Clarification on Gain Staging Advice from Tom Sides

So, I was reading the November 2014 issue of Electronic Musician and they interviewed veteran engineer Tom Sides. In the interview, he said the following:

“When setting the gain on a microphone, I usually start with no compressor in line and then set the peak level at about 4db from clipping or hitting the red on the Pro Tools channel meter.” (Emphasis mine.)

That seems to fly in the face of what I’ve learned about gain staging while recording. I was under the impression that you should aim for recording peaks somewhere in the -18dBFS to -12dBFS range (and probably not over -10dbFS). At first, I thought that maybe he was talking about 4dBu, which would equate to -20dBFS, but notice the part that I italicized in that quote: he specifically referred to the Pro Tools channel meter, which are dBFS meters.

Can anybody explain why such a respected pro might have suggested what seems like bad advice?

BTW, for anyone reading this who isn’t clear on the background re: gain staging, here’s an article from SOS that breaks it all down: http://bit.ly/1t3kfxj. And here’s a page on the Independent Recording Network site that goes into further detail, complete with interactive charts that you can click on for clarification: http://bit.ly/1t0RQHq.

I may be wrong, but I believe when people talk about using -18, it’s a reference to RMS levels, not peak. For me, -18 RMS usually works out to around -6 peak.

As in any A to D conversion, the closer to Full Scale the greater the resolution (more bits used) of the resulting file.

The Caveat here is that you DO NOT want to hit 0dBFS.

So it really depends on your source material.
If you can get away with -2 peaks with no chance of clipping - Great.

But that never happens in the real world, I think.

The -18 thing is a nice compromise for resolution and headroom.

Hugh

That’s an archaic practice from the old 16-bit days where the dynamic range was limited and the dithering was shoddy at best. Nowadays, the 24-bit environment has plenty of range and a ridiculously low noise floor so aiming for a -12dBFS to -18dBFS window for your peaks will leave you plenty of headroom for plugs and summing on the mix bus. Remembering the differences between analog and digital metering will break the habit of having to have everything as close to 0 as possible: -12dBFS being equivalent to +12dBu, etc.

FYI, my original post was accidentally deleted as Steinberg tried to flush out spammers, so I reposted my question…but it eventually got rescued from oblivion. As a result, I’ve consolidated both posts at the top. Carry on.

Some things come to mind:

  1. One advantage to recording with peaks much lower (e.g., at -10-12 dBFS) is that the errant loud bit won’t be recorded with distortion. BUt Mr. Sides probably knows his set up to the nth degree, and KNOWS how to keep everything clean even with a high peak recording level. Someone riding the faders during tracking to ensure that isn’t unlikely either!

  2. Another advantage to recording at those lower peaks is that A/D conversions are said to be more precise when not near their peak inputs (I don’t pretend to understand the math involved here, but that’s what people say). But I would guess that at his level the interfaces sound just fine when driven hard, maybe even better if he’s using tubes for distortion etc. So the “sound” he wants might be absent at those lower peak recording levels.

But, these are all “maybe” explanations … maybe someone here works with him? :slight_smile: