What is Gain Staging?

The term Gain Staging refers to setting the levels between stages of an audio system.  The real root of the term lies before digital recording and the ubiquitous computer based Digital Audio Workstation, but the concept is still applicable today.  Afterall, your audio has to be analog at some point, either on the way in from your instruments, microphones, mixers and interfaces, or on the way out to your speakers.

The whole point of gain staging is to optimize the signal levels to minimize noise and distortion.  In order to see what that means, you’ll need to start of with a key definition.

Signal to Noise Ratio

Don’t overthink this one, it is exactly as it says, the ratio of the signal you’re actually trying to produce compared to the noise present in the signal.

All electronics produce noise, it’s a simple fact of life.  The sources of the noise are all over the place, the typical contributors can be broken down like this, in roughly descending order of contribution:

  • Ambient Noise –  This happens only during microphone recording and is the background noise in your room while recording.  Fans, your computer, your family or roommates, the furnace, nearby traffic… any of these will add noise to your recording.  The only way to reduce this is to make your room quieter through soundproofing or otherwise eliminating these sources.
  • Electromagnetic Interference –  Your house, and pretty much every building supplied with power is full of 60Hz electromagnetic noise, and countless sources of much higher frequency signals from WiFi, cordless phones, cellular signals, whatever.  This is invariablly picked up on wires and within circuit boards.  This can be minimized through good design of your equipment, like grounded metal enclosures.  Cables are major culprits here, the longer the cable the more likely it will pick up noise.  Balanced cables by design eliminate that interference, use them everywhere you can.  If you have to use unbalanced cables, make sure they’re good quality and as short as possible.
  • Internal Equipment Noise –  All electronics will have some amounts of noise generated internally.  There are two main sources of this, the power supply design itself, and thermal noise from resistors.  The power supply is supposed to take the noisy 60Hz power from the wall and turn it into pristine DC voltage for the electronics to actully use.  Typically, these are designed pretty well, but a failing part or a cheapo chinese knock off can cause problems.  The thermal noise in resistors shows up as broadband white noise and is generated through some mystery of physics.  The details are beyond my level of education, but it’s a fact of life.  The higher the value of the resistor, the more thermal noise.  So a good electronic design will use the smallest resistors possible throughout.

Noise in a system is typically of a pretty constant amplitude once setup and powered on.  You can take the steps above to minimize it, but once it’s there the only thing you can control in the Signal to Noise ratio is the signal level itself.

The signal is what you actually want to send through and eventually hear.  Making this signal larger relative to the constant level noise = better signal to noise ratio.  This works up to a point, because all electronics also have a finite, limited signal level they can handle.  When you exceed this level, what happens is the tops and bottoms of the signal simply get cut off, or clipped.  That results in nasty distortion of your desired signal and should be avoided.  For your typical amplifier, this is determined by the power supply voltage, and the design.  In order to maximize your signal to noise ratio, all you have to do is pass through the maximum signal level possible for the amplifier.

Proper Gain Staging

So proper gain staging really comes down to maximizing your signal to noise ratio for each component in the signal chain.  The signal level should be set so the absolute largest peak hits the absolute highest signal level capable for each device in the chain.  There’s one thing mucking up that simple task though, dynamics.  The trick there becomes figuring out exactly what that largest possible signal will be.  Singers can be all over the place, based on how far they are away from the mic and how confidently they’re singing.  Percussion is terrible for consistency and can produce huge variations in signal level.  That means the best you can do is plan for the loudest signal you can test, and bring it down slightly in case that’s not really the loudest when the tape is rolling or the show is on.

If order to make it simple, your components in the chain will need a few things:

  • Input level adjustment
  • Signal meters
  • Output level adjustment

Lacking any of those and you’ll have a harder time pulling this off, though it can still be done with reasonable guesses.

Putting it in Practice

Let’s create a hypothetical chain of components:

Microphone -> Pre-Amp -> Compressor -> Equalizer -> Sound Card

In order to set levels, you’ll need to start at the left, the first part of the chain.  Obviously a microphone doesn’t have a level to adjust, so you’ll simply need to give it the loudest sound you think you’ll have happen.

The pre-amp will then be adjusted to with that loud sound, it’s output is near maximum.  Your typical mic pre will have a gain adjustment, at least a peak clip indicator, and an output level.  Turn the gain up until the clip light blinks, then back it back down a tad.  Crank the output level up as high as you can, depending on the next piece, our compressor.

Compressors may not have an input level but should have some metering to show you your levels.  The mic pre should be feeding it with as much signal level as it can handle.  You should only adjust the compressor settings after you have the signal level set for the input.  Then you’ll tweak the output level as high as you can for the next component, the equalizer.

EQ’s vary a lot so this part will depend on what your EQ has to offer.  Hopefully you’ve got some sort of peak or clip indicator.  Remember, boosting with the EQ adds gain so if you crank the level up so it’s on the edge of clipping when the EQ is flat, you won’t be able to boost at all   Turn it down a bit to allow you some room to boost what you need.  If you’ve got an output level adjustment crank it as high as you can for the next component, the sound card.

For sound cards, you’ll need to watch the input level in your software of choice.  Arm a channel in your DAW software, or use your sondcard’s control panel to monitor the input level and crank it up just shy of the peak 0dBFS level.  Then, turn it down a little more.  Not much sounds worse than a clipping analog to digital converter so leave yourself some room.

Now your set!  But if anything changes anywhere in the signal chain, all components downstream may need re-adjusted to keep your gain optimized.