Gain Staging the Art and Science

Before there were meters, there were ears. After all, distortion is what you hear when you overload a component in the signal path. The whole invention of the VU (Volume Unit) meter, which we use as a guide when setting levels, was based on how much voltage caused a distorted signal. A line was drawn and it was called 0VU. It was initally all very subjective. So to recommend that you use your ears is always going to be good advice for several reasons.

The VU meters were actually made to react slowly - because watching actual voltage meters was really too fast for the human eye - and tended to make engineers jumpy, edgy and nervous. The meter could distort while you blinked. So VU meters were built to actually fall back more slowly than the actual voltage would dictate. Simply so you could “see” it.

But now you have Peak Hold functions on most “computer renditions” of meters, so it is easy to tell when you have reached a CLIP level. But the following is true: signal can distort without clipping, and you don’t always hear distortion when the meter clips. So, again, it is good advice to always believe your ears when it comes to distortion. Your eyes - when it comes to sound - should not be the final judge (alone).

The process of gain staging is a science (although becoming a lost art, but it’s a science nonetheless). Each time the signal moves from one device to another in the audio chain, your ‘job’ is to make sure that the sending device is sending enough signal to provide the next device enough signal for good clean operation. And that when you send it on to the next device the same is true. If in a device you increase the energy of the signal (by boosting the EQ in a specific range, for example), you will need to ensure that you do not send on TOO MUCH, to the next stage. If you decrease the energy of the signal (by filtering some of the signal or compressing it) you will need to ensure that you do not send on TOO LITTLE signal to the next stage. (That’s Gain Staging).

If the signal is distorted in the first receiving device, it has no choice but to send the distorted signal on to the next device. So turning it down later in the chain does not get rid of the distortion.

Too many users only turn down the final signal thinking that “well, the final output is not clipping, I’m okay"… when nothing could be farther from the truth.

If signal is equated to water running through a series of pipes… You want to provide enough water pressure from the first device so that the next vessel in the chain has enough water (without spillage) so that it can do its thing and pass it on to the next stop. Each device should have enough waterflowing in, without spilling or wasting any so that by the time the water reaches the final out, it has been filtered, refined and is purified without any spillage on the way.

It is not always necessary to have an input meter, but hopefully there is some kind of overload or peak indicator. This can be helpful in determining what signal you are sending.

Typically the synth (source) has a nominal level setting. (You’d have to consult your devices manual for details)… but for example the Motif-series can be set for a nominal output of +0dB or +6dB at the output. How much is actually sent is a function of what signal is playing, who is playing, how much velocity is applied, how much EQ, what the filters are doing, etc, etc., etc. The nominal or normal output will be somewhere near the setting +0dB or +6dB with an average amount of PARTS playing and so forth. As you can see this is very general. If you are playing just a single sound, you are not going to be at 0dB or +6dB, you are likely to be under. Under is always better than over.

This is like the game show, “The Price is Right” you want to get close to the price but you cannot go over - go over and you are out of the game.

Input devices with Peak meters will tell you for example, that the peak light comes on when signal gets within -6dB of clipping. This is a good general indicator to an engineer just how hard they can hit this device with incoming signal. Again the goal is to give the device enough signal to do its thing without overloading it.

In the pipe there is either signal (water) or there is air (Noise). You want as much water in the signal flow as possible with a minimum of air. There is always going to be some air in the pipe (noise in the signal) but when you minimize it, it is negligible.

A blinking peak indicator is fine. A peak indicator that stays on for any significant length of time is bad, very bad. If your input device does not have a peak indicator of any kind, then you can use the input metering in your DAW.

The Cubase input meter will show you the maximum signal coming in from a device and has a PEAK HOLD function that will let you know even if you have to go to another screen or are busy playing… the highest value that the input reached will be shown (held) by the meter. If that value is in excess of 0.0db you clipped and should endeavor to do a better job of Gain Staging (means you should lower the incoming signal at the source).

Every CLIP does not necessarily result in a distorted sound (it may indeed be too fast register as distortion) but all the little clips wind up making your overall sound harsh and unpleasant. Ever hear someone’s mix and it just is not smooth, it is harsh and difficult to listen to, while other mixes on the same gear are as smooth as butter and cause no ear fatigue. All the little clips add up to an overall harshness in the sound.

So it pays to gain stage (simply put): police the output of each stage of your signal path, so that it delivers clean audio (not too little, and never too much) to the next stage in your signal path.

Signal flow is the key to audio nirvana. If you can follow the signal path, and know that it is traveling in a specific path, you can easily troubleshoot and correct your errors.

Hope that helps.