Learn
Auto Gain Staging: Keep Your Levels Consistent Automatically
Some mixes fight you from the first hour. The vocal jumps out in the chorus and vanishes in the verse, the drum bus is quietly clipping into a compressor, and by the time forty tracks are playing the master sits in the red. Most of that mess starts in one place: nobody set the level going into each plugin.
That job is called gain staging, and it no longer has to be done by hand. This guide explains what gain staging actually does, when doing it manually still makes sense, and how a dedicated tool like our plugin WaveBalance can take care of it while you get on with the mix.
The Basics
What gain staging actually is
Gain staging means setting a sensible signal level at every stage of the chain, so each processor receives the input it was designed for. Too quiet and you sit close to the noise floor; too hot and you clip, and digital clipping is abrupt and ugly rather than the gentle saturation analog tape gave you. The space you keep between your working level and the clip point is your headroom.
The common reference comes straight from the analog era. Studio gear was built to run around 0 VU, and when that world was translated to digital, 0 VU landed at roughly -18 dBFS. That is why so many engineers aim for an average level near -18 dBFS with peaks somewhere between -10 and -6 dBFS. It is a range, not a religion, but it is the range most equipment and most plugins assume.
Here is the mechanism that makes this matter in a modern DAW. A compressor's threshold is a fixed point: feed it a signal 8 dB hotter than intended and it clamps down far harder than the preset's designer ever heard. Saturation and tape emulations generate distortion as a function of input level, so the drive you hear depends directly on the level you send in. Analog-modeled EQs and channel strips are calibrated so their sweet spot sits around that same -18 dBFS operating level. Change the input level and you change how every one of these processors behaves, before you touch a single knob.
There is a second, quieter benefit. When every stage runs at unity, so a plugin's output matches its input, you can bypass it and judge honestly whether it helps. Louder nearly always sounds better to the ear, and sloppy staging lets that bias make your decisions for you.
And yes, your DAW mixes in 32-bit or 64-bit float internally, so the mixer itself will not clip between plugins. That argument is correct and it misses the point: the processors sitting on the mixer still care about their input level, and the float headroom vanishes the moment you export to a fixed-point file.
Two Ways To Do It
Manual gain staging vs automatic
Manual staging is not obsolete, and automatic staging is not cheating. They solve the same problem at different scales.
| By hand | With an auto gain staging plugin | |
|---|---|---|
| Time | Solo a track, watch the meter, trim, repeat for every track in the session | One analysis pass per track, or one pass for a whole group at once |
| Consistency | Depends on your patience and on the day | Same target, same result, every session |
| Dynamic sources | Clip gain section by section, or ride the fader yourself | A ride mode follows the level continuously |
| When the material changes | Re-trim by hand and hope you catch everything | Re-analyze in seconds |
| What you learn | A lot. You hear every track against a meter | Less. The plugin does the listening for you |
| Best for | Small sessions, learning the craft, final creative balance | Large sessions, uneven sources, repeatable mix prep |
One honest caveat: no tool makes balance decisions for you. Gain staging gets every track to a healthy operating level; the mix itself, which element sits in front and which sits behind, is still your faders and your taste. Staging just means those fader moves start from a predictable place.
In Practice
How to auto gain stage vocals, stems, and a full mix
The steps below use WaveBalance terminology, but the workflow applies to any DAW and any staging tool: set a target, measure, correct, and only then start processing.
Vocals
Vocals are the classic case for continuous control, because no single trim value fits a performance that whispers in the verse and belts in the chorus.
- Record sensibly first: average around -18 dBFS with peaks near -10. No plugin can remove distortion from a take that clipped the preamp.
- Insert the gain staging plugin as the first insert on the vocal track, before the compressor and any tone shaping.
- Pick VU or RMS metering and set the target, 0 VU being the traditional starting point.
- Switch to Ride mode so the correction keeps adapting through the performance. Slow the response until the movement feels like a careful fader ride, not pumping.
- Raise the threshold above the room tone so breaths and pauses do not get pulled upward.
- Now set up your compressor. Because it receives a steady level, it works on tone and energy instead of fighting jumps, and it usually needs far less gain reduction.
Stems and busses
Stems arrive at whatever level another session happened to leave them, so treat staging as step one, before you judge anything by ear.
- Lay out the stems and put one instance of the plugin first on each stem track, or early on each bus.
