The Big Squeeze – Compression Illustrated
There is no magic formula that can be applied to every track to achieve the perfect mix; however, there are a few elements of the process which are consistent and applicable to pretty much every track.
I have been an amateur music producer for nearly a decade now. From initial learnings on an 8 track digital recorder used in our schools Music Department to Audacity, REAPER, Cubase and for the last number of years, Ableton Live.
When it comes to mixing and producing music, there are a huge number of variables at play. There is no magic formula that can be applied to every track to achieve the perfect mix; however, there are a few elements of the process which are consistent and applicable to pretty much every track. This includes the holy trinity of mixing: EQ, Reverb and Compression.
What the heck is Compression?
“Compression = Tradeoffs. Lots of tradeoffs.”
Most people have some knowledge of Reverb and EQ; Reverb describes the effect that arises as a result of sound reflecting around the physical space in which the audio source resides (Most of us will simply see this as “More Reverb = More Echos”) and EQ refers to the equalisation of the track – how loud the bass/middle/treble frequencies are. Easy Peasy.
This leaves Compression; a key piece of the mixing puzzle and a facet of music technology that has been used in literally every single track that has graced the charts for the last 50 years (and beyond!) – So why do normal folk not really know about it? And why do amateur producers like me sometimes find it difficult to understand and use properly?
Without getting too bogged down in the details, compression is an audio effect that narrows the dynamic range of an input signal by a certain amount based on a set of predefined threshold (usually the volume of the input signal). In short “squishes” the peaks of the audio signal (more on that later) which enables greater dynamic control over a track which down the line can lead to a louder and higher quality overall mix.
In this article, I’d like to talk a bit about compression, explain what it is, how it works and then illustrate how the different parameters affect audio waveforms. This is going to be a little nerdy, but I hope we can embark on this journey together.
Ableton Live Compressor Plugin
The image above shows the default compressor plugin installed in Ableton Live 9. While you can get hardware compressors, this software compressor is ideal for the purposes of this article as we will be able to view our waveforms and see the effect of compression on our track.
There are a few variables in a standard compressor:
We can split the above factors into two main groups: transitionary and steady state. In this scenario, transitionary refers to the transition between the compressor turning on and turning off, whereas steady state refers to what the compressor does to the audio after it has been engaged.
Steady State properties:
The Steady State
Source: Universal Audio
The above image shows us how the compressor works in the steady state: As the input db rises, eventually a threshold is met where the compressor kicks in. After this point, increasing the input volume will lead to a proportionately lower output volume. The Ratio of the compressor controls the output volume and in the graph above is visualized by the gradient of the line after the threshold is reached.
For example, if we had a 5:1 ratio and a threshold of 3dB. If the input signal is louder than 3db, any volume above that will be reduced by 80%. The Knee describes how harshly this transition takes place, with a “hard knee” referring to aggressive onset of compression and “soft knee” providing a more gentle approach. It is important to note that in many cases, the compressor doesn’t “kick in” immediately after the input signal level exceeds the threshold (nor do we want it to, but more on this point later) and we actually have control over the time it takes for the compressor to complete the transition of OFF to ON (and vice versa) – this leads us nicely into our transitional properties; Attack and Release.
Attack refers to the amount of time it takes for the compressor to react to a transient (see below picture from ask.audio) or input signal. It’s a measure of how much time passes before the compressor kicks in and reduces the loudness of the signal (assuming that it is exceeding the threshold that we discussed above). Release on the other hand is how long it takes for the compressor to release the effect once the signal level drops below the threshold. If we think of a bass beat from a club, getting these times right is essential to ensure an energetic mix; we don’t want our release time to be so long that it bleeds into the next beat and we don’t want to “miss the beat” by having a long attack time. Having said that, in certain cases, we can go the other way and have attack and release times that are too short and create very artificial, pumpy and lifeless sounding mixes.
Let’s try to show the principles we talked about above in “real life” using waveforms. Below you will see a section of a guitar part that I recorded. There are 3 full “strums” here, the first is an upstroke, the second a downstroke and the third, another upstroke. As you can see, the second stroke is significantly louder than the other two. Here are their peak loudnesses relative to 0dB:
First Stroke: -14.30dB
Second Stroke: -2.61dB
Third Stroke: -10.16dB
Raw Guitar Signal: No Compression.
