Why boosting the EQ doesn't make your master louder

The most common surprise in mastering, and what it tells you about how the chain is built.

Here is a thing that confuses almost everyone the first time, and it confuses them for a good reason: it looks like a bug.

You push a low shelf up by 3 dB. The master obviously changed — you can hear it. But the loudness reading is exactly where it was. The true peak is exactly where it was. Nothing on the meters acknowledges what you just did.

The tool is not ignoring you. It is doing something specific, and understanding what separates people who fight their mastering chain from people who use it.

Loudness is a destination, not a result

A mastering chain ends the same way almost everywhere: a brick-wall limiter followed by normalisation to a loudness target.

That last stage is the whole answer. The chain does not measure how loud your track came out — it is told how loud to make it. Spotify wants roughly -14 LUFS. Apple Music wants nearer -16. The chain's final job is to arrive at that number, whatever it was handed.

So when you add 3 dB of bass:

1. 1.more energy enters the limiter, which works harder and pulls more gain down,
2. 2.the normalisation stage measures the result and applies whatever offset lands it on target,
3. 3.the meter reads the target. Because that is what a target is.

It is a thermostat. Open a window in a heated room and the radiator comes on; the room does not get colder. If you conclude the window did nothing, you have misread which part of the system is in charge.

So what did change?

Two things, and neither of them is level.

The timbre. This is the obvious one and the one you wanted. Relative balance between bands is exactly what EQ controls, and normalisation does not touch relative balance — it applies one gain figure to everything.

How hard the limiter is working. This is the subtle one, and it is where the real consequences live. That extra 3 dB of low end has to go somewhere. The limiter absorbs it, which means more gain reduction, which means more density, possibly more pumping, and a slightly different sense of movement in the track.

The meter says nothing changed. The character changed anyway. If you boost the low end aggressively and the whole mix starts to feel squashed and breathless, you are not hearing the EQ — you are hearing the limiter reacting to it.

The one number that does move

Watch the gain applied figure instead. If pushing a band up makes the chain pull more gain out, that is your feedback: it tells you how much extra work you just created.

Dynamic range moves too. A boost that drives the limiter harder will show up as a lower estimated LRA on the output, even though loudness and peak are pinned.

Those two are your instruments here. Loudness and true peak are not — they are contractual obligations to the streaming platform, and the chain meets them regardless of what you do upstream.

Why the chain is built this way

Because the alternative is worse.

If EQ changes were allowed to move the output level, every adjustment would be a loudness change and a tone change at once, and you could never evaluate either one. And Fletcher and Munson's equal-loudness contours guarantee how that ends: louder is perceived as better, more or less unconditionally. You would boost something, hear "more", and be unable to tell whether you improved the tone or merely turned it up.

Pinning the output level forces every comparison to be loudness-matched. It is the same discipline a mastering engineer imposes by hand when they level-match before auditioning a change — the chain just makes it non-optional.

It also means the deliverable is always correct. A master that meets -14 LUFS and -1.0 dBTP will pass platform ingestion whatever else you did to it, which is a guarantee worth more than the convenience of a moving meter.

The practical version

• ·Judge EQ by tone. Loudness and true peak will not tell you whether an EQ move was right; only your ears will.
• ·Watch gain applied and dynamic range, not loudness, when you want to know what a change cost.
• ·If it suddenly feels flat or pumps, the limiter is being pushed harder. Back off the boost, or lower the threshold that is feeding it.
• ·Loudness is the last thing to think about, because the chain has already handled it.

That last point is the real reversal. For thirty years, mastering was largely about loudness — how to get more, at what cost. Loudness normalisation removed the prize, and it turned the loudest question in audio engineering into a setting you pick from a menu at the start and never think about again.

Which leaves the interesting part: what the record should actually sound like.