Heavy limiting changes the tone of a master

The master that got thinner on its own

You push the for the last few decibels, match the loudness and compare. The limited version has less weight in the kick and the bass, and the hats and the vocal's seem further forward. On the next song the same move does the opposite and the top end goes dull. You never touched an .

The limiter changes the tone because it applies one gain to the whole mix at each moment, and which moments it turns down depends on which part of the spectrum makes the peaks.

Why it matters

Tone moves with the drive

If you hear the thinner low end and reach for a low shelf, you can start a loop. When the low end is what drives the limiter, more low end in front of it means more gain reduction on every kick, and the extra lows come straight back off the peaks. Put the boost after the limiter instead and you raise peaks above the ceiling you set.

Knowing what drives the limiter tells you where the fix belongs. It also explains why a master can sound balanced at 2 dB of gain reduction and lopsided at 6.

Figure 1. A simulated limiter taking 6 dB off each kick with a 50 ms release. The kick loses its peak, and the steady highs, which never came near the ceiling, dip on every kick as well because the gain is shared.

The science

One gain, shared by every frequency

A limiter measures the peak level of the full signal, computes the reduction needed to keep it under the ceiling, and smooths that gain over time (Zölzer, 2011; Giannoulis, Massberg and Reiss, 2012). At any instant every frequency gets the same gain, so a single moment keeps its spectral balance. What changes is the balance over time, because the parts of the mix do not have their energy in the same moments.

For one band $b$ of the mix, with signal $x_b$ and limiter gain $g$ , the change in its average level is the gain weighted by when that band has its energy:

\Delta L_b = 10 \log_{10} \frac{\sum_t g(t)^2 \, x_b(t)^2}{\sum_t x_b(t)^2}

A band whose energy sits in the moments the limiter turns down loses the most. A band whose energy is spread evenly loses roughly the average gain reduction, which can be much smaller. Makeup gain raises every band by the same amount, so it cannot undo the difference between them.

I ran that calculation on two loops through the limiter in Figure 1 (instant attack, 50 ms release), pushed until the biggest peaks lost 6 dB. In the first, a kick on every beat at 120 made the peaks over steady high-frequency noise standing in for hats and air. The lows lost 4.2 dB of average level and the highs 0.7 dB, a tilt of 3.5 dB toward the top. The tilt grew with the drive: 1.0 dB at 2 dB of reduction, 2.2 dB at 4 and 4.9 dB at 8. In the second loop a bright snare on beats 2 and 4 made the peaks over a softer kick on 1 and 3 and a sustained bass. The highs lost 4.2 dB and the lows 0.3 dB, so the master went 3.9 dB darker. When I replaced the first loop's steady highs with short bursts that start on each kick, the tilt shrank, because the highs then lost level in the same moments as the kick.

Figure 2. Level each band loses, averaged over a loop, computed with the limiter from Figure 1 taking 6 dB off the biggest peaks. When the kick makes the peaks, the lows lose 4.2 dB and the steady highs 0.7 dB, so the master tilts 3.5 dB brighter. When a bright snare makes the peaks, the highs lose 4.2 dB, the lows 0.3 dB, and the master tilts 3.9 dB darker.

Those figures depend on the levels I chose for each part, and real limiters use look-ahead and gentler gain curves, so your numbers will differ. The direction follows the same rule: the part of the spectrum that makes the peaks loses the most level. A fast release adds a second effect, from the gain moving within each bass cycle, which brightens the low end in a rougher way, as the lesson on limiter release explains.

Push the drive and listen to the tone rather than the level, since the demo keeps the loudness matched.

Listen

Drive a limiter and listen to the release

A drum loop or a real mix driven into a limiter, matched in loudness to the original. Push the drive and change the release to hear the drums and the tone change.

Loading demo…

Demo volume100%

Start with your speakers or headphones low.

Try it

Find the trigger, then move it

Loop the loudest chorus. On the master put a flat EQ, then the limiter with a -1 dBTP ceiling, then a gain plugin and a loudness meter showing short-term LUFS.
Drive the limiter to about 6 dB of gain reduction on the biggest hits. Use the gain plugin to match the short-term reading of the bypassed chain, then compare limited and bypassed and write down what moved: weight, brightness, vocal level.
On the EQ, cut a low shelf by 2 dB below 100 Hz and watch the gain reduction. Reset it, then cut a high shelf by 2 dB above 5 kHz and watch again.
Whichever cut lowers the gain reduction more points to the part of the spectrum that drives the limiter. Reset the EQ to flat.
Control that part earlier. For a kick or bass, use clip gain on the loudest hits, a clipper on the drum bus, or a high-pass on sub energy below what the song needs. For a snare or cymbal, tame its peaks in the mix.
Drive the limiter again to the same short-term loudness as step 2 and compare both limited versions with the bypassed mix at matched loudness.

0 of 6 steps done

If those peaks were the trigger, the version where they were controlled first needs less gain reduction for the same loudness and stays closer to the tonal balance of the unlimited mix. If it does not, look elsewhere in the chain for the tone change.

Common mistake

EQing the limiter's side effect

Correcting the limiter's tone with more EQ on the master usually makes it worse. A low boost in front of a kick-driven limiter feeds the trigger. A high cut to tame the brighter top makes the master duller during the parts where the limiter is not working. Fix what drives the limiter, and the tone mostly stops moving.

The second mistake is judging tone at different levels. Turned up, a master seems to have more bass, because the ear's flatten as level rises (ISO, 2023). A limited master compared louder than the original will seem fuller than it is. Match the loudness first; the lesson on monitoring level covers how much level shifts the balance you hear.

Takeaway

Watch what the limiter is listening to

Treat the final limiter as part of your tone controls. Before you push it, find out which part of the mix makes the peaks, and give that part its own peak control in the mix or on its bus. Then drive the limiter only as far as the tonal balance holds at matched loudness. I would rather lose half a decibel of loudness than spend it on a master that tilts every time the kick hits.

Sources

Show all 3

Giannoulis, D., Massberg, M., & Reiss, J. D. (2012). Digital dynamic range compressor design: A tutorial and analysis. Journal of the Audio Engineering Society, 60(6), 399-408.
International Organization for Standardization. (2023). ISO 226:2023 Acoustics: Normal equal-loudness-level contours. ISO.
Zölzer, U. (Ed.). (2011). DAFX: Digital Audio Effects (2nd ed.). Wiley.

Heavy limiting changes the tone of a master

In short

The master that got thinner on its own

Tone moves with the drive

One gain, shared by every frequency

Drive a limiter and listen to the release

Find the trigger, then move it

EQing the limiter's side effect

Watch what the limiter is listening to

Check yourself

Sources

Next lesson · 39 of 44

Written by Virzy Guns

Later in Mixing & Mastering