How to use an audio compressor?
You have undoubtedly heard of dynamics compression: it is an essential effect for mixing, just like equalization.
But do you really know what audio compression is for? Do you know how to adjust the attack, release, threshold, or ratio settings on your plugins?
Maybe not, and that’s normal: it’s quite a complex topic.
The problem is that if you don’t know exactly how each of these dynamics compression parameters affects your music, you risk deteriorating your mixes rather than improving them.
I therefore recommend taking a few minutes to follow this tutorial on the various settings of your compressors, so that you can adjust them yourself: whether you are using a mastering compressor or an analog audio compressor, whether you are working on Cubase, Ableton, or Reaper — the settings are generally always the same!
More concretely, here are the topics we will cover:
- What is an audio compressor?
- What is an audio compressor used for?
- Audio compression: the settings to know
- In the end, how much should you compress?
What is an audio compressor?
It is primarily an effect, just like equalizers, reverbs, distortions, etc.
It can take the form of a plugin or an external/hardware effect module.
Generally speaking, and although there are many possible ways to use a compressor, it allows you to reduce the dynamics range of a recording or a complete mix. That is to say, reduce the gap between the loudest and the quietest sounds in the track.
Hence the name: a compressor compresses the sound.
For example, if we have a vocal track with a significant level variation between words, we can level the sound by attenuating the louder parts.
Here is an example in images:
In the top image, there is no audio compression: the signal (the singer’s voice, for example) alternates between significant peaks and quieter elements.
In the bottom image, we used compression to attenuate these peaks. Indeed, they are now at a level closer to the rest of the recording. The dynamics range has therefore been reduced, and the sound is perceived more uniformly.
This is why we talk about compressor and dynamics compression.
By the way, I strongly recommend watching this video on the subject; it will be much clearer:
What is an audio compressor used for?
You might say: okay, but what is the purpose of a dynamic compressor? Why do we need it in a mix?
So, I propose a theoretical example…
Imagine you want to mix two sounds:
- An instrumental track (guitar, keyboards…) with a fairly constant level (
If you mix the vocals so that the quieter parts are audible above the instrumental track, then the louder parts will, in fact, be way too loud. They may even exceed the level of 0dBFs and saturate your speakers.
Conversely, if you optimize your mix so that the louder parts are at a correct volume, then the softly spoken words will be nearly inaudible among the denser sound of the instrumental track.
So what to do?
This is precisely where audio compression will help us, as it will control the peaks and reduce the level differences of the vocal track. It will then be possible to mix the two instruments in a much more pleasant way.
Let’s take an audio example (with the kind permission of James May, the composer of this piece titled “All Souls Moon”). Listen to the passage without and then with compression.
If you don’t hear the difference very well, pay attention to the syllables highlighted in orange in the lyrics below:
“This all souls moon confides
In both the worlds
This all souls moon abides both the worlds”In the uncompressed version, these syllables stand out too strongly, making it difficult to listen.
On the other hand, when compression is applied to the vocals, the sound level of the vocal track seems much smoother and more homogeneous, blending better with the rest of the instruments.
In short, here is a concrete example of using compressors.
Of course, there are many possible contexts, but one way or another, the compressor allows you to control the dynamic range and, very often, prevent certain parts of the signal from standing out too much compared to others.
Note also that there are multiband compression techniques, which allow for targeted compression of certain frequency bands. These are advanced approaches, but they can be very useful for solving specific mixing problems.
Audio Compression: The Settings to Know
Of course, to control this type of effects, you need to know how to properly set your dynamic compressor (and place it correctly in relation to EQ).
This involves adjusting a number of settings: the threshold, the ratio, the knee, the attack, the release, and the output gain.
Yes, that’s a lot of things :)! But we will go through all of this in detail right away…
The Compression Threshold
The Threshold parameter (seuil in French) is particularly important for successful compression.
It is simply the level in decibels (dBFS) at which the dynamic compressor starts to operate – in other words, to attenuate the signal.
For example, if your recording peaks at -3 dBFS and you set your threshold to 0 dBFS, then the signal will not be compressed. Indeed, the threshold is higher than the signal (0 dBFS > -3 dBFS). Conversely, if you set it to -12 dBFS, then the part of the signal above this threshold will be subject to compression.
In image form, it might be a bit clearer:
Only the part of the signal exceeding the threshold will be compressed How do you choose at what level to set this threshold?
It simply depends on what you want to do with your compressor.
If you want to only attenuate the peaks, then you will set it quite high. A bit like in the image above, in fact, even if the numbers will be completely different.
If, on the other hand, you want to reduce the dynamic range of your recording more broadly (to level a vocal track, for example), then it will be interesting to try a lower threshold, so that the compressor affects a larger part of the signal.
Note: the threshold level you set necessarily depends on the sound level of your raw track. As a consequence, it is impossible to give theoretical values that could correspond to a specific application.
The compression ratio
The Ratio is simply the parameter that controls the amount of audio compression applied to the signal that exceeds the threshold.
If the ratio is low, then the attenuation will be low. Conversely, if the ratio is high, the attenuation will be high.
In absolute terms, I could stop the explanation here: it’s almost enough to know that. But to illustrate in more detail what the ratio corresponds to, I suggest you stay with me for a few more lines. 🙂
Let’s take the example of a 4:1 ratio (that’s how ratios are written).
This means that if an audio signal exceeds the threshold by 4 decibels, it will be compressed so that it is only 1 decibel above the threshold. In other words, the level of the signal above the threshold is divided by four.
If the ratio is 8:1, then the level of any signal exceeding the threshold will be divided by eight. Et cetera.
So, if you see a 1:1 ratio, it means that no compression is applied: when 1 dB enters the compressor, 1 dB comes out. This is referred to as a unity ratio.
