Gain staging · Armon.ai

What it is and why it matters

Gain staging is the management of levels along the entire chain: from the sample or recording, to the channel, plugins, fader, buses and master. In a DAW like FL Studio the internal engine works in floating point (32-bit float): going above 0 dBFS inside the mixer doesn't permanently ruin the signal as long as you bring it back below before the output or export. So why worry about it?

Plugins with analog behavior (emulations of compressors, EQs, saturators, consoles, tape) are calibrated: many have their "0 VU" at −18 dBFS. Above that level they saturate and change character; far above, they sound bad. Below, they are "clean".
Compressor and gate thresholds depend on the input level: chaotic levels make presets and settings useless.
Headroom for the master: if the buses already arrive close to 0 dBFS, the master clips or the limiter works excessively.
Export and converters: the final file and the outputs to the interface are fixed-point: above 0 dBFS there is real clipping.
Faster work: with consistent levels, the faders stay near 0 dB where they are most precise.

Reference levels

Point	Peak	Average (RMS or short-term)
Recording (24-bit)	−12/−6 dBFS	−24/−18 dBFS
Single track in the mix (pre-fader)	−12/−6 dBFS	about −18 dBFS (−20 to leave more room)
Bus (drums, music, vox)	−10/−6 dBFS	−18/−14 dBFS
Master (premix for mastering)	−6/−3 dBFS	integrated about −20/−14 LUFS
Final master	ceiling −1 dBTP	according to the target (see "Loudness by platform")

With 24 bits you have about 144 dB of theoretical range: recording "hot" isn't necessary; peaks at −12/−6 dBFS leave margin for louder performances.

Procedure in FL Studio

Sources: normalize or adjust the samples (in Edison or in the sample properties) so they don't go beyond −6 dBFS; for synths use the volume of the Channel Rack or the synth's master.
Faders at 0 dB: leave the mixer faders at 0 dB at the start and set the level before the plugins, with the channel volume or with Fruity Balance in the first slot (trim).
Measure: watch the mixer meters (peak) and use Wave Candy in Peak meter mode with RMS mode (300 ms window) or VU for the average, or the ITU-R BS.1770 mode for loudness; or Fruity dB Meter. Target: average about −18 dBFS on the main tracks.
Plugin by plugin: each plugin should output at about the same level it receives (output gain/make-up), so the A/B comparison is honest and the next stage receives the expected level.
Buses: the sum of the tracks is higher: lower the tracks or the bus input trim, not the master fader.
Master: no plugins (or only metering) while balancing; mix peaks around −6/−3 dBFS. If it clips, lower all the tracks (multiple selection) or the buses, not the master.
Before exporting the premix: peaks −6/−3 dBFS, no limiter (if you are sending it to a mastering engineer) or master chain active if you master it yourself.

Calibrating your listening

Working always at the same listening volume helps you judge: many engineers calibrate their monitors so that pink noise at −20 dBFS RMS produces 79–85 dB SPL (depending on room size); Bob Katz's K-System ties metering level to listening level (K-20, K-14, K-12) for more or less dynamic genres. In a home room you often work quieter (70–78 dB SPL) and check louder every so often.

Gain staging and loudness

Gain staging doesn't decide the final loudness: that comes in mastering. A mix averaging −18 dBFS with peaks at −6 dBFS is perfect for the master; the master will bring it to −14, −10 or −7 LUFS.

By genre

Dynamic genres (jazz, orchestral): leave more headroom (peaks −10/−6 dBFS even on the buses).
Dense genres (trap, EDM): watch the buildup of 808 and kick; keep the low-end bus under control (peaks −8/−6 dBFS) before the clipper and master.

Common mistakes

Lowering only the master fader to avoid clipping: the plugins on the tracks are still running too hot.
Analog emulations fed at −3 dBFS: unwanted distortion and compression.
Plugins that increase the volume (EQ, saturation) without compensation: a chain that keeps growing and skewed comparisons.
Normalizing everything to 0 dBFS: no headroom for processing.
Confusing peak and average: a kick can have peaks at −6 dBFS and an average of −20 dBFS; a pad peaks at −10 and averages −14.

How to hear it

On a compressor or saturator emulation, try the same sound at −24, −18 and −8 dBFS average: hear the change of character.
A/B comparison with compensated output: without compensation, the louder one always wins.

Starting settings

Parameter	Value	Why
Average per track	About −18 dBFS (RMS/VU)	Calibration zone of analog emulations.
Peaks per track	−12/−6 dBFS	Headroom for buses and master.
Premix peaks	−6/−3 dBFS	Margin for mastering.
Initial faders	0 dB	Precision and consistent levels.
Trim	Fruity Balance in the first slot	Level before the plugins.
Every plugin	Output ≈ input	Honest comparisons and the expected level for the next stage.
Recording bit depth	24-bit	Headroom without noise.

Frequencies to watch

Range	What to do
Below 100 Hz	Kick, 808 and bass consume most of the headroom: check the bus peaks here.
Whole spectrum	Gain staging is not equalization: it is level; but EQ boosts also raise the level and must be compensated.

Example chain

Step	Tool	Settings	Why
Source	Channel Rack (volume) / Edison (normalization to −6 dBFS)	Peaks below −6 dBFS	Consistent foundation.
Input trim	Fruity Balance	Average about −18 dBFS	Plugins in their zone.
Processing	EQ, compression, saturation with compensated output	Output ≈ input	Stable chain.
Faders and buses	FL mixer	Bus peaks −10/−6 dBFS	Headroom.
Measurement	Wave Candy (Peak/RMS/BS.1770) / Fruity dB Meter / Emphasis (LUFS)	Peak and average	Objective control.
Master	Metering only during the mix	Peaks −6/−3 dBFS	Ready for mastering.

How to check the result

Every track: peak −12/−6 dBFS, average around −18 dBFS.
Buses: peaks −10/−6 dBFS; premix master: peaks −6/−3 dBFS, no clipping.
Plugins: input/output level difference within ±1 dB (except for deliberate choices).
Premix: integrated loudness about −20/−14 LUFS, crest factor 12–18 dB.