To start with, let’s see what happens when we set the threshold of our compressor to -10db, let’s set a fairly high compression setting of 5:1 so we can see the results. We are expecting to see a large reduction in the second strum, a small reduction in the third strum and no reduction in the first strum. We don’t really care about Attack and Release so we will just set them to as low a number as possible.
Raw Guitar Signal: 5:1 Ratio, -10dB ratio
The above image shows our compressed waveform. Let’s use an imagediff tool to see the difference:
Red: Uncompressed Signal. White: Compressed Signal
In the above image, the red waveform is the uncompressed signal and the white waveform is the compressed signal. It is clear that the compressor is working as expected and the high peaks that can be seen on the second strum have been significantly reduced in loudness. This demonstrates two of our variables: Threshold and Ratio. Where Threshold dictates the point at which the compressor engages and ratio dictates the severity of the dynamic reduction.
Now let’s play with the Attack Settings. We will do a comparison between four scenarios:
Scenario #1: Attack 0.01ms (Very Fast!)
Scenario #2: Attack 1ms Fast
Scenario #3: Attack 10ms Reasonably slow
Scenario #4: Compression disabled.
Images from left to right: Attack 0.01ms, Attack 1ms, Attack 10ms, Uncompressed Signal
The four images left to right show the four scenarios in order. What do we see in these waveforms? The faster the attack, the quicker the onset of “squished waves”: A 0.01ms attack will crush all of the peaks that exceed the loudness dictated by the threshold, but as we increase the attack time, we see more dynamic range get introduced to the signal until finally, we see the full dynamic range in the original uncompressed signal on the right.
In changing the Release time we see a different phenomenon; a slow release time causes the compressor to remain on after the initial transient, causing some additional dynamic reduction even after the signal level drops below the threshold. An extremely fast release will turn off the compressor like a switch once the signal level drops below the threshold. The image below from masteringthemix.com illustrates this excellently.
Effect of varying Release Time. Source: masteringthemix.com
Now that we have seen some waveforms, let’s ask the real question…
How does this apply to music?
“If we are mixing a track for the club, we don’t really care about the dynamic range of the kick drum; just hit it hard, all the time!”
As mixing engineers, we want good control of the dynamics in a track because this gives us more control over the overall track loudness and can help us mix the instruments together. However, if we reduce the dynamic range too much then we risk making the track sound artificial.
For instance, earlier in this article we talked about the meaning of Ratio and Attack; we saw that a high ratio and fast attack led to heavily and rapidly compressed (or squished) waveforms once the signal exceeded our threshold level. In a way, this is good as we have a huge amount of control over the instrument loudness, however, too severe a reduction in dynamic range can make the instrument sound artificial, which we also want to avoid.
If we go the other way with a very low ratio and slow attack, we have more dynamic range, but less control, which makes it difficult to manage the volume of the track in the mix and thus control the overall track loudness.
To compound the problem, we want to apply compression differently for each instrument, for example, if we are mixing a track for the club, we don’t really care about the dynamic range of the kick drum (hit it hard, all the time), whereas if we are mixing a piano sonata, the dynamic range plays a huge role in capturing the emotion of the music.
Last but not least is Release, which is another really important control on the compressor – if we have a release time that is too rapid, we get that pumpy, artificial sounding mix, which is ideal if we are talking about our club kick drum but is positively disastrous for our piano sonata. Similarly, if the release time is too slow, our kick drum will sound loose and lose a lot of the punch and impact as the compressor will not disengage properly after the initial transient. With classical music, however, this might work a little better as we can capture the character of the notes as they decay after the initial transient.
Well, that about wraps it up for Compressors 101. I hope that you have found this interesting (or at least informative). In summary, I think that compressors bring a huge amount of creative control to the table of the mix engineer; it also raises some tough questions surrounding control and musicality, the big one, of course, being the choice between dynamic range and track loudness.
In short, if you take only one line away from this article, let it be this:
Compression = Tradeoffs. Lots of tradeoffs.
We live in a fascinating world of music technology; it is a field which is constantly and rapidly developing and innovating and it can be completely overwhelming, with thousands of plugins and hardware options available to do all sorts of signal processing. This is why I believe it really helps to get back to the basics. This article started when I opened a simple waveform to experiment with the stuff that I thought I knew and now that I’ve taken a close look at the waves, I feel I know a little more.
Want to get started with Audio?
Written by Stephen Brennan