Similarly, an infinite ratio ∞:1 will indicate a limiter type operation: the entire output signal will be below the threshold, no matter what happens.
But then, how do you choose the compression ratio to apply during mixing?
Objectively, it really depends on the original recording: there is no simple method to apply. I recommend adjusting your ratio until it sounds good.
However, to guide you a bit, here are some indications:
- 1.5:1 to 2:1 — light dynamic compression. Typically the kind of ratio used during mastering.
- 2:1 to 5:1 — medium compression. Generally, it’s on this kind of values that you are likely to fall.
- above 5:1 — strong compression. The effect is quite pronounced. More for percussion or for voices with significant dynamic variations.
The knee
The knee is also an important parameter of audio compression, but it is sometimes a bit forgotten, a bit set aside, even though it truly impacts the behavior of compressors.
The knee is simply a setting that controls the transition area around the compression threshold.
It’s hard to explain the concept more clearly than with a diagram:
But then what is the impact on the sound?
Well:
- if the knee is “hard”, the effect of compression will be more easily audible: dynamic compression will either be activated or deactivated. There is no transition area.
- if the knee is “soft”, the effect of compression will be more gradual: it will therefore potentially appear more natural or musical (personally, I set my knee this way most of the time).
Furthermore, note that while some plugins simply give you access to a choice of hard or soft knee, others provide knobs that allow you to precisely adjust the knee from hard to soft.
Examples of knee settings on two different plugins The attack of compression
The fourth parameter to absolutely know is the Attack (attack in English).
It is basically the response time of the dynamic compressor.
And to be a bit more precise: it is the time taken by the compressor to compress the signal to the level determined by the ratio.
For example, if you have a ratio of 2:1 and an attack set to 10 milliseconds (ms), this means that the compressor, when it detects a signal exceeding the threshold, will take 10 ms before fully applying the 2:1 ratio (dividing the signal above the threshold by 2 — I hope you read the previous part about the ratio 😉 ).
An image might be a bit clearer:
Of course, this crucial setting will mainly depend on what you want to do with your audio compressor. If you want to compress the peaks of a signal, then a fast attack will be appropriate. Conversely, if you want to highlight these peaks by attenuating the signal once the peaks have passed, then a slow attack will be recommended.
Here are some examples of durations:
- < 5 ms — fast attack, transients are directly compressed
- from 5 ms to 30 ms — medium attack
- > 30 ms — slow attack, transients will be minimally affected by compression.
For more information and audio examples, check out this video:
The Release of Compression
Now we have the Release setting (which can be translated as release in French).
Like the attack, this is measured in seconds/milliseconds. It is the time taken by the compressor to stop functioning once the signal has fallen below the threshold. In other words, to return to the unit ratio 1:1.
You understand: if the release is not instantaneous, it means that your dynamic compressor can continue to compress the input signal even if it is below the threshold.
In image:
Consequently, it is of course important to set the release correctly. If it is too fast, you will generate a pumping effect; in other words, you will hear the compressor raising the sound level after having reduced it for the peaks. If it is too slow, the compressor will not have time to return to zero before the next peak, and your track will lose dynamic range significantly.
The notions of “fast” and “slow” can be a bit abstract, as they partly depend on the tempo of the music. That said, to give you a rough idea:
- < 100 ms — fast release
- > 300 ms — slow release
Make-Up Gain or Output Gain
Example of the Make-Up Gain setting of the Thillseeker VBL plugin And finally, the Make-Up Gain.
Sometimes simply called “Output Gain”, this is a setting that is also found on most compressors.
When we compress a sound, as we have seen, we attenuate a part of the signal. This means that, whatever happens, the output signal is weaker than the input signal.
The Make-Up Gain will solve this problem, as it will allow you to control the output signal level.
How to set it?
Everyone has their technique, in principle.
However, a good practice is to adjust it so that the average sound level is the same with and without the effect. In other words, if you disable the compressor, you should not hear a change in volume. The goal is to facilitate the comparison before/after compression.
The Sidechain
And finally, let’s look at the sidechain setting.
The sidechain is primarily a compression technique, but it is also the name of a setting commonly found on dynamic compressors.
When you place a dynamic compressor on a track, it will compress the signal of the track.
Logical.
But with the sidechain option, you will be able to trigger the compressor based on another signal, which can typically be:
- the same signal as the first track on which a filter was first applied;
- the signal from another track that has nothing to do with it.
A common use of this option is to trigger the audio compression of a track (like bass, for example) every time the kick hits, making it easier to bring out the latter.
For more information on the subject, I recommend reading my detailed article on sidechain.
In the end, how much should you compress?
It’s a big debate, and it would be quite inappropriate to give an absolute rule.
Depending on the style, the artist, the track, dynamic compression can be applied more lightly or more aggressively.
However, once again, it is important to have a rough idea.
Most compressors, whether plugins or external effects, have indicators of gain reduction (gain reduction or GR). This measure will indicate the degree of compression you have applied.
Generally speaking, try to stay below 6 or 8 dB of gain reduction. If you exceed this number, your settings may be too extreme.
Moreover, for mastering, many sound engineers tend to say that you should barely see the needles move. The associated gain reduction is then minimal, but effective enough to add some sonic “glue.”
On the other hand, some compression techniques such as parallel compression involve completely squashing the signal — a whole different story! 🙂
In any case, I can only advise you to experiment as much as possible.
Take a simple drum track, add a compression plugin, and vary the settings one by one. Take the time to identify the differences and variations based on the elements I summarized a bit earlier.
As you go along, you will see, it will seem easier and easier!
In conclusion
There you go, you now have all the necessary information to start using your audio compressors on your mixes, knowing what the different settings are for :)!
► Keep reading by checking out my similar article on equalization.
