and the debate continues I guess…

Although reading the other articles on gain structure I suppose the right way to adjust gain structure correctly is determined by what you are trying to achieve whether it is low noise and maximum system volume or low distortion.

p.s. using optimal gain structure doesn’t mean you are mixing to peak, you can still mix to “0VU” what ever that is these days

I agree with Eilon here..

You can use “Optimal Gain” with plenty of headroom and
and less chance of distortion. You just have to make sure
the amps are sized properly. If Amps are sufficient, then
we can run “Unity Gain” at the console. But “Optimal”
gain throughout the system..

This strikes a balance between headroom and noise floor.

This article is factually incorrect!  There are plenty of good articles on proper gain structure.  This isn’t one of them.

When the other writes

“Hotter signal = less system headroom. Less headroom = poorer sound quality.

Therefore, it’s better to set amp inputs wide open and control the system volume at the console.”

It’s just simply not true.  If the amp produces full output at +4 input and the console or driving device’s output is +24 and you just turn down the fader on the console you are decreasing your signal to noise ratio by 20 dB!  That is HUGE.  Especially to avoid non-linearities that may or more likely do not exist in a circuit driven at half or less of it’s peak voltage.

I’m sorry but this just goes to undo the good work on gain structure education that has been done already by people, including myself, that actually understand the subject.

I’m running an inexpensive Behringer amp for my monitors and they are turned up 3/4 or so.  Are you saying that I should turn these amps all the way up and turn down my monitor sends to make the most of my gain structure?

RE: John Whiteker

Not necessarily.  This approach is primarily for the mains.  My follow-up article will cover good gain structure on the console itself, which will have much to do with how we end up setting the sensitivity on monitor amps driven from the FOH console.  It may well be that some attenuation would be in order to keep from blowing out those on stage, but that can only be determined once a good console gain structure is established.

You can find the real thing of proper gain structure in the book of Bob Katz. 6 dB headroom is not enough. +4 dBu set it for 0 VU is not a good choice if you don’t have good gear. Let’s see what should be the clip point if +4 is set it like 0 VU. How many db we need ? 20 dbFS diference betwen RMS and peak + 6 db in case of balanced-unbalanced adapt, 6 dB necesary for every solid state gear. 20+6+6=32 (when it change the class of function). So if +4 dBu is O VU the clip point should be at… yeah, that’s right = +36 dBu. So how many gear junks (amps/consoles pre’s) of these days have +36 dBu. Who ? Bull shits of Yamaha consoles will never have +36 dBu.
So keep in mind that diference betwen RMS and peak is not 10 dBFS. Maybe on the Bennasi songs, but not in live. So +4 dBu as 0 VU is not a good choice if you don’t have good gear. Allmost all gear (junks) have +24dBu so your 0 VU should be - 4 not +4.

” The input sensitivity shall be 5.27 Vrms at the 26 dB setting and 2.67 Vrms at the 32 dB setting. “

The QSC PowerLight 3 Series PL380 specifications.
The relation betwen gain-gear is the key. Good gear -> +4 dBu=0 VU. Cheap gear -> less than +4 dBu for 0 VU

I know gain structure is important.
I am not a pro, but at my church we had very
poor gain before feedback, a sound contractor fiddled with the settings, eq, etc..
One thing I did later improved our gain before feedback, I turned down the controls on the power amps by about 6db.
The system performed a lot better.
This may reduce our headroom, but it is headroom we at this point have not approached.
Yeah the room needs acoustic treatment too, they won’t spend the money! That alone would have helped.

Bob, its is a bit off-topic but I think you inadvertently fixed your problem in a way, As you might know the louder a system is the more it interacts with a space, so by lowering your output you also reduced the interaction with that space, there by reducing reverb time and peaks (hot spots) in the sound field which cause feedback,
I mean, that is the same idea behind nearfield monitoring - reduce level by monitoring close there by reducing acoustic interferences and achieve a more faithful representation of the mix.

“Today, however, headroom is often used to describe the difference between peak and RMS (root mean square) values.”

That’s completely incorrect.

The difference between peak and RMS is crest factor, and it depends mostly on the signal.

Headroom is simply the difference between the signal level and the maximum possible signal in the circuit. The maximum is set by the power-supply rails in analog or by the ADC reference voltage in digital.

