No Mana

Compressors and Limiters

Added 2025-07-15 21:26:41 +0000 UTC

compressors and limiters are effects that reduce the dynamic range of a sound. they turn your waveform closer resembling a sausage by making the amplitude of a waveform homologous through time.

when you feed sound into a compressor, if the sound is loud enough above its threshold, the compressor takes time to react (the attack) to output the sound to its threshold sound level. limiters on the other hand, work so that the gain reduction is immediate (if not, almost), so a sound never reaches above a certain threshold.

both effects share a release parameter. this is the time it takes for the sound to return to its original, unaffected volume. let's go over all the vital parameters.

Threshold (dB): the loudness point at which your compressor or limiter begins to start ducking the sound.
Ratio (X:X): how much ducking you want.
Attack (ms): how fast you want the compressor to react to incoming sound.
Release (ms): how fast you want the compressor or limiter to return to sounds original volume at output.

for example, let's say a compressor has the following parameters:

threshold: -15dB
ratio: 3:1
attack: 10ms
release: 60ms

this is basically saying:

"if a sound enters the compressor at -10 dB (which is 5 dB above the threshold), the compressor will reduce the level of the signal so that only 1/3 of the amount above the threshold gets through. that means instead of the signal being 5 dB over, it will be 1.67 dB over.

it will take 10 milliseconds for the compressor to reach full gain reduction after the signal exceeds the threshold, and if the signal drops below the threshold again, it will take 60 milliseconds for the compressor to stop compressing and return to normal volume."

to simplify/recap the functions, the attack and release parameters are just tools to tell the compressor how fast you want it to react to loudness. ratio tells the compressor how much you want to squash it whenever loudness reaches over the decibel threshold.

compressors and limiters are used for many things.

controlling unpredictable volume, common in unprocessed vocals, recorded instruments and synthesizers with a lot of automation.
shaping transients, often in drums. you can get a transient more apparent if the attack is slow enough to not "catch" the transient, but fast enough to duck the rest of the sound's body. this makes the drum snappier. limiters can be used to shape transients as well, especially in cases where transients are too clicky or snappy.
sidechaining external sounds. you can use the sound of another source to activate the compressor on a sound. this other source is often a kick, used to duck other larger sounds such as melodic sounds or basslines.
making things feel larger in the mix. when the quieter parts are closer to the volume of the louder parts, the sound overall can occupy more dynamic space and therefore feel larger and apparent.

while compressor/limiting is often appealing especially when seeking a larger sound, there are just as many times where it might not be used:

when you want the body of a transient to pop through the mix
when you want breath and space in a mix
when you want sounds to be as god intended them
when you are trying to avoid over-compression (or limiting) artifacts (i.e. amplitude shakiness)
when volume automation is a better solution

types of compression:

Downward compression: most commonly used - sounds above threshold gets compressed.
Upward compression: rarely ever used - sounds below threshold are compressed towards the threshold volume, and therefore louder.
Expansion: also rarely used, same mechanics of compression except the ratio is inversed, therefore increasing dynamic range as opposed to decreasing it.
Parallel compression (a.k.a. New York compression): mixes a heavily compressed and a lightly (if at all) compressed of the same signal. this helps bring up quieter details of the mix while maintaining dynamic range.
Multiband compression: separate compression settings for each frequency band.
Sidechain compression: compression using an external signal to trigger the effect.
Transient shaping: using an attack on percussive sounds to shape the transient, usually making it sound snappier.

compression is used quite differently across genres.

typically in EDM, it's common to compress more than normal things. as a result, the mix has less dynamic range and producers are able to achieve louder masters. some pop music do the same, reaching LUFs at -5 or even -4. most of the art of compression is in the vocals, doing things like parallel compression to bring out the airiness, handling peaks with a bit of limiting but allowing a small amount of dynamic room where the compression isn't doing anything so that the vowels, consonants and sibilances are still legible.

while many strive for loudness, deadmau5 is an exception to the genre, prioritizing dynamics and dimensionality of the mix. as opposed to having a mix sound heavy, the benefit of quieter mixes is being able to place sounds in front or behind, and being able to get certain sounds in the mix to pop through. while the master isn't quite as loud, ironically listeners benefit from feeling immersed into the mix because of that dynamic representation.

this is why older music is often praised as holy grails of mixes. every bit of dynamics were preserved, not compromised to achieve a louder master. some contemporary releases still achieve this today.

here's how i rate using compressors/limiters on each type of sound:

kicks and snares (2/10) - i might compress a kick/snare to enhance the transient, or limit to control peaks, but i don't find shaping the dynamics with a compressor is the cleanest way as opposed to throwing it in a sampler and managing the adsr that way. compressing kicks or snares also warrants another issue whenever they are played in bursts or rolls - the first sound may benefit from the transient shaping but the rest are entirely ducked.
hats, cymbal, rides one-shot (3/10) - sometimes limiting is used to control the peaks, depending on the hat or cymbal. heavy compression may be used to extend the tail of a one-shot.
top end loops (3/10) - compression used but not common. may be applied if the loop isn't mixed well, and limiting may be applied to control the peaks. over compressing top end can also be a great creative decision if you want to get the mix to sound airy, but can hurt the mix if your top end is a major driving force.
full drum loops (4/10) - compression used sometimes, if you want to glue things together. you can get the body of the kick and snare to be the same volume, while getting the entire loop to sound a bit snappier depending on how you set up your attack and release times.
impacts and ambient drones (2/10) - ambient samples and impacts often have a very predictable dynamic range. sometimes i will want to limit impacts if the transient is too high, otherwise many times i typically throw those types of sounds on the project without any processing, just eq or cutting lows if necessary.
FX (4/10) - these types of sounds vary widely, but most of them only last a short moment in a song so i don't usually have to use compression on them unless i want to change the tone of them with multiband compression or EQ. depending on the sound i may use limiting to control the peaks. such sounds can be vocal FX, reverb shots, miscellaneous synthesizer sounds or foley. there's typically no pattern on using compression with FX and ultimately it will have to come from your own judgment to decide whether or not you want to use it.
raw vocals (10/10) - these always need compression. the dynamics on raw vocals are rarely at a range where it seems musical. compression can never be added as a crutch here and you must pay careful attention to the parameters because the desired dynamics depend on the style, vocalist, and how it was recorded. slower compression can control the dynamics while maintaining legibility, while faster compression can add weight and airiness. you can achieve both with parallel compression. however if you're grabbing vocals from splice, or stems from a remix competition, they are often already processed and don't need further compression.
recorded string and wind instruments (7/10) - classical music often requires as much dynamic range as possible, and compression can make them lose their piano/forte dynamic performances. compression can be used to fit them into a mix a bit better, depending on the song, however it is not always necessary. these sounds are often mid heavy in raw recordings, so multiband compression may be used to bring out the highs.
lead synthesizers and edm basslines (5/10) - multiband compression can fatten the sound at the end stages, but the overall dynamics of most synthesizers are already quite bricked because distortion is often applied to them. the dynamic range of synthesizers are often more predictable than recorded instruments.
guitars (7/10) - the dynamic unpredictability is comparable to vocals, although slightly less so, especially considering that electric guitars go through amps and the range is often lost there. adding compression can help glue it to the mix, but less compression (slower settings) can help maintain the naturality of its dynamic stages (transients, release, etc).