top of page

09. Compressor

There is kind of a myth around compressor saying that it makes everything sound louder. While it is kind of true in some situations, it is not really how a compressor works.

In this video, we'll see exactly how a compressor works, what parameters you have to control it, and different kinds of compression you can find.

A compressor is actually designed to make things sound quieter by turning down the volume of the loudest parts of a sound.

screenshot 1

So the dynamic of the sound, which is the difference in volume between the loudest and the quietest parts, would be reduced. The dynamic would be compressed, hence the name compressor.

And because quiet parts of the sound will be closer to the loud parts, in terms of volume, everything would sound closer to the ears.
And it can also bring forward some details in the sound that would otherwise be in the background.

There are 5 parameters that you should find on any compressors.

screenshot 2

- The threshold, which the point above which the sound will be compressed. If the sound if not loud enough to reach the threshold, then it won't be compressed

- The ratio, which set how much the sound will be compressed. The higher the ratio, the more compressed the sound. If a sound exceeds the threshold by 10dB and the ratio is at 2:1, then this excess will be divided by 2, so the sound will only peak 5dB above the threshold. If the ratio is set to 4:1, the sound will be compressed by a factor 4, so it will peak at 2,5dB above the threshold.

- Makeup gain or output gain, which allows you to make the overall sound louder to make up for the loss in volume due to the compression.

- Attack, will set how fast the compressor effect kicks in after the signal passed the threshold. So with an attack at 0 ms the sound would be compressed instantly. And with an attack at 100ms the compressor would go from zero to full compression in 100ms after the signal reached the threshold. This is useful to let transients through before the compressor kicks in, we'll see that in a moment.

- Release, which tells how fast the compressor falls back after the signal have returned below the threshold. So with a long release the compressor will stay active for a little while.

It can be hard to actually hear the effect of a compressor at first, so don't hesitate to crank up some knobs to get the other settings right before turning back down the first knobs you cranked up.

screenshot 3

For example  you crank up the ratio and turn the threshold down to exagerate the effect and hear it more easily. Then it can be easier to set the attack and release to values that you like. Then turn the threshold back up to affect only the part of the sound you want to affect and then turn the ratio back down to a more subtle value. And finally you can turn up the makeup to taste.

The thing is when you compress a sound, it will fill the frequency spectrum more consistently, because the quiet elements of the sound will brought up relatively to the loud elements, so the sound would appear thicker. This can be used to make an element cut through the mix without making it actually louder, but it can also make it clash with other instruments in the mix, so there's a fine balance to find there.

Now let's see the effect of a compressor on a percussive sound like a kick drum. Let's be a bit extreme with it so we can hear better what it does, so I'll put a high ratio and a low threshold.
Now let's hear how the attack affects the kick drum.

screenshot 4

As you can hear, a very short attack actually makes the kick quite dull, it doesn't it as hard. That's because with a short attack, the transient, so the very first part of the sound, high in energy, is brought down by the compressor. So to keep the energy of the kick we need to let the transient through before the compressor kicks in. So with a longer attack, the transients have the time to go through before the rest of the sound is compressed.

screenshot 5

And then you want to adjust the release. It sets how long it takes to the sound to go back to normal volume once it went back under the threshold. It all depends on the tempo and the feel of the track, so it is easier to set this one in context, with the other instruments playing. Listen carefully, trust your ears and stop when is feels good.

So the rule of thumb to remember is that generally, a longer attack with a shorter release is good to keep the dynamic of a sound, so the transient have the time to go through before the compressor kicks in, which preserves the energy of the sound.
And a short attack with long release is good to control a sound, if you feel it needs to be more consistent.
It is often a matter of finding the right balance between the two.

Now, this is the main principle of a compressor, but there are different kind of compressors, or several ways to use them


PARALLEL COMPRESSION:

You could use it as as Parallel effect, which means you would split the signal in 2 tracks, to have one channel that is processed by the compressor and one channel that is not. One way to do that is to use a return track.

screenshot 6

You put your compressor on a return track and then send your track to this return effect. This way, the sound of your track will be layered with the effect in your return track.

Another way to do that in ableton is to create an effect rack with 2 chains, one with no effect and one with a compressor.

screenshot 7

This way you would have your original sound doubled by its compressed version, which can be a good way to add beefiness to a sound like a kick for exemple.

GLUE COMPRESSOR:

You can apply a compressor to a group of tracks to affect several instruments together.
This can help tie up those tracks together in a more cohesive ensemble. It can make it sound more like one block that is well held together.

screenshot 8

You could use group compression to glue all your drum tracks together, or all your lead track, or to glue your bass with the kick drum for example.

This is more on the mixing side of things but compressing a group of tracks can also highlight frequencies issues. So use it in combinaison with an EQ on each track of the group, so you can avoid having frequencies of one track clashing with the same frequency range of another track.

Compressors designed to be used on a group of tracks are often called Glue compressors, but nothing prevents You from using a glues compressor on a single track or using a regular compressor on a group. They are essentially the same thing.

MULTI BAND:

You could also split a signal into different bandwidth, to have only the lows in one channel, only the mid in a second channel and only the trebles in a third one. This way you could apply different compressions to each channel.

screenshot 9

Some compressors have this split channel system already built in, they're known as multi band compressor. They work like several compressors, one per frequency band, and usually they let you set the split point between two bands where you want. So they are very handy to have more control for your compression, if you need that level of precision.

