Do I understand what compressors do?

M

maximedupre

New member
I've been through a couple of compressor explanations on different sites, but I still don't think I really understand what compressors do. This is what I understood (please tell me what I understood wrong and also some other tips to help me :D).

So basically, from what I understand, compressors don't have any effect unless the volume of a certain sound exceeds a number of decibels (that is set by the threshold). If the volume exceeds that number of decibels, the sound is compressed and the volume is decreased depending on the ratio (I don't have a ratio knob, instead I have an "amount" knob that goes from 0-100%, so I guess 100% would let 100% of the volume go through and 50 % would let 50% of the volume that passes the threshold go through, etc). The compression is done by increasing the volume of the highs but not the lows (when passing the threshold, more and more lows are not seeing their volume increased depending on the ration/amount). Also, the Gain Reduction is the number of decibels that got cut down by the compressor. The Knee just specifies how intense the compression should be. All of theses things I just said, I am not too sure if i am right or wrong, so please correct me if I'm wrong. I pretty much understand all of the knobs (''threshold", "Knee", "Amount", "Attack", "Release", "Gain"), but not so much the behaviour of the compressor.

Please help a future producer :D
 
maximedupre said:
The compression is done by increasing the volume of the highs but not the lows (when passing the threshold, more and more lows are not seeing their volume increased depending on the ration/amount).
Click to expand...

Got it all but this.

A compressor does not increase anything(make up gain aside), it decreases it.
Any sound over the threshold gets lowered in volume.

if your working with bass heavy audio, it may seem like the compressor boosts the highs. This is because the bass frequencies are hitting the compressor before the high frequencies do resulting in the lowering of the bass and not the treble frequencies.
 
Last edited:
Compression is an effect that is associated with amplitude, and I will have to state that amplitude is commonly confused with volume. It is not the same. If you have delved into oscillation, you will notice that the crest and trough of the wave sometimes may "clip" due to high AMPLITUDE. You are increasing the size of the wave within the oscillation, and causing distortion. A compressor allows you to manage, and limit the wave so that clipping does not occur. Compressors can sometimes have effects associated with the device, but the concept if fairly simple. For instance, if you add multiband compression to your master track, you can manipulate the hertz range of your low end, mid, treb... etc... They are able to limit the amount of those frequencies coming through in your recording AFTER the compression has been applied.

Simply put, a compressor is a device that allows you to increase the VOLUME of a recording by manipulating the AMPLITUDE.

That is all I know about it.
 
Last edited:
PresentProducer said:
The boys'll be here soon....
Click to expand...
Yes we will..... :)

maximedupre said:
I've been through a couple of compressor explanations on different sites, but I still don't think I really understand what compressors do. This is what I understood (please tell me what I understood wrong and also some other tips to help me :D).

So basically, from what I understand, compressors don't have any effect unless the volume of a certain sound exceeds a number of decibels (that is set by the threshold).
Click to expand...

True

If the volume exceeds that number of decibels, the sound is compressed and the volume is decreased depending on the ratio (I don't have a ratio knob, instead I have an "amount" knob that goes from 0-100%, so I guess 100% would let 100% of the volume go through and 50 % would let 50% of the volume that passes the threshold go through, etc).
Click to expand...

knowing your daw/plugin/model/make would help untangle this for you but I am going to guess it is logic based on some of the words and phrasing that you are using

A compressor is a device that is used for taming dynamics: for making louder signals softer.

The first setting is the threshold at which the compressor begins to act.

Depending on the model/make of compressor, this can be
  • a peak value rendered as a nominal value below 0dbfs (for that device)
  • a RMS value expressed as a nominal value (can be positive or negative)

The compression is done by increasing the volume of the highs but not the lows (when passing the threshold, more and more lows are not seeing their volume increased depending on the ration/amount).
Click to expand...

Totally wrong - a compressor is not frequency dependent unless it is part of a multi-band compressor which is a 3 or 4 way crossover filter bank followed by separate compressors.

