Practical Mixing

How to use REAPER's stock limiter without squashing your track

By StudioRescue Editorial

Published October 16, 2026

Almost every home producer hits the same wall. The mix sounds fine, you render it, you play it next to a released track, and yours is noticeably quieter and smaller. The obvious move is to put a limiter on the master and push it. That does make the track louder. It also, past a certain point, makes it flat, breathless and tiring, and most people cannot tell exactly where that point was.

This article is about the limiter that already ships with REAPER, what each of its controls actually does according to the official manual, and how to find the point where you should stop. It is deliberately not a full mastering guide.

What ReaLimit actually does

ReaLimit ships with REAPER, so there is nothing to install. Add it to the FX chain of your master track, as the last plugin in the chain.

The REAPER User Guide describes the mechanism plainly, and it is worth reading twice because it explains the squashing problem on its own. ReaLimit lets you set a threshold and a ceiling, and it then transfers sound from the threshold to the ceiling. Sound that comes in at the threshold goes out at the ceiling. No signal can leave above the ceiling, and anything above the threshold is compressed.

The manual's own worked example: if the threshold is set to -6 dB and the ceiling to 0 dB, the overall signal is boosted by +6 dB to carry the threshold up to the ceiling, and input above -6 dB is compressed so it peaks at exactly 0 dB.

Set it up in this order

Why the ceiling is not 0 dB

A sample peak of exactly 0 dBFS is not the same as the actual peak of the waveform between samples, and lossy encoding on streaming platforms can push a file slightly over. Spotify's own guidance for artists is to keep true peak below -1 dB, and below -2 dB if your master is louder than their normalization target, because louder tracks are more prone to distortion when encoded. Leaving a dB of room costs you nothing you can hear.

ReaLimit has a True peak option, which the manual describes as choosing whether to show true peak, a more accurate reading than plain peak. Turn it on so the number you are reading is the one that matters. Double-clicking either fader restores it to its default value, which is a quick way out if you lose your place.

Release is in dB per second, not milliseconds

This one catches people out, because most limiters elsewhere label release in milliseconds and a lot of writing about ReaLimit repeats that assumption. In ReaLimit the release field is expressed in dB/sec: it governs how quickly the limiter stops working once the signal drops back below the threshold.

The manual gives the two failure modes directly. Too long and you will hear audible pumping, where the whole track breathes in time with the loudest elements. Too short and you get distortion artefacts. If you are not sure, leave it where it starts and fix loudness with the threshold instead. Release is a control to reach for when you have already heard a specific problem, not one to set preemptively.

The remaining controls

There is also an in and out mode on the right-click menu offering stereo, multi-mono and multi-stereo. In stereo mode all channels are compressed based on an analysis of them summed together, in multi-mono each channel is analysed and compressed separately, and in multi-stereo each pair is treated independently. For an ordinary stereo song or voice track, stereo mode is what you want, because it keeps the two sides moving together.

How much gain reduction is too much

There is no universal number, and anyone who gives you one is guessing. What is reliable is the method: judge it by what you hear when you bypass, not by the number on the meter.

A useful working budget for a home mix is to keep reduction into single digits on the loudest moments and to see the meter resting at zero through the quieter sections. If the limiter is working constantly from start to finish, it has stopped being a limiter and become a loudness effect, and the cost shows up as a track that sounds smaller the longer you listen.

If you need a lot of reduction to reach a level you are happy with, the honest read is that the limiter is being asked to fix the mix. Pull it back and look at the balance first. If vocals are the part that keeps disappearing, setting the vocal level against the instrumental solves more than the limiter will.

Why chasing loudness is mostly self-defeating

The reason your track sounds quieter than a released one is usually not that you need more limiting. It is that you are comparing at matched fader positions rather than matched loudness, and streaming platforms remove most of the difference anyway.

Spotify normalizes playback to -14 LUFS under the ITU 1770 standard, applying negative gain to tracks mastered louder than that and positive gain to quieter ones. So a track crushed to be loud is turned back down on playback, and what survives the turn-down is the reduced dynamic range you paid for it with. The measurement that decides this is loudness over time, not peak level, which is why a limiter alone is a poor tool for the job.

If you want to know what your track actually measures rather than guessing, install a free LUFS meter and read the integrated value over the whole track. For spoken word the target is a different number entirely, covered in the podcast loudness guide.

What a limiter will not do

• It will not fix a mix that is out of balance. Limiting an unbalanced mix makes the loudest element louder relative to everything else, which usually makes the balance worse rather than better.
• It will not undo clipping or distortion that is already in the file. Damage recorded on the way in is a repair job, not a mastering one. See how to fix clipped and distorted audio in REAPER.
• It will not add width, depth or excitement. Those come from arrangement, mix decisions and the recordings themselves.
• It will not make a quiet, thin mix sound full. It will make a quiet, thin mix louder and still thin.
• It will not guarantee your track survives encoding if the ceiling is at 0 dB. That is what the headroom is for.

If it still does not sound right

1. It sounds louder but duller. You are limiting too hard. Raise the threshold until the transients come back, and accept the lower level for now.
2. The whole track breathes or pumps. Release is too slow for this material, or you are asking for far too much reduction. Reduce the amount first, then adjust release.
3. There is a gritty edge on loud moments. Release may be too fast, producing distortion artefacts. It can also be clipping upstream, so bypass the limiter and check whether the edge is already there.
4. The render is louder than what you monitored. Check that nothing sits after ReaLimit in the master chain, and check whether any render-time normalization is enabled. The render settings guide covers where those options live.
5. It is still quieter than commercial tracks. That is expected and largely irrelevant once platform normalization is applied. Compare at matched loudness before you decide you have a problem.