Short answer: A godawful racket where all frequencies are represented equally. The goal is to get the exact same racket reaching your ears.

--------------

Long answer: Pink noise is audio signal with equal power per 1/3 octave. (White noise has equal energy per *frequency.*) So when you "pink" a room, you are sending every frequency through the speakers at the same level. (It isn't, really, but that's close enough for now.)

In reality, the system a) probably will not be able to produce it exactly that way and b) the room has audio characteristics of its own that alter the sound after it comes out of the system. By changing the EQ you are sending through the speakers, you seeking to reproduce that flat pink noise level.

Why? Well, every room is different (and slightly different on every day based on numerous factors), so you never know what you are going to get. For example, let's say you are in a room with hollow walls. This is common in older buildings where a performance space has been retrofitted by building out the walls and running cabling behind them, etc. The floor might also have been built up for a rake, so the audience may be sitting above a hollow floor as well. Chances are, the room is going to have a real problem with something around 300Hz. You're gonna have that frequency runing back and forth accross the room and getting in all your mics and it's gonna feedback something vicious. When you pink, you watch your spectrum analyzer and you find out that particular problem and pull it way down out of the EQ, so you are less likely to have problems while the performance is happening.

It won't be perfect, because different parts of the room will respond differently, but it will get you in the ballpark.

If you want to pink your system without the room, you can buy a pink noise CD and compensate for what your PA actually produces. It's a start; later you'll make changes for the room you're in. White noise can do in a pinch, but it's not the best choice for a live setup.

The latest studio design vogue is moving away from rooms with a flat frequency response in favor of character, but for a live setup, it's the most predictable and manageable. In live sound, predictable is oh so good, 'cause you don't get a second take!

For samples, go here and click on the different "noise components" to the left.