1. What is the optimal gain structure for signals? Is 0dbu the lowest noise of the amp ideally or does it matter much? For example, if I have a hot signal coming in to a mixer due to some hard to control issues is it better to try to lower several of the gains before the signal to reduce it coming in or is it ok just to turn down the mic/level input gain as much as needed? (there seems to be a loss of fine control when this is done but is it ok as far as sound quality assuming nothing is distorting?)

2. When a peak led lights up how serious is it? Obviously just a peak here and there is not life threatening BUT is it optimal to try and get a signal in that has has plenty of headroom so the "clipping" doesn't occur? (which will give less distortion and overall a cleaner mix)

3. How does one go about setting up the gain structures when there are a huge number in the signal path? Each one can reek havoc in many ways(ruin ears, create distortion, unbalance the mix, reduce sensitivity, etc....). Is there some theoretically based method that the pro's go by that lead to the quickest and best sound?

I know people that just sort start turning things up or go by previous settings and it seems wrong on many levels(ultimately just unsafe in case you have a real hot signal coming in and not knowing it).

Just looking for a few tips on these things...

Gain structure is about optimisng the average signal level so that it is comfortably positioned sufficiently above the noise floor that noise isn't a problem, and well below the clipping level so that transients aren't damaged.

In analogue equipment the 'ideal' average level is the notional line level which is typically either 0dBu or +4dBu in pro gear, and -10dBV in semi-pro gear. If the equipment uses VU meters, 0VU is the notional line level.

In general, professional equipment will have a noise floor around 90dB below the nominal line level, and won't clip until 20 or 24dB above -- giving a total dynanmic range of around 115dB.

If you look at the specs of a typical digital converter -- even budget models -- you'll see that they are specified with a dynamic range of between 110 and 120dB, with most being about 15dB these days... exactly the same as a good professional analogue system. Consequently, it makes same to use the same operating level concepts, allowing a good 20dB headroom avove the nominal operating level.

In other words, ananlogue 0VU should equate to about -20dBFS.

Analogue VU meters don't reveal transient peak levels well at all, whereas digital sample meters are far more indicative. As a result, expect to see transient peaks kicking up to about -10dBFS, with the average level around -20dBFS.

There is no reason to get the peak over lights flashing on digital equipment. Doing so introduces anharmonic distortion which isn't nice. There are cases when some analogue overload distortion can be musically creative, but that's harmonic distortion and a completely different thing!

You can control the input level to bring it up or down to the nominal line level (0VU) by using te gain input control and/or input pad first and foremost. Use the channel PFL (if available) to check the signal level through the channel electronics and adjust the input gain to achieve a sensible level.

Ideally, the channel faders should all end up around the unity 0dB mark with the mix roughly right. When that is achieved, you have optimised the signal-noise and headroom for each source channel.

If you find you're having to pull the input faders down a lot on analogue desks, or the output faders down on digital ones, you've got the gain structure wrong.

Hugh

--------------------
Technical Editor, Sound On Sound

There is an essay in what you have asked so a short response each directed to your points:

1) Use input gain controls (including pads if needed on mics and preamps/mixers) to back off the signal level, you can get situations where this is not enough and you overload the first electronic components at the input stage. If you can back off the source signal a bit it would help. Look up 'PFL'ing' it's what pro engineers understand when they record and mix. (note I did not say 'do') cause once you know you know it by heart after a while if you work the same gear and you can see it on master output metering as well as you get better and more familiar.

2) Not all LED peak lights are the same, analogue? digital? for starters. Also many mixer/preamp LED's light up 4dB before distortion occurs, you need to know what the peak light actually means on every bit of kit you use.

3) Every signal gets PFL'd individually, think of them as isolated sources at first, you are not mixing you are optimizing your gain structure pre fader for each and every signal being recorded/mixed. Then you can think about your monitor mix and balances. 1 by 1, step by step is the way to go. Optimizing gain structure is fundamental and first because everything coming after relies on it, headphones mixes, monitor mix, recordings, the whole kaboodle.

TOP TIP, when recording always peak 5dB lower than the level the musicians give you cause it almost always goes upwards in the actual take.

Have a read of this too:

Gain Structure

cheers

SafeandSound Mastering

Hi
Thanks very much for your explanation on ‘Gain structure’…
Most of us,coming from the analog world,have always assumed that close to ‘0dB...’ was ’the' rule.
Manufacturers still indicate to record(or use gear) close to this mark,in most of their manuals.They just forget to tell us about the difference between ‘analog’(0dBvu) and ’24 bits digital’(0dBFS)…
Thank you for making that clear.

That doesn’t mean that,recording(in digital world) close to ‘0dBFS’(but without clipping) is wrong,just that we are using a reduced headroom range for mixing.I’m I right?

Thanks again!
Cheers.

*
Sorry,0dBVu

It's actually just 0VU...

No, there's nothing wrong technically with peaking close to 0dBFS. In fact it maximises the signal-noise ratio. However, it does make life difficult when tracking and mixing because of the risk of clipping, and most analogue equipment struggles to deal with such high average levels.

The bottom line is that here is absolutely no practical benefit or advantage, and everything works better when yo leave a sensible headroom margin.

Hugh

--------------------
Technical Editor, Sound On Sound

Quote Hugh Robjohns:

---

.. everything works better when yo leave a sensible headroom margin.

Hugh

---

Don Corleone or the Pope could not have said it better. Whatever your paying this guy it's not enough.

Thank you Hugh!
‘0VU’ got it.

Cheers!

Thanks guys, it helps a little. I see the big issue in mixing live or studio is trying to maximize both gain and headroom... These are competing variables and there is an optimal point(ideally). I'm still a bit confused on actually how to go about but I will make a different thread.