Gain Staging (i.e. +6 db = 0?)
 

Sorry to post a question that seems so elementary but I think I'm losing it.

I've always been careful to set my audio channels so they do not peak above zero anywhere in the channel (including internal clipping within the plug-ins) and to set the channel levels so that the bus never peaked above zero even without having to attenuate the bus with a master fader. My friend claims that with 6 or 12 db of headroom, it is better to peak above zero and approach the true zero which is +6 or +12 db (depending on your protools setting). Even when he bounces to disc, he doesn't bother reducing the master fader to peak below zero.

I feel that my approach is much more sophisticated, yet, I can't hear any clipping in his mixes and they almost sound better (richer) at those high gains than they do after I've bothered to "fix" his gain staging.

What do you all target for:
1. Peak level on your master bus (ie for bouncing)?
2. Peak level on aux channels?
3. Peak level inside an audio track (ie at the input of the plug-ins)? I have horrible noises using wave q10 eq even without any red clip alerts when there is an eq gain and the overall level is attenuated on the output fader.
4. Is is better to keep channel output low or to use a master fader on your buses.

Thanks!

Re: Gain Staging (i.e. +6 db = 0?)
 

-6bBFS as a max throughout.

The reason for this is ‘Intersample Peaks’ this is a truly massive subject so I suggest you google it, for more detailed information…

But in a nutshell treating digital as analogue and recording into it as ‘hot as possible’ back in the 16 bit days was a major contributor to why early digital recordings (and some current miss informed ones) suck … (together with poor a/d converters)

Now in this 24bit world you would have to be insane to record higher than -6dBfS, it gives the headroom to allow the wavform to not generate a distorted square wave when the sequential samples that are close to 0dBFS are joined together with the Nyquist curve and go above 0dBFS.

This clipping and distortion will not show up on traditional PT metering and will only show up on an ‘Oversampling Meter’ like the TL Master Meter.

And don’t worry about losing dynamic range, you would have to lower the peak level to -48dBFS to give you what would effectively be a 16 bit recording… So there is plenty of room there before you go anywhere near the noise floor…

Cheers



N

Re: Gain Staging (i.e. +6 db = 0?)
 

Thanks. That is helpful. The internet is full of discussions on this such as:
http://recforums.prosoundweb.com/ind.../t/15038/6263/
and the protools mix bus is nicely explained by digidesign here:
http://akmedia.digidesign.com/suppor...ixer_26688.pdf
And it seems the consensus is to keep levels low because there is room to spare.
And it seems that while the mixer has 12 db headroom (due to 48bit internal precision), the bottlenecks would occur when a) bouncing to disc, b) outputting to a DAC, or c) feeding the signal into a plug-in. These are all instances where the 12db headroom is removed, if my understanding is correct. I guess I was just surprised that I can not hear the problems with a hot mix, even peaking at +3 on the master fader - perhaps my ears are what's lacking.

Thanks!