Makeup gain in plain English
A compressor has one job, turning sound down when it gets louder than the threshold. It never turns anything up. So the moment it starts working, the sound coming out is quieter than the sound going in, and the harder it works the quieter it gets.
Makeup gain is a volume control at the very end of the compressor that puts that level back. It turns everything up by the same number of decibels, the loud peaks the compressor squashed and the quiet parts it left alone. Because the peaks were turned down first and the quiet parts weren't, you end up with a sound at the same loudness as before and a smaller gap between its loud and quiet moments.
That smaller gap is what you wanted from the compressor. Without makeup gain it's still there, but at a lower level, so you'd have to push the fader up to hear it. With makeup gain the quiet detail in a vocal comes forward and the loud words stop jumping out, at the level you started with.
If you're new to compressors, the compression lesson covers every control. This page stays on the last one.
Why does a compressor need makeup gain?
Say a vocal's loudest word peaks at −1 dBFS and you set a −24 dB threshold at 4:1. That word is 23 dB over the threshold, so it comes out about 5.75 dB over, roughly 17 dB lower than it went in. Quieter words get turned down less, and the parts under the threshold not at all. The average level of the whole vocal falls by somewhere in between.
The lab shows the same thing on its vocal loop. At those settings the gain reduction meter reaches 10.8 dB on the loudest moments (the compressor takes a few milliseconds to react and the tallest peaks are shorter than that, so the full 17 dB never arrives), and the vocal as a whole comes out about 6 dB quieter. Leave it there and you'll end up pushing the channel fader up by 6 dB to put the vocal back where it was in the mix. That's makeup gain done the long way. The control on the compressor does it in one place, right where the level was lost.
The other reason is fair comparison. When you switch a compressor off and on to judge it, a 6 dB drop in level is the first thing you hear, and it drowns out whatever the compression did to the shape of the sound. Match the levels and the only difference left is the compression itself.
Reading the lab for makeup gain
The lab above is the same compressor as the main compression lesson, with the knee and release fixed so the steps can stay on level.
On the Whole loop picture, the pink compressed shape is drawn at the same scale as the pale original, so you can see it shrink as you compress and grow as you add makeup.
The Loudest peak out readout shows the peak level coming out after makeup gain. Keep an eye on it, because matching the loudness doesn't always keep the peaks where they were.
The crest factor is the gap between the peak and the average. Compression changes it. Makeup gain can't, because turning the whole sound up by 6 dB raises the peaks and the average by the same 6 dB. With the mix at 100%, move the Makeup slider and watch Loudest peak out rise by exactly as much as you add; it's the quickest proof that makeup gain is plain volume.
The Original and Compressed switch in this lab normally plays the original at the same average (RMS) level as the compressed sound, so the lab can't fool you even when the makeup is wrong. Your DAW's bypass button doesn't do that: when you switch a compressor off in a real session, you hear whatever level difference your makeup setting has left. Step 3 turns the lab's match off so you can hear it that way too.
How much makeup gain should you add?
Enough to make the compressed sound as loud as the uncompressed one, and no more. That's easy to miss, because the obvious number to copy is the wrong one.
The gain reduction meter shows the most the compressor is taking off at any moment, which happens on the loudest peaks. Most of the sound is turned down by less, and some of it isn't touched. So if you dial in makeup equal to the biggest gain reduction reading, you over-correct: the quiet and middle parts get a boost they never lost. In the lab, 11 dB of makeup to answer a 10.8 dB reading leaves the vocal about 5 dB louder than the original.
The matched setting for that vocal is 6 dB, a little over half the meter's biggest reading. With a lower threshold and a higher ratio the fraction goes up, because more of the sound spends more of its time over the threshold. This is what the lab measured on the vocal loop, with the release at 100 ms and a 6 dB knee:
| Threshold and ratio | Attack | Biggest gain reduction | Makeup to match | Peak after makeup |
|---|---|---|---|---|
| −24 dB at 4:1 | 10 ms | 10.8 dB | +6 dB | −4.9 dBFS |
| −24 dB at 4:1 | 1 ms | 14.5 dB | +8 dB | −6.1 dBFS |
| −36 dB at 8:1 | 1 ms | 27.4 dB | +19 dB | −5.7 dBFS |
The original vocal peaks at −0.9 dBFS. In every row the compressed vocal is as loud as the original with lower peaks, which is the trade a compressor gives you when the makeup is right.