We describe the degree to which the compressor works (after the signal is above the threshold setting) in terms of its compression ratio:
  • 1:1 means no change
  • 2:1 means 2db in 1db out
  • 4:1 means 4db in 1db out
  • 12:1 means 12db in 1db out
  • 24:1 means 24db in 1db out (effectively a limiter which is ideally ∞:1)

If we were to have a ratio of 1:2 (some folks would prefer that to read 0.5:1), we would have an expander, which is the opposite of a compressor, making the louds louder.....

Also, the Gain Reduction is the number of decibels that got cut down by the compressor.
Click to expand...

The nett effect is to lower the high intensity portions of a signal, whilst maintaining the low intensity portions of the signal at the same levels - this is the gain reduction.

The Knee just specifies how intense the compression should be.
Click to expand...

The knee is the point at which the compression curve changes angle from 1:1 to the actual compression ratio set (comes form the older use of the word knee to mean angle) - so the knee is the angle at the threshold value.

the following image shows an ideal hypothetical compressor that has its threshold set at -20db. Each line shows the effect of different ratios as described above. There is no attempt to include make-up gain in this image.

compressorRatios.png


Based on use of the word as the label for a knob control further down, it would seem reasonable to assert that the knee is being used as a substitute for the compression ratio.

All of theses things I just said, I am not too sure if i am right or wrong, so please correct me if I'm wrong. I pretty much understand all of the knobs (''threshold", "Knee", "Amount", "Attack", "Release", "Gain"), but not so much the behaviour of the compressor.
Click to expand...

The last knob, the gain knob is used as follows:

To compensate for a perceived drop in relative loudness, some manufacturers also added a makeup gain stage - essentially a small amplifier that increases the overall signal level; this is graduated in db, so that a precise increase in overall levels can be achieved.

This can have the perceptual affect of making the high level frequencies sound louder. In fact what is happening is the full power spectrum is increased uniformly at all frequencies, it is just that our hearing is more sensitive to high frequencies compared to low frequencies and so we perceive an increase in high frequency energy.

jaysummers75 said:
Simply put, a compressor is a device that allows you to increase the VOLUME of a recording by manipulating the AMPLITUDE.
Click to expand...

I can't let that stand as it is a misrepresentation - volume is not increased as a result of using a compressor - what a compressor does allow you to do is tame the signal and therefore increase the fader/gain of that sound in the mix without it clipping - but it does so at the cost of dynamic range.

the out and out use of compressors as a creative tool is not in dispute but it is hardy ever required to achieve it's original purpose of squashing the dynamic range of a signal.

In fact, I find it incredible that most folks need to use them for that any more as the dynamic range available in 16 bit audio is 96db and in 24 bit audio it is 144db (greater than the extremes of human hearing 0db-130db) - if you need to compress a signal to tame it then it is far too hot (and most likely clipping) to begin with and your approach to mixing is not taking into account your available dynamic range
 
Last edited:
bandcoach already covered most of the details, so now I'm just going to make an even easier overall explanation...

I guess the simplest way to explain a compressor is to simply think it as an automatic volume knob - when something's high enough, it's automatically turned down. That's it. The rest are just nuances (albeit very important ones) of how said turning down is done.

And while the "knee" parameter was already explained, I think it's even easier to just look at an image of it:

basics_01.jpg
 
Holy crap! Thank you guys, this is very helpful. I will probably come back to this post once in a time just to refresh my memory. I did not understand every single thing (hi bandcoach), but I'm getting better and better and things will come naturally with time ;).
 
Kinda funny after all that amount of detail as to what a compressor does that was posted before you, you still felt like there was something left out and had to say that. :o
 
Last edited:
bandcoach said:
which bits do you mean (I can simplify the language if necessary)
Click to expand...

Ok so re-reading it a couple time helped alot. I'm not very good in English, so bare with me!

So the compressor is used to tame the signals (whether it's the lows or the highs) louder than the threshold and leave the softer signals untouched. The gain reduction is the number of decibels that got cut down of from the louder signals. Is that correct?

Also I don't understand this:
what a compressor does allow you to do is tame the signal and therefore increase the fader/gain of that sound in the mix without it clipping
Click to expand...

