return to tranceaddict TranceAddict Forums Archive > DJing / Production / Promotion > DJ Booth

The Compressor
View this Thread in Original format
I still dont fully understand compressors but ever since I discovered that compressors are supposed to be used on every element of your song, I've been using it like crazy.

Should a compressor be used on a kick?... I mean, the reason I ask is because they reckon that if you have a compressor on everything, you have a better chance of hearing every element of your tune.
Tony Morello
Most people compress the hell out of the drums. Try not to over do it though.
Originally posted by Tony Morello

um... no
Oh yeah I forgot. LOL go to the producers forum, its below the DJ booth. They'll help you out there.

a compressor is a tool you use to limit/control dynamic range.

what is dynamic range? if you look at a spectrum analysis of a sound, it is the difference between the loudest and quietest point, as represented by the tallest peak and the lowest valley.

[edit: if you want to know what spectrum analysis is, i would recommend downloading voxengo SPAN (dont worry, its freeware) and run it as a VST in a digital audio workstation. if you dont have one you can use the fl studio demo (which is also free).

just import any mp3 or wav and run the spectrum analyser to get a visual image of the entire song at any given moment as gain (dB) over frequency (hz).]

what a compressor does is physically squash the peaks and valleys closer together. the overall effect of this depends on the sound being compressed although a by product of compression is a gain in headroom which can equate to more loudness.

i mentioned that its action depends on the sound you compress. if you look at a spectrum analysis of a drum machine kick drum its usually a big hump around 40 to 120 hz before tailing off rapidly thereafter. synthetic kick drums are built using sine waves which have very little harmonic content. in the case of a kick drum, presence is much more important than dynamic so its very common for producers to compress kick drums - it squashes the loudest and softest point of the kick together and gains up, making the overall sound much louder and the entire range of bass frequency sounds much more continuously. compressing a kick drum to extremes will eventually result in a dull hard, thudding kind of sound. actually, this is partly where hard house gets its name and its intended effect - many of the percussive elements and the basslines in hard house are compressed very very hard to get this hard knocking effect.

if you look at a spectrum analysis of a woman singing, you will see the spectrum starts at around 120 to 200 hz and consists of lots of sharp spikes between 500 and 8000hz. vocal delivery is usually rich in complex harmonic series' and this is reflected in the peaks and valleys. compressing this sound will make the peaks and valleys closer together. in vocals, dynamic is more important than presense because many of the subtleties of the human voice (and other harmonically complex instruments) result from the harmonic intervals and the resonances it creates (which result in the sudden large emphasis of certain frequencies in an ordered fashion - as a series of peaks). compressing a vocal will make it sound much louder but also much flatter - the extent of this effect depends on the amount of compression you apply but it will eventually turn into a hissy, dull, very flat mush.

so what works best? this depends on the extent of the compression and the desired effect. if you want to make a sound louder, without any concern as to the changes in dynamic range that will occur, then brickwalling a compressor is one way of doing this.

if you want to specifically reduce the dynamic range of a sound (i.e. to make a thin sounding bassline warmer and 'bigger' sounding) then varyign degrees of compression will work.

if you want to limit the gain of an audio track (i.e. if you have a audio channel that clips and goes over 0 dB) you can use a compressor to supress gain above a specified threshold (ill explain this in a minute). often these are called limiters although the mechanics are the same, the typical uses for a limiter are different to those of a compressor.

how do you use a compressor?

all compressors have the same basic controls. threshold. ratio and gain are the important ones.

the threshold determines the level (in dB) in which the compressor takes effect. the lower you set the threshold, the more of the sound it will compress. the compressor will do nothing if you set the threshold higher than the peak level of the signal.

so lets say you put a kick drum through a compressor and set the threshold to -10dB. the kick drum signal peaks at -15 dB (this is actually very quiet for a kick). in this instance the compressor will do thing because the level that the compressor takes effect is higher than the loudest part of the signal.

if the kick drum signal peaks at -9 dB then we are in business but it will only compress the very loudest peak of the drum - only that signal which goes above -10 dB. in this case its only 1 dB so the compressor will have an effect but it will be very small, since it is only acting on a tiny bit of the kick drum.

if the kick drum signal peaks at -1 dB then there is a whole 9 dB of the kick drum that will be effected by the compressor. this will be a fairly big chunk of the loudest part of the kick.

but basically you just need to remember that whenever any part of the signal (in frequency terms [hz]) rises above -10dB, the compressor takes effect.

the extent of this effect is determined by the ratio which tells the compressor how much it should supress the signal that exceeds the threshold.

a ratio of 1:1 does nothing.

at 2:1, if the kick drum sound goes over the threshold by 2 dB (in this example kick drum will therefore peak at -8dB), it will compress it enough to reduce the sound by half this value - i.e. 1 dB. thus, the compressor will bring the peak level of the sound down to -9 dB. still using a 2:1 ratio, if the peak signal of the kick drum exceeds the threshold by 10 dB (so that it peaks on 0 dB), then the compressor will supress the peak level of the signal to -5 dB.

with a ratio of 10:1 it means that the kick drum peak signal has to sound 10 dB over the threshold in order for the output to increase by 1 dB.