Use the fraction as a first guess, then match by ear: loop a busy section and move the makeup while you flip the compressor on and off, until neither version jumps out as louder. A LUFS or RMS meter after the compressor helps. A peak meter is the wrong tool for this, since compression changes the peaks and the average by different amounts.
Makeup gain and your peaks
On the vocal, matching the loudness left the peaks about 4 dB lower than the original's. That's the usual result on sustained sounds, where the peaks are the loudest syllables and the compressor catches them.
Drums can go the other way. With the same −24 dB threshold, 4:1 ratio and 10 ms attack, the front of each drum hit gets through before the compressor reacts, while the body of the hit is turned down. Matching the loudness then takes about 6 dB of makeup, and those untouched transients go up by 6 dB with everything else. Step 9 in the lab plays it on the drum loop, and the Loudest peak out readout goes above +2 dBFS on the loudest hits, peaks the original never had.
Most DAWs process audio in floating point, so a peak over 0 dBFS inside a channel isn't stored as clipping. The next plug-in still sees a hotter signal than it was built for, though, and the master output and any 16-bit or 24-bit export will clip at 0 dBFS. If makeup gain sends a drum bus over, a faster attack, a little less makeup or a clipper or limiter after the compressor will sort it out. The headroom lesson covers how much space to leave.
Re-match after every change
Makeup gain is a fixed number, and the amount the compressor takes off depends on every other setting. Lower the threshold, raise the ratio or speed up the attack and the compressor takes more off, so the makeup you set a minute ago is now too low. Ease any of them back and it's too high.
In the lab, changing the attack from 10 ms to 1 ms drops the vocal 2 dB below the original, because the compressor now catches the front of each word as well. Adding 2 dB of makeup puts it back.
Once the makeup is matched, every later change to the threshold, ratio or attack needs a new match before you compare again. It takes seconds, and skipping it means you're judging loudness instead of compression.
Makeup gain turns the quiet parts up too
Compression brings detail forward this way, and it brings noise forward the same way. The quiet parts of a sound get the full makeup gain and very little gain reduction, so after compression they end up louder than they were.
Push the vocal hard in the lab, with a −36 dB threshold at 8:1, and it needs 19 dB of makeup to match. The loudest words lost up to 27 dB and come out about 8 dB quieter than before. The quietest moments lost only about 11 dB, so they gain 7 or 8. Breaths, lip noise, headphone spill and the hum of a computer fan live in those quiet moments, and they rise with everything else.
When heavy compression and makeup make a recording sound noisy, the compressor is doing what you asked. Ease the threshold up, or clean up the quiet parts first with a noise gate or by cutting the gaps between phrases.
Makeup gain vs input gain
Some compressors have a gain control at the start as well as the end, and the two do very different jobs. Input gain raises the level before the compressor measures it, so more of the sound goes over the threshold and the compressor works harder. Makeup gain raises the level after the compressor has done its work. It changes how loud the result is and has no effect on how much compression you get.
Classic hardware shows the difference well. The Universal Audio 1176 has no threshold knob at all: you turn the Input knob up to drive it into compression and set the Output knob to bring the level back, so Output is its makeup gain. The Teletronix LA-2A works the same way under different names, with Peak Reduction for how hard it compresses and Gain for the makeup.
Plug-ins with a threshold knob usually call the end control Makeup, Make-up, Output or Gain. Check where it sits before you trust it. A control labelled Gain that comes before the part that measures the level is input gain, and turning it up adds compression.
What is auto gain?
Auto gain, or automatic makeup gain, is a switch that sets the makeup for you. Logic Pro's Compressor has an Auto Gain setting alongside its Make Up knob, and Ableton Live's Compressor has a Makeup button. They save a step, but they don't promise a matched level.
Designs differ, and your plug-in's manual is the place to find out which one you have. Some work the makeup out from the threshold and ratio alone. They calculate how far a loud peak would be turned down and add back some or all of it, which repeats the mistake of copying the gain reduction meter: on a sound that rarely goes over the threshold, the boost is too big. Others measure the level going in and coming out and keep the two close, which is nearer to what you'd do by ear.
Auto gain gets in the way when you automate the threshold during a song. With a threshold-based design, a move meant to add compression also turns the vocal up. It's also no help on a sound you want louder after compression, such as a drum bus you're pushing harder into the mix, because you have to set that level yourself anyway.
If you use auto gain, treat its level as a first guess and check it by flipping the compressor off and on. When it lands somewhere you don't like, turn it off and set the makeup by hand.
