Gain staging means keeping the signal at a sensible level at every point in the chain. The practical target: average levels around -18 dBFS and peaks around -6 dBFS on individual tracks, with the mix bus peaking near -6 dBFS and no limiter on it. Not because anything breaks otherwise, but because processors sound different at different input levels, and because headroom is the thing you run out of last.
In a modern DAW running 32-bit or 64-bit floating point internally, the mix bus does not clip. You can push a channel to +20 dBFS, pull the fader down afterwards, and the audio is mathematically intact.
Clipping happens at the fixed points:
- The analogue-to-digital converter on the way in. Once a recorded take clips at the interface, it is clipped permanently.
- The digital-to-analogue converter on the way out. Your monitors will distort if the master exceeds 0 dBFS, even if the file is fine.
- File export in a fixed-point format. A 16-bit or 24-bit WAV cannot represent anything above 0 dBFS. Samples above it are truncated.
- Lossy encoding. MP3 and AAC encoders can produce output several tenths of a dB louder than the source, so a master peaking at exactly 0.0 dBFS will clip after conversion. Leave -1 dBTP of true-peak margin.
- Any plugin with a fixed internal ceiling, which includes some hardware emulations.
Analogue consoles had a nominal operating level with roughly 18 to 20 dB of headroom above it before distortion. Plugins that emulate analogue circuits inherit that design: their saturation, transformer modelling and compressor knees are calibrated for signals in that range.
Feed such a plugin a signal peaking at -1 dBFS and you are driving a modelled circuit far past the level it was designed around. Sometimes that is exactly the sound you want. Often it just sounds harsh, and you did not choose it.
Digital-only plugins — clean EQs, linear-phase filters, straightforward digital delays — do not care about input level at all.
| Stage | Target | Note |
|---|
| Recording input | Peaks -12 to -6 dBFS | Never touch 0 dBFS; a clipped take is unrecoverable |
| Individual track | Average around -18 dBFS | Use clip gain or a trim plugin, not the fader |
| Group and bus | Peaks around -10 dBFS | Keeps bus compressors in range |
| Mix bus before mastering | Peaks around -6 dBFS | No limiter |
| Final master | -1 dBTP true peak | Loudness targets are a separate question |
Use clip gain or a trim plugin, not the channel fader. The fader is after most of the chain. If a plugin is receiving too much signal, pulling the fader down does nothing about it. Set the level going into the chain with clip gain or an input trim, and reserve the fader for balance.
Compensate output on every plugin you add. A compressor doing 5 dB of gain reduction with 5 dB of makeup gain leaves the level roughly where it was, so the next plugin sees what you expect. Without this, levels drift upward through the chain and the last processor is being hammered.
Start faders at unity, build up from the loudest element. Set the kick or lead vocal at a sensible level first and balance everything against it, rather than pushing every fader upward in turn until the master clips.
Watch the master with everything soloed off. If the master meter is peaking above -3 dBFS during a mix, you are mixing into a corner. Pull down a group fader, not the master.
Headroom is optionality. A mix with plenty of it can be made louder, saturated, bus-compressed, or handed to a mastering engineer. A mix already touching 0 dBFS can only be made quieter or squashed. Gain staging is not about correctness — it is about leaving yourself somewhere to go.
Related reading: LUFS and loudness explained, what is mixing, and sample rate and bit depth.
Start making music in Veena
Free, browser-based, no downloads required.
Try Veena Free