Vocal Gain Staging: The Level Your Vocal Should Hit Before Any Plugin
Vocal mixing · 4 min read
Get the vocal to about -18 dBFS average, with peaks around -10 to -6 dBFS, before the first plugin in the chain. That is 0 VU on almost every analog-modelled compressor and saturator, and those plugins change character with input level — so if you dial settings at the wrong level, every number you chose is answering the wrong question.
Target average
-18 dBFS
= 0 VU · peaks -10 to -6 dBFS
The 60-second answer
At the mic, set the preamp so the loudest phrase peaks -10 to -6 dBFS. In the mix, put a gain or trim plugin first in the chain with a meter after it, then set the average of the busiest phrase to -18 dBFS. Reset to -18 dBFS after the compressor with makeup gain. Use the trim, never the fader.
Why -18 dBFS and not something hotter
Two reasons, and only one is about headroom. First, the analog reference: most plugin models are calibrated so -18 dBFS corresponds to 0 VU, the level the original hardware was built around. Second, 24-bit recording has a noise floor near -144 dBFS, so there is no signal-to-noise argument for tracking hot — that argument died with 16-bit. Peaks at -6 dBFS cost you nothing and cover a singer who suddenly gets excited.
Why the fader does not gain stage anything
In every DAW the channel fader sits after the inserts. Pull it down 6 dB and the plugins still receive exactly the level they did before — you only changed what leaves the channel. So if your compressor is hit too hard, the fader will not help, and if your saturator sounds too clean, it will not help either. Gain staging happens with a trim plugin at the top of the chain or with clip gain. Both sit before the inserts, which is the point.
Gain staging a vocal, step by step
| # | Do this |
|---|---|
| 01 |
At the mic: peaks -10 to -6 dBFS Have the singer do the loudest line in the song and set the preamp there. Never aim for -3 dBFS and never chase the top of the meter. Tracking peaks: -10 to -6 dBFS |
| 02 |
Put a trim plugin first in the chain Before the tuner, before the EQ, before everything. This is the one control you use to set input level. Trim: slot 1 |
| 03 |
Add a meter after the trim A VU meter calibrated to -18 dBFS = 0 VU is easiest to read, because its 300 ms integration matches how you hear syllables. -18 dBFS = 0 VU |
| 04 |
Set the busiest phrase to average -18 dBFS Play the loudest section, not the whole song. The needle should sit near 0 VU with occasional swings to +3. Average: -18 dBFS |
| 05 |
Reset the level after the compressor Makeup gain until the output average is back at -18 dBFS, so the next plugin sees the level it expects. Post-compressor: -18 dBFS |
| 06 |
Use clip gain for big offsets, and leave master headroom Anything over about 6 dB belongs on the audio, not the trim. Keep mix bus peaks at -6 to -3 dBFS. Master: -6 to -3 dBFS |
The trick: the trim in front of an analog model is a tone control
Take any analog-modelled compressor or saturator you own. Put a gain plugin in front of it and another behind it. Push the front one up 6 dB and pull the back one down 6 dB. The output level is identical and the sound is not: more harmonic content, more gain reduction, a different attack curve, sometimes a different frequency balance. That is the model doing its job, because real circuits are nonlinear and their behaviour is defined against an internal 0 VU reference. Two things follow. First, a preset you copied from someone else only means something at the input level they were using, which is why the same plugin with the same numbers sounds great on their session and wrong on yours. Second, you now own a free tone control: trim high for grit, low for clean, compensating on the output so the level never changes while you compare.
Stop guessing the chain.
The Vocal Chain Bible is 88 real vocal chains from Grammy-winning engineers — every plugin, every setting, in order, with free alternatives for each one. 5+ years, 100+ interviews, one PDF.
Get the Vocal Chain Bible — $48 →Mistakes that kill it
- Tracking at -3 dBFS peaks. No extra quality, no extra resolution, and one loud ad-lib away from a clipped converter you cannot undo.
- Using the fader to gain stage. It is post-insert. Every plugin still gets the old level and nothing about the tone changes.
- Trusting 32-bit float to save you. The mix bus will not clip internally, but analog-modelled plugins still change character with input level, long before any clipping.
- Setting the level with no meter. Peak and loudness meters read the same vocal 12 dB apart. Pick one reference and use it for everything.
Gain staging is boring and it is the reason two engineers with identical plugins get different results. Set the level once, at the top, and every setting after it is a real decision.
Keep going
FAQ
What level should a vocal be before mixing?
About -18 dBFS average with peaks around -10 to -6 dBFS. That puts you at 0 VU on analog-modelled plugins with plenty of headroom left.
Does gain staging still matter in a 32-bit float DAW?
Yes. The mix bus will not clip internally, but every analog-modelled plugin is nonlinear and responds to input level, so its character changes with it.
Should I use clip gain or a trim plugin?
Clip gain for anything bigger than about 6 dB or when sections need different amounts. A trim plugin for the one static offset that feeds the whole chain.
Who's behind this
Ivan Meshcheriakov · IvanFromRnD
5+ years ago I moved to LA with zero connections and started building from nothing — 14-hour days, meeting people and interning at studios. It took me 5+ years to meet and interview 100+ Grammy-winning engineers — then I turned it all into the Vocal Chain Bible.
Read my full story →