Gain staging

Definition

Gain staging is the management of level at every point in the signal chain (source, preamp, converter, clip gain, inserts, fader, bus, master) so that each stage works in its own optimal range: far enough above the noise floor and far enough below the point of saturation or clipping.

The gain of a stage is G = 20 · log10(V_out / V_in) dB; the gains of stages in series add up in dB. Reference units:

• dBu: 0 dBu = 0.775 V RMS; the professional nominal level is +4 dBu (1.23 V);
• dBV: 0 dBV = 1 V RMS; the consumer nominal level is −10 dBV (−7.8 dBu);
• dBFS: referenced to digital full scale.

Alignment between analog and digital sets which dBu level corresponds to which dBFS level: EBU R68: 0 dBu = −18 dBFS; SMPTE RP155: +4 dBu = −20 dBFS. In a 32-bit floating-point DAW the internal bus does not clip (the representable range exceeds 1500 dB), but gain staging is still necessary for three reasons: the converter and fixed-point files have a limit at 0 dBFS; plugins that emulate analog circuits are calibrated around a nominal level (usually −18 dBFS = 0 VU) and saturate above it; compressors, gates and saturators act according to the level they receive.

Summing tracks: N uncorrelated signals of equal level raise the level by 10 · log10(N) dB; N identical in-phase signals raise it by 20 · log10(N) dB.

Parameters and units

• Preamp gain: 20–60 dB for dynamic and condenser microphones.
• Recording level (24-bit): peaks −12/−6 dBFS, RMS −24/−18 dBFS.
• Clip gain / trim: level correction before the inserts, in dB.
• Nominal level in plugins: −18 dBFS RMS ≈ 0 VU (+4 dBu analog).
• Fader: working zone −10/+3 dB; fader resolution is greatest around 0 dB.
• Master bus pre-mastering: peaks −6/−3 dBFS.

Musical function

Correct gain staging makes the behavior of processors predictable: compressor thresholds fall within the values of the scale, the saturation of analog plugins is a choice and not a side effect, the faders stay in a comfortable zone and the master reaches mastering with margin. It does not change the sound by itself: two mixes with different levels but consistent gain staging sound the same after normalization; what changes is the way each stage reacts to the signal.

Practical example

A vocal recorded with peaks at −3 dBFS and RMS at −14 dBFS:

1. clip gain −6 dB → peaks −9 dBFS, RMS −20 dBFS;
2. 1176 emulation calibrated at −18 dBFS = 0 VU: the vocal now enters 2 dB below nominal and the compression responds as on the original circuit;
3. fader at −2 dB in the mix.

Mix summing: 40 tracks with RMS −20 dBFS each, uncorrelated, produce 10 · log10(40) = +16 dB on the bus, that is RMS −4 dBFS and peaks above 0 dBFS. Lowering each track by 12 dB with clip gain brings the bus down to −16 dBFS RMS with peaks around −5 dBFS, a value suitable for mastering. Kick and bass hitting together in phase behave instead like correlated signals: two −6 dBFS peaks summed reach 0 dBFS.

Common mistakes

• Normalizing all tracks to 0 dBFS: the bus saturates and analog plugins distort.
• Compensating for tracks that are too hot with faders at −25/−30 dB: the signal still reaches the inserts too loud before the fader.
• Lowering only the master fader: in floating point it prevents clipping of the file, but plugins inserted on the master still receive a signal that is too loud.
• Recording with the preamp clipping: analog distortion cannot be recovered afterward.
• Judging a processor without compensating for gain: the louder signal always seems better.

How to check it

• App analysis: analyze the individual tracks and the mix; read Sample peak, RMS and Clipped sections (≥3 samples at full scale). Tracks with RMS −24/−18 dBFS and a pre-mastering mix with peaks −6/−3 dBFS indicate correct gain staging.
• DAW meters: in FL Studio, check the meter of each mixer channel before and after the inserts; Wave Candy in meter mode shows peak and RMS.
• VU meter plugin: inserted first on the track, with 0 VU = −18 dBFS, it shows whether the level is close to nominal.
• Level-matched comparison: use the plugin's output gain to match the level in bypass before judging its effect.