What is compression?
Compression is a concept a lot of people struggle with at first.
Basically, a compressor reduces the dynamic range of an audio signal. In other words, it makes loud sounds quieter and quiet sounds louder. This makes it easier to mix the audio in with the rest of the track.
When is compression needed?
Let’s listen to this short vocal sample, while taking a look at the waveform it produces.
As you can hear (and see), there’s quite a bit of dynamic range in this sample. The word ‘lot’ for example, is much louder than the words ‘got a’.
Let’s say we were to try to mix this vocal into a track. If we turned the volume fader of the vocal up, so the quieter parts would all be clearly audible, then the loud parts would ‘jump’ out from the rest of the track. When we turn the vocal’s volume fader down however, the loud parts are now at a good volume in comparison to the rest of the track, but the quieter parts get drowned out by other sounds in the track. (leads, bass lines, effects, etc…)
This is where compression comes in.
The FL Studio Fruity Compressor
Let’s insert a fruity compressor on the vocal’s mixer track. To do this, click on the little arrow and select ‘insert’ ‘fruity compressor.
As you can see the compressor has 6 knobs : Threshold, Ratio, Gain, Attack, Release and Type.
As explained earlier, a compressor makes loud sounds quieter. This is how it works: when the volume of an audio signal goes over a certain threshold, the volume of the audio gets reduced. The part of the audio signal that doesn’t exceed this set threshold, is unaffected by the compressor.
For example, if we set the threshold at –3db (by turning the knob), everything above -3db will get compressed, when we put it at -12db, a lot more of the signal will get compressed.
So, we now know that sound that exceeds a set threshold, gets reduced in volume. The ratio will determine how strong this reduction is going to be. As you can see, by default the ratio on the fruity compressor is set to 1.0:1. This means that for every 1db the input level goes over the threshold, the output level will be 1db. In other words, no compression will take place. When we set it to 5.0:1, the output level will be 1db for every 5db the input level goes over the threshold.
I’ll give you an example. If our threshold is set at -10db, the input level reaches -2db and the ratio is set to 2.0:1,then every 2db that goes over -10db gets reduced to 1db. In this case, the audio exceeds the threshold by 8 db. The ratio is 2.0:1, 8/2=4 so after compression is applied, the audio level will only exceed the threshold by 4db. In other words, the original input level of-2db will have an output level of -6db. The loud part has now been made quieter, while leaving the quieter parts unaffected. Are you beginning to see the picture?
Let’s try this out on our vocal sample from before. As you now understand, we want to reduce the dynamic range of the sample, so there won’t be so much variation in volume throughout.
For demonstration purposes, I’m going to apply some rather heavy compression.
I’ll set the threshold to -20db and the ratio to 3.0:1. As you can see, the original sample is peaking at -4db, so the 16db that go over the threshold of -20, will be compressed. Everything below -20db will remain unaffected.
Because such a large portion of the audio is being compressed (and thus reduced in volume), the overall volume of the sample is going to drop quite a bit. We can compensate for this loss of volume with the gain (or makeup gain) knob.
After the compression has been applied, this is what the sample now sounds and looks like.
Now take a look at the uncompressed waveform from before again.
Overall, there is now a lot less difference between the peaks and dips in the audio.
There are still some small peaks in the waveform that the compressor didn’t seem to ‘catch’ however.
This has everything to do with the attack setting on the compressor. By default, the attack time on the fruity compressor is set to 15,0 milliseconds. This means that it takes the compressor 15 milliseconds to compress the audio input from the moment that it exceeds the threshold.
Remember, the compressor is only active when the sound goes over the threshold. So as soon as it drops back under this threshold, the compressor will switch off again. The amount of time for the compressor to go from active to inactive is determined by the release.
In our example the compressor didn’t activate soon enough to catch some little peaks.
If we want to get rid of these peaks, we would have to set a faster attack time. ( This isn’t always desirable. Sometimes you want the sound to have a sharp, punchy attack before it gets compressed).
Let’s set the attack time to 0,0 ms. Compression will now take place from the very instant when the threshold is surpassed.
As you can see and hear, this time the compressor did catch the peaks and they are now gone.
The audio signal we’re left with has a lot less dynamic range than the one we started out with, and will therefore be a lot easier to mix in with the rest of a track. No sounds are jumping out, neither are there sounds which will get drowned out.
The last knob on the compressor is ‘Type’. There are 8 different types of compression to choose from. Hard, Medium, Vintage, Soft, Hard/R, Medium/R, Vintage/R, Soft/R.
The main difference between these types is the ‘knee’.
The knee determines the db range, above and below the threshold, in which the compression changes from 1.0:0 to the set ratio. A hard knee setting for example, means that the compression will take place immediately after the threshold is exceeded, whereas with a soft knee, the compression is more gradually applied.
Play around with the different types to hear what they do.
In conclusion, a compressor is a great way to even out a sound and give it more punch.
But don’t overdo it. An over compressed audio signal won’t sound good at all.