Gain staging does not make your mix better — here is what it makes possible

Gain staging carries a reputation as a mystical discipline: you are told to hit a level, never told why, and the reason you are given is usually wrong. So you either ignore it or perform it as a ritual. Both cost you something.

First, the myth

"You need headroom or the summing will clip."

In a modern DAW, no. Nearly every station sums in floating point: levels beyond 0 dBFS are represented without damage, and simply pulling the master fader down brings them back intact. An internal bus reading +6 dB has destroyed nothing.

What actually clips are the outputs and the non-floating stages: your interface's converter, the final integer file encode, and any plugin that clips internally at its own zero. That is where margin matters — not in the summing engine.

Burying that myth is worth doing, because it hides the three reasons that do hold up.

Reason 1: non-linear processors listen to your level

A linear digital EQ does not care what you feed it: drop 10 dB going in, add 10 dB coming out, and you get exactly the same sound back. A compressor, a saturator, a tape or preamp emulation does not work that way. Their behaviour depends on input level — that is what non-linear means.

The direct consequence: on those processors, sending 12 dB more does not make the sound louder, it makes the processing different. If your tracks arrive at wildly different levels, you are not comparing settings, you are comparing operating points. You are turning knobs blind.

That is reason one, and on its own it justifies the whole discipline.

Reason 2: analogue emulations have a calibration point

Many modelled plugins are calibrated around a reference level inherited from broadcast: a given digital level corresponds to 0 VU on the original machine — that is, to the point where it sounds the way it is supposed to sound.

Two alignment conventions coexist depending on where the standard came from: −18 dBFS on the European side, −20 dBFS on the American side. That argument is not yours to settle. What concerns you is that your plugin's manufacturer documents its own calibration, and that this is the only figure that counts for that plugin.

Go read it in the manual. A number recalled from memory on a forum is worth nothing here, and I am not going to hand you another one.

Reason 3: the resolution of your faders

A DAW fader is not linear. Around unity, one pixel is a fraction of a dB; at the bottom of the throw, the same pixel is several dB. If every track ends up crushed at the bottom because it arrived too hot, you lose precision at exactly the moment you need it most: fine balancing.

That is not a physics argument, it is an ergonomics argument. It is just as real.

What your peak meter is not telling you

A peak meter measures instantaneous peaks. It measures neither perceived loudness nor what the delivery platform will see.

Two classic traps. A very low-frequency instrument reaches 0 dBFS long before it is perceived as loud, because the ear is less sensitive down there — so you clip it while thinking it is too quiet. And more importantly, a signal can stay below 0 dBFS on every sampled value while exceeding it between two samples. That is what a true-peak meter measures, and that is what your delivery chain will see.

As for streaming platform loudness targets: they exist, they are published, and they move. I am not quoting a number here that I have not verified at the source on the day you read this page. Go and get it from the platform itself before you calibrate a master around it.

The order of operations

1. Fix it at the source. A track recorded into clipping is not repaired by negative gain; the damage is in the file.
2. Normalise the head of the chain, before any processing, with an input gain or a gain plugin — not with the channel fader, which sits at the end of the chain and changes nothing about what your plugins see.
3. Aim for the level your plugin documents, not a universal number.
4. Balance with faders, once the operating points are set.
5. Check the output on a true-peak meter before export.

What it will never fix

Impeccable gain staging does not correct a bad balance, does not reconcile two out-of-phase bass parts, does not declutter an overloaded arrangement, and does not make an average take interesting. It does not make your mix better.

It does something else, and that is why it is worth the time it costs: it makes your next decisions land on what you think they land on. Without it, you make the right calls in the wrong place — and you do not even know that is what is happening.