What do you mean with "increase the fader/gain of that sound"? I thought it was lowering the volume of a sound (because the louder signals got cut down a bit), not increasing it. (I probably just don't understand what you said)
 
Last edited:
maximedupre said:
bandcoach said:
which bits do you mean (I can simplify the language if necessary)
Click to expand...

Ok so re-reading it a couple time helped alot. I'm not very good in English, so bare with me!

So the compressor is used to tame the signals (whether it's the lows or the highs) louder than the threshold and leave the softer signals untouched. The gain reduction is the number of decibels that got cut down of from the louder signals. Is that correct?
Click to expand...

essentially, yes!

Also I don't understand this:
what a compressor does allow you to do is tame the signal and therefore increase the fader/gain of that sound in the mix without it clipping
Click to expand...

What do you mean with "increase the fader/gain of that sound"? I thought it was lowering the volume of a sound (because the louder signals got cut down a bit), not increasing it. (I probably just don't understand what you said)
Click to expand...

After compressing a sound it may be lower in overall volume than you would like. You can address this by

1) using makeup gain in the compressor
2) raising the channel fader (aka channel volume)
3) adjusting the gain structure in other ways such as using a gain insert (as logic calls it)
 
Last edited:
Hah! The compressor is pretty clear in my mind now. Also, your website is really helpful for learning music theory :). Thank you!
 
Compression is an effect that gives amplitude to a track that would otherwise have a low volume. This is what I understand of it so far. As a sound wave the crest and trough may reach "clipping" limits, in other words it may "distort" and sound trashy. Well, compression allows you to increase the AMPLITUDE of the sound wave without the distortion. This will allow you to bring to the forefront any sound that you wish in the mix. The most common that I hear about is the snare drum. You want it to crack, and make an impression, but you don't want to have it overwhelm everything else. Compression gives a track "punch", and that is what I have learned about it so far.
 
Increasing the amplitude only comes after the compression phase of the application of this device.

Not all compressors allow you to increase the amplitude, i.e. makeup gain is to present on all devices though it is on many.

A compressor is used to reduce the signal level first, it compresses the signal range from it's full, untouched state, into a lower overall amplitude range.

The problem with this is that it may reduce the perceived volume of the signal. So we need to boost the overall amplitude of the signal, bringing the low level signal up and bringing the high level signal back to where it was before compression was applied

This is true whether it is a full-range compressor or a multi-band compressor - a multi-band compressor simply uses crossover technology to split the input signal into different frequency bands that can be compressed and have makeup gain applied differently.
 
I'll try to explain how a compressor works mathematically because words change person to person, language to language, but math is universal.


Lets say you have the following unit-less values representing the volume intensity of a track at a certain point, taken at 5 different times.
12,7,9,11,10

The mean of these numbers is
9.8
The standard deviation is: (standard deviation is the square root of variance which is the square of the difference between each value and the average over the average)
4.1

Now if we were to apply a compressor with the following setting:
-threshold 7
-ratio 4
This is what would happen


12,7,9,11,10
- (7) <- This would be your threshold. We want only values that are above the threshold to be affected.
________________
5,0,2,4,3



5,0,2,4,3
________
4 <- This would be your ratio

=1.25,0,0.5,1,1.33



1.25,0,0.5,1,1.33
+ (7) <- Here we add back the threshold value that we originally subtracted
_________________
8.25,7,7.5,8,8.33



After the compressor has done it's main work this is what you'd be left with for your new volumes:
8.25,7,7.5,8,8.33

The mean of these new numbers is:
7.8

Therefore we will increase the output gain by 2 so we have the same mean as the original.
8.25,7,7.5,8,8.33
+ 2
___________________
10.25,9,9.5,10,10.33


Our final values are:
10.25,9,9.5,10,10.33


now if we take the standard deviation of these new numbers we will get:
1.9
Which is less than the initial S.D of 4.1 meaning that the difference between individual values is now far less.
The higher the ratio, the smaller the standard deviation becomes.

Math is everywhere.
 
You must log in or register to reply here.
Back
Top