- If your tool supports grouping, link the instances and share the target across them, keeping per-track response and threshold free. In WaveBalance this is the group workflow with Analyze All.
- Run one analysis pass over a representative section of the song, not the quiet intro.
- Lock the corrections. A static, held correction is right here: stems are usually internally consistent, they are just delivered at the wrong level.
- Only now balance the stems with faders and add bus processing. And if a stem was delivered clipped, ask for a re-export; turning a clipped file down makes it quieter, not cleaner.
A full mix, before mastering
Mix bus headroom is not something you create at the end. It falls out naturally when the tracks underneath are staged.
- Stage every track to the same operating level as it enters its chain, using static corrections for stable sources and ride for uneven ones.
- Keep faders near unity while you build the balance; that is where they have the most resolution, and staging is what makes it possible.
- Watch the mix bus. With tracks staged around -18 dBFS average, a full mix tends to land with peaks around -6 to -3 dBFS on its own.
- Leave 3 to 6 dB of headroom on the mix bus for the mastering stage, and do not chase loudness here; that is mastering's job.
- If the bus still runs hot, resist pulling the master fader down. The fader sits after your bus processing, so it hides the problem instead of fixing the level going into it.
Demonstration
Watch auto gain staging in action
The official WaveBalance walkthrough shows the whole workflow on real material in a few minutes.
What the walkthrough covers
- Setting a target level and choosing between the VU, RMS, and LUFS measurement modes
- Running an Analyze pass on a track and locking the correction so it holds while you mix
- Ride mode keeping an uneven source at a steady operating level in real time
- Reading the applied gain display next to the input and output meters
- Grouping several instances and staging related tracks together
Where WaveBalance Fits
The workflow above, as a plugin
We built WaveBalance to do exactly what this guide describes. You set a target, it measures the incoming signal and applies a smooth correction, and it stays out of the way of everything downstream. Two modes cover the two situations that come up in practice: Analyze & Hold measures a representative section, locks the correction, and keeps it fixed, while Ride keeps adapting in real time for vocals and other uneven sources.
It meters in dBFS, VU, RMS, or LUFS, with the VU mode calibrated to the classic 0 VU = -18 dBFS standard this guide is built around. A threshold control stops it reacting to silence and room tone, a ceiling keeps the corrected output safely away from clipping, and instance grouping lets you stage a whole drum kit or a folder of stems in one pass with Analyze All.
WaveBalance is €14.99 as a one-time purchase, in VST3 and AAX on Windows and VST3, AU and AAX on macOS. There is a free 7-day trial.
FAQ
Gain staging questions, answered
Do I still need gain staging if my DAW uses 32-bit float?
The mixer itself will not clip internally, that part of the argument is true. But plugins that model analog gear are calibrated to sound right around an operating level of roughly -18 dBFS, and a compressor reacts completely differently when the input arrives 10 dB hotter than its threshold expects. The float headroom also disappears the moment you export to a fixed-point file. Gain staging still matters, just for different reasons than in the analog days.
Does auto gain staging reduce dynamics?
No. A static correction is a single fixed gain offset, exactly like turning a trim knob once, so the dynamics inside the performance are untouched. Ride mode adjusts gain slowly and smoothly, which is closer to a hand on a fader than to compression. Neither reshapes individual notes or transients the way a compressor does.
Where in the chain should a gain staging plugin go?
First insert, or as close to the start of the chain as you can get. The point is to control the level before it reaches processors that care about their input, such as compressors, saturators, tape emulations, and analog-style EQs. Placing it after those defeats the purpose.
What level should I gain stage to?
For most music work, aim for an average level around -18 dBFS, which corresponds to 0 VU on a classic meter, with peaks landing somewhere between -10 and -6 dBFS. Treat these as ranges, not laws. The goal is a healthy operating level with headroom to spare, not an exact number on every track.
Should I use peak, RMS, or LUFS readings for gain staging?
VU or RMS is the most natural reference for mixing, because it follows average energy roughly the way your ear does. LUFS suits spoken word and loudness-oriented delivery work. Plain dBFS peak readings are mostly useful for checking headroom, since two sounds with identical peaks can have wildly different perceived levels.
What is the difference between gain staging and normalizing?
Normalizing is a one-time file operation that raises a whole clip until its highest peak hits a chosen value. It knows nothing about average level or what your plugin chain expects. Gain staging is about the level flowing into each processor while you mix. A normalized file can still be badly gain staged.