agree twice a day :D.

G.
 
You must log in or register to reply here.
Back
Top