OTT:

OTT is a particular preset that got so popular, I thought I needed to include in here. OTT stands for "over the top" and It was originally a preset from ableton's multi-band compressor, but then it got so popular it's been made as a free third party vst plug in by xfer, the guys who made serum.

screenshot 10

And it's really weird because it really squashes the audio like there's no tomorrow, which would be a bad idea in a lot of contexts. 
If you do that on every track, it would feel like you brought every sound closer to you ears and every sound would be competing against each other. That could end up in a very messy mix.
But in the phase of sound design, you can use it like a magnifier to reveal all the texture of the sound that it normally in the background.

Because it compresses the lows, the mids and the trebles independently, and it compresses them so much, everything would be brought up to the same level. And in some music genres, it's become a common practice to put several in a row if you really really want to squash that sound.

So in the end it's a nice thing to try out.
And then you can filter or eq afterward, because it will really push up all the frequencies in the spectrum.

I think it works better on synth sounds than recorded samples, because, with a recording, it would really reveal and bring up a lot of the background noises.

screenshot 11

On the OTT the main parameters would be the time knobs,  which acts a bit like a combination of attack and release.
And the amount knob, which acts a bit like between a general ratio and a dry/wet. The amount knob is not really a dry/wet though. If you put several OTTs in a row with an amount of 0, you would still hear a transformation in the sound.
I think it is due to the way the multiband compressor separates each band of frequency, but anyways that's still a cool thing to try.

Then you have the level of each band that you can use as a small EQ.

The OTT is a particular case, not for all sounds but very popular. So we'll probably talk about it again

SIDECHAIN:

Another way to use a compressor that is very popular in electronic music is to use it in sidechain.

screenshot 12

When a compressor is set to sidechain, it takes an external signal as an input. Then it is this external signal that will trigger the compressor. The most common application is to put a compressor on the bass, and put the kick track in the sidechain input. So everytime the kick plays, it will trigger the compressor which will compress the bass. So the bass will duck every time the kick plays, creating a pumping effect.

Personally, I usually prefer to use an extra track dedicated to trigger sidechain compressors. Basically I would create a midi track, and create a synth that make a clicky sound, by turning down its envelope ' s attack, sustain and release and then setting a very short decay.

screenshot 13

Then I'd draw a note every time a kick is played. It's very easy if the kick is already played by a midi instrument so I just have to duplicate its midi track.
And then I mute this new midi track and use it to trigger my sidechain compressors.

Because the sound of this synth is so short, it's easier to make the release synced to the tempo. You would put the attack very low so the compressor engages right away when it receives the click, and then the release sets how long it takes to the bass to come back at normal level. 

So if you want that time to be of a quarter note, just take 60 000, which is the number of ms in a minute, and divide it by your tempo, which is the number of quarter notes in a minute.
It will give you the time for 1 quarter note, which you can put as the release time.
This way you can have a release time synced to your tempo.

screenshot 14

(If you want the bass to go back up after a 8th note, just divide that value by two. Divide that by two again and you'll have the length of a 16th note.)

Another way to mimic that in Ableton, for a 4 to the floor rhythm, that I find cleaner sometimes, is to put an auto pan effect.

screenshot 15

It's a panning effect, but if you put the phase at 0, sync it to the tempo and use a reverse sawtooth wave form, you can mimic the pumping effect. Then you can play with the shape knob to change the way the bass is coming back up.

Here I'm showing it with a bass and a kick, but it can help in a mix for any instrument to set priorities. For example, you could use it to duck some tracks down when the singer's voice is coming in.

DE-ESSER:

A compressor can also be used as a de-esser which can reduce the sybillances if a voice recording. It's used to reduce the level of the sss and the chh sounds.
To do that you need some kind of multi band compressor to compress only the frequencies where these sybillances live.

Usually the chh sounds are around 5000 to 6000 Hz, and the sss sounds are usually around 7000 to 9000 Hz. (So you could use these values to set a multiband compressor - or a dynamic EQ - so it only compresses those bands of frequencies.)

In Ableton's regular compressor, when you open the sidechain tab with this little arrow, you also have access to an EQ section.

screenshot 16

There you can select the bandpass mode, so only a band of frequencies will trigger the compressor.
This is how the de-esser preset has been built.

But if only the sibilances will trigger the compressor, it's the whole sound that Will be compressed, but it is still a good way to lower some of the harsh parts of the sound.

LIMITER:

A limiter is usually identified as a separate effect, but it works in a similar way than a compressor.

If you crank the ratio of a compressor all the way up until it says "infinite" ratio, you have what is called a "brick wall" compressor.

screenshot 17

This way, no sound would go passed the threshold, because anything that would go past it would be brought down at the the threshold level.
And this is exactly what limiter does. It prevent everything from going higher that a certain level. 
Usually it is used on a track or on the master bus with a threshold at 0dB, to make sure you don't get any clipping.

CONCLUSION:

Overall a compressor will reduce the dynamics of a sound while bringing it closer to the ear of the listener. Maybe this video was more on the mixing side than actual sound design, but I hope it's been useful to some of you.

Compressor is kind of mysterious effect, you hear a lot about it, but it's kind of hard to actually hear what it does. Or at least that something I struggled a lot with when I was starting out. 

screenshot 18

screenshot 19

screenshot 20

screenshot 21

screenshot 22

screenshot 23

screenshot 24

screenshot 25

screenshot 26

screenshot 27

screenshot 28

screenshot 29

screenshot 30
bottom of page