Headroom in the final mix

S

smokeAndMirrors

New member
All the advice I've read suggests that for mastering purposes there should be at least 6dB, sometimes even 10dB, headroom in the track being mastered. As a relative novice, this puzzles me. It generally means that I am cranking up the threshold on the loudness maximizer stage by a large amount more than a small amount. Also, the computer scientist in me says "less volume = fewer bits = reduced fidelity" so the loudness maximizer is basically adding noise as well as level. Why not leave less headroom so less interpolation is required?
 
Even if there is 0db headroom, the mastering engineer can reduce the gain on the track input to bit shift it all down at his end. Whether s/he or you does that, it makes no difference. The fidelity lost is probably not noticeable. The numbers equate to spl and frequencies of sound come from the rate those numbers go negative and back again, the fidelity lost equates to small differences in the volume of those frequencies and it will be lost for all frequencies in the same proportion.

I think the point is to not compress everything so much that there's no dynamic range left in your tracks for the mastering engineer to play with.

Edit - obviously I'm talking about digital only and am guessing and assuming left, right and centre.
 
(Geez, why does everyone always automatically assume "volume" is the center of everything?)

[/digression]

But while we're on it...

"Volume" is the last thing the mastering engineer (well, an experienced, practical, "real" for lack of a better term type) is worried about. A monkey with a limiter can do "loud" and there are a million different ways to get there.

An awful lot can happen before that stage. Everything else happens before that stage. ANALOG processing may very well happen before that stage (where the chain may be calibrated for signals far beyond digital headroom).

[/Digression again - Sorry]

ADMITTEDLY: Some of these numbers I have no idea. *SOME* headroom. Clear, clean, non-limited or excessively compressed headroom. I personally don't care whether it's 1dB, 6, 10 or 20. Once you're beyond 20dB you're in the "meh, that could be hotter" range. In 24 bit, you can be PEAKING at -48dBFS and still have the essential resolution of a compact disc. So the whole "losing bits" thing really doesn't fly unless you're working on recordings of ants and spiders talking to each other while cannons are going off (and of course, using gear that can take advantage of such a dynamic range).

So - SOME headroom. How much doesn't make a difference. It's far more important to have all that headroom when tracking anyway. But it's also important to keep some of that headroom through every possible stage.
 
it just plain sounds CLEANER to me,
when there is at least 6-8db of headroom.

i like RMS of anywhere between -22 to -16, depending on the material..

but peaks no higher than -6

i can't quantify this in any way other than to say that, to my ears, it sounds cleaner.

it is easy enough to get back to hitting the ceiling,
so there is really no down side to having plenty of headroom.


once you mix this way,
and get used to the sound of somewhat more OPEN dynamics,
i can almost guarantee you will prefer it,
to the sound of tight and intense compression and limiting.



you almost have to go back to audiophile recordings from before the loudness wars started,
to RE-Calibrate your ears


Crosby, Stills & Nash, Daylight Again

Eric Johnson, Ah Via Musicom

Fleetwood Mac, Fleetwood Mac and Rumours

The Allman Brothers Band at Fillmore East

Steely Dan, Gaucho

R.E.M.: AUTOMATIC FOR THE PEOPLE



but there are some more recent ones too...

check out the 'honor roll':

Honor Roll - Digital Domain: CD Mastering | Mastered for iTunes | Audio Mastering | Blu-Ray Mastering
 
smokeAndMirrors said:
All the advice I've read suggests that for mastering purposes there should be at least 6dB, sometimes even 10dB, headroom in the track being mastered. As a relative novice, this puzzles me. It generally means that I am cranking up the threshold on the loudness maximizer stage by a large amount more than a small amount.
Click to expand...
Reads like you're doing the finalizing (AKA 'mastering'). In that case you'd be the one bringing it up to whatever final level.
The idea is keep reasonable amounts of headroom during the mix process -and for the most part leave it there if you're handing it off to someone else to master.
Otherwise simple adjustments on the pre-gain of your master bus, up into your final compressor and limiter.

Also, the computer scientist in me says "less volume = fewer bits = reduced fidelity" so the loudness maximizer is basically adding noise as well as level. Why not leave less headroom so less interpolation is required?
Click to expand...
No, the system works as designed - I.e, at normal levels. The room above is the headroom so the whole process upstream can function cleanly and correctly.
Another way to look at it is you pick -with some forethought and understanding.. when and where you begin to 'push it'.
 
smokeAndMirrors said:
Also, the computer scientist in me says "less volume = fewer bits = reduced fidelity" so the loudness maximizer is basically adding noise as well as level. Why not leave less headroom so less interpolation is required?
Click to expand...

The computer scientist in you is operating on false information. More bits does not equal more resolution, it means lower noise floor (and as Massive points out it's extremely low). Each bit represents a doubling of voltage (I think, ashcat, ecc, someone please correct if wrong) which equals 6.02dB.
 
bouldersoundguy said:
The computer scientist in you is operating on false information. More bits does not equal more resolution, it means lower noise floor (and as Massive points out it's extremely low). Each bit represents a doubling of voltage (I think, ashcat, ecc, someone please correct if wrong) which equals 6.02dB.
Click to expand...
I think it's more a halving of voltage (-6db). 0dbfs is a given voltage, and each bit allows you to split that up (...errr, down???) finer and finer. Like Massive said, 24bits is huge dynamic range, and while the OP's concern is real, it's a bit overstated.
 
You must log in or register to reply here.
Back
Top