Yes, Virginia, System Gain Structure Matters - Here’s Why
There should always be headroom left unused in a circuit, ideally never less than 6 dB. The closer a circuit runs to its maximum, the greater the non-linearities (distortion)
+- Print Email Share Comments (20) RSS RSS

It’s time for me to come out of the closet: I’m a dyed-in-the-wool unity gain structure freak.  There, I said it and I sure do feel better. Catharsis is a good thing. 

Now, for those of you who aren’t sure what the heck I’m talking about, I’ll explain, and maybe get a few of you over to the dark side with me.

First some clarification. Gain in sound parlance is the act of adding to (positive), subtracting from (negative), or leaving unchanged (unity) the amplitude of an audio signal. 

The first gain stage in the typical signal path is the channel preamp, followed by the channel fader, bus master, and main master. This is then followed by all the gear that’s patched between the console and the amps. 

The method by which you set the many gain stages in the signal path after the preamp is known as gain structure.

There are two prevailing methods of establishing the gain structure of a sound system - unity and optimal. (There are also sound systems that don’t follow any established gain structure whatsoever.) 

Unity structure means that the gain for every piece of gear in the signal chain is set at a 1/1 ratio.  That is, the level of the signal going out is the same as that coming in. 

Optimal gain structure means setting the clipping point of every piece of gear in the signal chain at the same level. The outcome of this is that any one particular piece may have positive, unity or negative gain, because the amount of headroom in individual components varies by manufacturer and product tier.

The only obvious practical result in a gain structure comparison is that optimal gain structure yields an overall lower apparent noise floor in a sound system. But what’s not so readily apparent is that optimal gain structure reduces the available headroom, which is a big no-no.

Headroom is an oft-misunderstood concept. Back in the day, headroom meant the difference between the level at which a circuit would clip and the current peak level of the signal coming from that circuit.  If a circuit clipped at +22 dBu, and the signal was currently peaking at +10 dBu, then there was 12 dB of headroom. 


Comments (20) Most recent displayed first | All comments in chronological order
Posted by Bob Lee (QSC)  on  04/30/09  at  12:49 PM
James, if you want to offer a certain technique you've learned and cite that it makes some sense procedurally or logistically or something, go for it. If you want to cite some technical justification, though, please research it first. Don't rely on assumptions or long-held beliefs that may actually be myth. There are many technical resources out there, more numerous and accessible than back when I first started writing pro audio articles. Use them and present real, useful information, and you can produce good articles. Dismiss the cumulative knowledge that many in the pro audio community have gathered and spread misinformation instead, and you'll have knowledgeable people taking issue with what you say. Don't make that out to be a personality conflict.

I'm not going to give a wink and a nod to erroneous information just because you're a QSC user. There are many other QSC users out there besides yourself--and non-users as well--who deserve factual information.

Posted by Andy Peters  on  04/29/09  at  11:13 PM
Here I go, ignoring my own advice.

James --

a) Please provide links to examples of Bob's impositions. I've been reading his articles and forum posts for as long as Doug (!) and he's always restrained and well-informed. One suspects your animosity results from him calling you out. You do yourself no favors by continuing to post inciteful instead of insightful comments.

b) How old was the particular QSC amp from which that "long forgotten input board" came? One suspects it's pretty ancient. A lot of old amps with single-ended inputs did have a simple pot at the front end (and I'm sure Pat Quilter would rather forget such designs), but put simply, all modern amplifiers with balanced inputs have a differential input stage followed by the attenuator. Perhaps you might wish to tell us why this is the case.

Anyways -- the door of the LAB is always open, if you wish to discuss this, and any other, live-sound topic.

Good luck.

Posted by James Cadwallader  on  04/29/09  at  08:28 PM
The least deserving of all? I've seen Mr. Lee impose himself upon others by wading into places uninvited wielding his imperialistic objectivism like a club more than once.

Oh, by the way, just this weekend I was digging through a box of old parts looking for some supplies for a project I was working on. I came across a long forgotten input board taken from a QSC amp. Want to know the funny thing? The traces from the input jacks went directly to a pair of pots with nary an active device in the path. Is that not a passive attenuator? I then cut the pot section of the board off for use in my project.

Posted by Andy Peters  on  04/29/09  at  05:01 PM
Hey, Doug,

You know better than to feed a troll ;)

-a

Posted by Doug Fowler  on  04/29/09  at  04:44 PM
>>>>>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.

Sir, Bob Lee has been providing world class customer support via the web since before PSW existed, and since before the LAB existed, way back when a group of us met on the CompuServe LSMAG forum.

Your comment is so uninformed it is laughable. I mean really, of all the people you could make an uninformed jab at, you picked the least deserving of all.

It looks like there is a pattern here. You post highly questionable 'information', people who have around a loooooooong time take you to task for it, and you push back. Word to the wise - know who you're talking to. You have not bothered to do that.


+ View all comments on this article

Commenting is not available in this weblog entry.