Also: “When the knobs on the amp are turned down, the signal from the console has to be hotter. And this means: Hotter signal = less system headroom. Less headroom = poorer sound quality.”

Also not true. If, as you point out, the input sensitivity of a typical amplifier is about a volt, then a console mix output that lives around 0 dBVU (which leaves about 16 dB or more console output headroom) clips the amp. The options here are to a) pull down the console master faders to keep the amp from clipping, b) reduce all input trims and/or channel faders to keep from clipping the amp, or c) dial down the amp input level control.

On one one hand, the first two options degrade signal-to-noise ratio. On the other hand, option c) does not, and is used most often in optimized systems. Why? Turning down the amp input level also turns down the noise from everything preceding it in the signal chain. And, again note that a console output level of 0 dBVU still has a ton of headroom!

On the gripping hand—turning down the amp as noted in option C of course reduces output level, so the operator might just goose up the mixer output to compensate. This, then, is a case of Not Enough Rig For The Gig, meaning you need more and/or bigger amps and more speakers.

The vast majority of pro power amps do NOT have passive input attenuators. Instead, most have an active balanced input section, followed by the gain control (usually configured as an attenuator), followed by the amp’s power section, where most of the voltage gain and all of its current-handling reserve lies.

It is not wise to run a power amp at full gain unless you have an unusually weak signal going into it. For optimum signal to noise ratio, it’s smarter to reduce the amplifier gain and instead run a hotter signal through the system, while being careful to avoid clipping. Headroom is good to have, but you do not need a minimum of 6 dB at all.

My response to Andy Peters can be found here:
http://www.prosoundweb.com/article/how_and_why_unity_mixing_can_make_all_the_difference_in_the_world/
Bob,
OK, you added to my understanding of an amplifiers input stage, and if signal/noise ratio is your primary criteria for excellent audio, then by all means reduce the amp sensitivity.  But, where on the rising curve does it cross the falling curve of audio quality due to reducing the system headroom?  That one’s not so easy to pinpoint.  People complain more about poor sound than they do about a slight hiss.

That would be reducing the amps’ gain.

What exactly is the “falling curve of audio quality due to reducing the system headroom”?

Avoid clipping the mixer and other gear in the signal path; you won’t have any reduction in audio quality by having the peaks hit 0.5 dB below clipping instead of 6 dB. And you’ll have that much more signal above the noise floor.

You don’t get improvement in audio quality by running your signal levels disproportionately lower than what the gear can handle. This is especially true if there is digital gear in the chain; in fact, running input signals well below the sensitivity of the A-D converters sacrifices audio quality.

To-may-to, to-mah-to.
   
I’m sure my next claim will send you into an apoplectic fit so violent that you deposit spittle all over your monitor and send your coffee cup flying as you scramble for the keyboard.  No reduction in audio quality?  Your knowledge is in direct opposition to my experience.  I’ll go on record saying that consoles need even more than 6dB of unused headroom at the summing amp for them to sound best.  Leave the input “gain” wide open on the amplifiers and you’re forced to leave headroom in the mixer if you understand the value of unity mixing.
   
I find it interesting that the official spokesperson for QSC deems it appropriate to call into question the operational concepts of one man.  I’m also surprised that QSC doesn’t have any problems insulting a customer and owner of QSC products.

I’m not official anything. I’m just pointing out technical errors so that at least others can do things better. If people start thinking factually incorrect things, like amp gain controls are passive input attenuators, or that mixing boards need at least 6 dB of headroom, it’ll obstruct their better understanding what actually goes on in their gear, and that’s more likely to interfere with coming up with correct solutions to problems. Doing tech support for many years makes me keenly aware of how people can get derailed by incorrect information.

If a console needs at least 6 dB of headroom to sound best, the designer needs to go back to the drawing board or back to school.

BTW, I’m not apoplectic. I also don’t mean to insult anyone, I’m just pointing out common misconceptions.

My advice is to run an amp at full gain if you need to, such as when the source has a weak output capability (such as an iPod). If the source is pro equipment like a mixer, spreaker processor, etc., you’ll make better use of the headroom if you turn the gain down to better match the source’s output capability.