Makeup gain in parallel compression
With parallel compression you blend a heavily compressed copy under the untouched original. The makeup gain on the compressed copy decides how much of it you hear, so here it works partly as a blend control.
In the lab, makeup gain is applied to the compressed signal before it's blended with the dry one. Match the compressed copy first, then pull the mix down, and the blend stays close to the original loudness. With the vocal at 50% mix and 19 dB of makeup, the lab measures the blend at 0.2 dB below the original. Some plug-ins apply their output gain after the blend instead, so check yours before you trust the mix knob to leave the level alone.
When you set it up the old way, with a send to a separate bus, the compressed bus has its own fader and the makeup on that compressor matters less. Set the compressor's makeup so the bus is roughly as loud as the dry track, then bring the bus fader up from silence until you hear it. The parallel compression lesson covers both methods.
How to set makeup gain, step by step
- Turn auto gain off and set the makeup to 0 dBStart from a known place, so you can hear exactly how much level the compressor takes away.
- Set the compressor firstGet the threshold, ratio, attack and release where you want them by watching the gain reduction meter and listening to the shape of the sound. The level doesn't matter yet.
- Guess the makeupStart at about half the biggest gain reduction reading for gentle compression, a bit more for heavy compression.
- Match it by ear on a loud sectionLoop eight bars or so and flip the compressor off and on, nudging the makeup in 0.5 dB steps until neither version sounds louder.
- Check the output peaksLook at the compressor's output meter, or the next plug-in's input meter. If the peaks are higher than before, decide whether the next stage can take them.
- Re-match whenever you change somethingAny move to the threshold, ratio or attack changes how much level the compressor takes away.
Common makeup gain mistakes
Judging compression while it's louder
A slightly louder version usually sounds better in a quick comparison, so a compressor with too much makeup tends to win the A/B and teaches you nothing. If a compressor sounds better, check that it still sounds better at the same loudness before you keep it.
Copying the gain reduction reading
The meter shows what happens on the peaks. Makeup equal to that number leaves the sound louder than it started, by about 5 dB on the vocal in the lab.
Using makeup gain to mix
Makeup gain is for undoing the compressor's level change. When you want a vocal louder in the mix, use the fader, so the compressor's bypass still tells you something true and the plug-ins after it keep getting the level they were set up with.
Pushing the next plug-in too hard
Saturation, tape and console plug-ins react to the level they receive. Twelve decibels of makeup ahead of a saturator gives it twelve decibels more drive, and much more distortion than you set it up for. The gain staging lesson covers keeping levels sensible from one plug-in to the next.
Forgetting it's there on a preset
A compressor preset made for a loud, dense recording might carry 8 or 10 dB of makeup. On a quieter recording that preset barely compresses, and the makeup just makes the track louder. Check the makeup every time you load one.
Train your ears
The quiz under the lab covers the controls and asks whether the attack is fast or slow, or how high the ratio is. For each round it measures the level and sets the makeup to match, so loudness never gives the answer away. That's the matched comparison this page is about, done for you, and it's a good way to learn what compression sounds like once the volume change is out of the picture.
Questions people ask
Does makeup gain add compression?
No. On almost every compressor, makeup gain comes after the part that measures the level, so turning it up changes the volume and nothing else. The gain reduction meter doesn't move when you change it. Input gain is different: turning that up pushes more of the sound over the threshold.
Should makeup gain equal the gain reduction?
Usually not. The meter shows the most the compressor takes off, which only happens on the loudest peaks, and most of the sound is turned down by less. In the lab, the matched setting came out at roughly 55% to 70% of the biggest reading.
Is auto gain good or bad?
It's fine as a first guess. Auto gain often gets the level wrong, so check it against the bypassed sound and correct it by hand. Turn it off when you want a particular amount of extra level, or when you're automating the threshold.
Why does my compressed track sound worse when I match the level?
Because the extra loudness was doing the work before. A louder version usually wins a quick comparison. If the matched version loses, the compression isn't helping yet, so try a slower attack, a higher threshold or a lower ratio.
Where is makeup gain on an LA-2A or 1176?
On the Teletronix LA-2A it's the Gain knob, next to Peak Reduction. On the Universal Audio 1176 it's the Output knob; the Input knob is what drives the 1176 into compression, because it has no threshold control.
Can makeup gain cause clipping?
Yes. It raises the peaks along with everything else, and if the compressor lets transients through, matching the loudness can push them past 0 dBFS. Watch the output meter. If the peaks go over, use a faster attack, less makeup, or a limiter later in the chain.