Proper Gain Staging in a Digital World
Lets Cram Some More 1’s in Those 0’s
Gain is the measure of the ability for a circuit to increase the amplitude of an audio signal. Gain staging is the process of optimizing the level of your sound signal to be recorded, broadcasted, or amplified into a PA system. This is important to get the full dynamic range of your audio signal while avoiding the noise floor.
Warm and Fuzzy: Analog Distortion
In the world of analog recording audio is stored onto magnetic tape by arranging magnetically charged ferris particles onto a strip of mylar. Analog recordings were a bit more forgiving of over modulated levels as the end result was a harmonically saturated distortion that is interpreted psychoacousticly as warmth. For this reason recording to two inch analog tape remains a coveted boutique media option in commercial recording facilities. Analog equipment supports an extended headroom meaning that you have the ability to drive a channel “Hotter” before clipping than the digital equivalent.
Cold and Sad: Digital Distortion
When you record audio digitally the analog signal is first converted from an electrical signal to a series of numbers that represent the amplitude of the analog wave form in a moment of time. This is done by using a binary system of code consisting of two digits “1” and “0” Digital, get it? In conventional digital audio these numbers are arranged in “words” consisting of 16 or 24 digits which look like this: “1101011000111010” These strings of digits represent a sample of amplitude for an instance of time. All digital audio is comprised of these samples and the sample rate of digital audio refers to how many times per second the audio signal is sampled during the conversion process. The most common sample rates are 44.1kHz, 96kHz, and 192kHz. CD’s are encoded with a bitdepth of 16 i.e. “1101011000111010” at a sample rate of 44.1kHz. This means that a 16 digit sample is taken 44,100 times per second.
The absolute max level that can be achieved with digital audio is 0dB. Any signal that surpasses 0dB is a clip as there will be no digital number to represent the level. All other levels are displayed as –XdB So what does all that astronaut speak mean for you? DON’T HIT ZERO! No matter what your bit-depth or sample rate the highest achievable level is 0dB as your DAW and all digital audio playback devices are not able reproduce anything louder, therefore instead of the comparable warm distortion in an analog device, clipping digital audio results in nothing other than playback failure awful sound.
To avoid this problem is it important to keep an eye on your level meters. Most level meters have a clip indicator where the top red LED of the meter remains lit until it is reset. When this happens it often means you may have to have another go at your take, pass, or mixdown.
Floor to Ceiling
Another concern with gain staging is a level that is too low. After making the necessary boosts or compression to a channel of audio that is recorded too low, the noise generated from the microphone preamp, summing amp, and other circuitry in your signal chain is amplified as well, resulting in a much noisier version of your intended recording. So it is very important that you have levels that are hot enough to remain clean while avoiding clipping.
It is good to shoot for a max level that is well under your ceiling of 0db while remaining hot. A good benchmark for the sharp transients of a snare for instance is to peak at -6dB while the majority of your tracks should hover around -12dB peaking occasionally at -6dB. The use of a good “brick wall” limiter is a good way to wrangle in those stray transient signals that can otherwise clip your track.
During the mixdown phase it is important to maintain an adequate amount of headroom so that your final mix can be mastered. About -4dB is a nice target to shoot for a pre-mastered mix. This will result in ample headroom to really squeeze everything out of your mix.