experiment with this without using the gain to get an idea of how it works - it makes more sense when you actually do it. so import a kick drum and make sure its peaking at around -5 dB. set the threshold really really low - like -30 dB. if your compressor has attack and release envelope settings, make sure they are all turned right down to 0. ill get to this in a minute.

in this example you will notice a dramatic effect when you change the compression ratio because you are acting on all of the signal which exceeds -30 dB (this is a huge 25 dB in total)

set the ratio to 2:1. you will instantly notice that the kick has suddenly gone really really quiet - the kick will sound like a dull, hard thud with pretty much no bass whatsoever. dont worry about that, its normal. this is what the gain control is for.

you now compensate for the effect of the compression by increasing the overall gain of the sound up until it is peaking approximately where it was originally (-5 dB). you will notice the sound is different after you do this - do an A/B comparison by turning off/bypassing the compressor altogether - become familiar with what overcompression sounds like because thats what you have done - totally overcompressed a kick. the bass end is alot louder and its alot warmer than it was before. it also sounds duller and flatter.

so what about the other controls on a compressor like attack/release and knee?

the attack setting determines how long it takes for the compressor to fully compress the sound once the signal passes the threshold.

so if you set the attack to 500 milliseconds and the signal passes over the threshold, it would take the compressor 500 ms before it fully exerts its effect on the sound. an attack of 0 ms means the compression starts immediately when any part of the input signal peaks above the threshold.

the release determines how long the compressor takes to literally 'release' its action once the signal dips below the threshold. if you set it really long, like 5000 ms, then the compressor will linger and hold its effect for 5000 ms after the peak signal falls below the threshold again. this is good for when you have lots of sounds going on through the same compressor with sudden and unexpected changes in level - you can use a compressor with high release to keep everything smooth and uniform and prevent sounds from suddenly jumping into the mix really loudly. a release of 0 ms means that the compressor will stop compression the exact moment the input signal falls below the threshold again.

the knee? this is a setting which is sort of hard to explain but if you know what filter resonance does to a filter, compressor knee does that for compressors - the knee affects only the signal that is right around the threshold.

with a 'hard' knee setting, signals below the threshold are not compressed at all, and when they exceed the threshold the gain suddenly starts being reduced by exactly the amount tdictated by the ratio. with a 'soft' knee, ss the signal level approaches the threshold it's reduced in level slightly, and the reduction gradually increases until the level crosses the threshold.

this might all sound a bit difficult and abstract but its actually very simple - it will make more sense when you start using it properly.

alot of people put compressors on their tracks and wonder why nothing seems to happen - usually this is because they dont realise that:

1) a compressor has no effect whatsoever when the threshold is set higher than the peak level of the input signal.

2) a compressor has no effect with a compression ratio of 1:1

3) transient, very short sounds like kick drums do not get compressed if you set your compressor envelope attack longer than the duration of the kick. for instance - if your kick lasts 0.3 seconds (300 ms) and you set your compressor attack to 400 ms, then the kick has finished playing 100 ms before the compressor fully comes into action!
^ Great post....f'(x):toothless , Makes more sense when I use it in soundforge now :)
great post. esp the part about the ratio. never knew about that.

I knew alot of that stuff... but thanks man you did help :)


See if you can answer me this one...

Why is it that when I set the attack to 0.0 (like with nearly everything I compress) and set the decay to a short amount of time the first kick of the pattern is always louder?
probably because the release time you set is overlapping onto the attack phase (the click part) of the next kick drum hit and is muting it. solution - decrease the compressor release.

the click part of the kick drum is generally always the part with the highest amplitude.

also, it is wrong to think that using a compressor on every part of a track will make it all stand out in the mix more. the idea isnt to make every element of your mix heard. its important to have many parts of your track quieter. otherwise you will never be able to create the illusion of space and depth in your tracks.

in digital audio, compressing everything hard, on the contrary will flatten the entire mix and make nothing stand out. in order to make something stand out you need to have certain instruments recessed and panned out. thus the compressed instrument will stand out further from the back row instruments. certain effects like reverb essentially make this difference in level and off centredness sound more like a difference in physical space. in fact, adding reverb to quieter instruments makes them seem even further away. which is sometimes what you want. comparitively it will also make the louder, compressed sound seem closer to you. you have to think of using a compressor as a balancing act since its action has a relative effect on the instruments and sounds that are NOT compressed. this gives a sense of space to your mix but its important to have dominant instruments and recessed ones to varying degrees. you cant have everything up front in the mix because then your brain cant perceive any sound is up front or recessed - it sounds like its all on the same row. only twice as loud because you are squeezing every bit of dynamic out of everything.

Thanks man, you've been very helpful! :):):)
no problemo.

just pass on what you learn in the production studio forum. i have no aversion to sharing knowledge - more than likely i learned it from someone else who was gracious enough to share what they have learned. a few people around here like thy and rb2k1 and mr.mysteryman (although i havent seen you active in the producers forum for a while now :mad: ) have my respect for this very reason.

if everyone did that i think we would all be better producers and better DJs. period. since we are all going through a near constant learning curve. theres always some info and tips and tricks you can pick up from someone around here, regardless of their ability.
Privacy Statement