what is audio compression

What is Audio Compression – Know right to Mix Right!

Recently, I went to Indore for my cousin brother’s wedding. It was a phenomenal week with the entire extended family around. There was music, delicious food, laughter, and frolic all around.

Indore is also my birthplace. A significant part of my childhood has memories of this city, so a heavy wave of nostalgia hits me whenever I visit. Also, my Nana ji (my grandfather who isn’t with us anymore), holds a very special place in this abundant pool of memories.

Now why am I telling you this? You’re here to know what audio compression is.

Well, I was trying to jog my memory and find the first instance when I observed volume control, and it hit me. It was during our family time in the evening, and this was back in 2005 (before Netflix). We’d all gather to watch a sitcom on television and every time the blaring ads came up, Nana ji would either reduce the volume quickly or just mute it right away!

It was manual volume control or in technical terms, known as manual gain riding.

Now if I had to define audio compression in three words, I’d say – Automated Volume Control. Because that’s what is does, right? It automates the volume control that was previously being done by the remote control, manually.

If you try to understand the meaning of audio compression in a technical way (right in the beginning), it won’t hold a place in your head but I’m pretty sure this story will.

Further in this article, you’ll get to know:

The REAL NEED of Compression

To understand the need for compression, we first need to understand:

What is Dynamic Range in Audio?

what is dynamic range in audio

Dynamic Range is the difference between the softest and the loudest element in a song.

Nana ji could mute the TV audio during the blaring ad and push the volume up when the sitcom came back but as a recording or mixing engineer, you can’t do that during a song, can you? You can’t mute the snare just because it’s too loud or you can’t ask the vocalist to sing loudly when it’s supposed to be sung softly.

Higher the difference between the loudest and softest element, higher is the dynamic range, which translates to uneven volume within a song.

Now to make the soft element audible, a basic solution is to increase its volume but at the same time the loud element would become even louder, right? And if that goes above 0 dBFS (audio ceiling in a DAW), digital clipping will occur and that is heard as a jittery, distorted sound which must be avoided at all costs.

To solve this problem, the concept of audio compression was introduced.

What does an Audio Compressor do?

The fundamental purpose of an Audio Compressor is to reduce the Dynamic Range i.e., reduce the volume of the loud elements and at the same time, increase the volume of the soft elements so that both can be heard clearly.

Now if anyone asks you what an audio compressor does, this is what you can tell them:

  1. It brings up the volume level of the soft elements and reduces the volume level of the loud ones.
  2. It reduces the Dynamic Range.
  3. It controls the volume fluctuation.

Check out this video to hear how audio compression works on a vocal sample and a basic drum loop…

what does an audio compressor do

When to use Compression in a Mix?

To understand this further, we first need to understand:

What is Crest Factor?

And to define Crest Factor, we first need to understand:

  1. What is Peak value?
  2. What is RMS value?
what is peak value
The dull green maximum reading is the Peak Value.
what is rms value
The bright green reading is the RMS Value.

To put it in simple terms:

  1. Peak Value is the highest reading on the metre.
  2. RMS Value is the average reading on the metre.

In musical terms, a simple example of Peak Value would be the transient of a snare.

Quick Tip: Can you guess which of these values do human ears perceive better? It's the average or RMS Value because it signifies the loudness of audio over a long period of time, and that is how we define an audio sample or a song, right? It's either loud or soft to our ears. We never speak of that momentary split-second peak.

Crest Factor is the difference between Peak and RMS value.

Higher the crest factor, higher is the Dynamic Range. And if the Dynamic Range is high, it roughly indicates that mostly soft elements are present, but at times, loud elements exist as well. In such a case, we obviously can’t push the volume up because of clipping. So the only option left is to use compression. Crest Factor directly indicates whether we should compress the audio or not.

Here's a short hack: If your Max Crest Factor is more than 8-10 dB, using compression on it becomes important.

This is a general guideline, not a universal rule.

This video has an excellent example of how vastly apart can the crest factor be in a violin sample and a drum loop, and how that helps you to decide whether you should use compression or not…

what is crest factor

How to use a Compressor?

Before we start exploring the parameters of a compression unit, we need to understand what a transient is.

Let’s take an example of a snare drum.

what is a transient

This is how the waveform of a single snare hit looks like. It must have taken place in a second but imagine if you heard it in slow motion or in this case saw the zoomed-in waveform, you’ll notice that the initial acoustic energy is quite high (grey rectangle) than the rest of it.

That initial high or the crack of the snare is what is known as a transient.

A metre hits the peak during a transient, any kind of clipping that may occur is mainly because of the transient.

With the help of compression, we attempt to duck (reduce) the volume of the transient and increase the volume of the sustain to create a balance. That’s the primary reason why a snare drum sounds different in a recording than in real life.

Moving on to the real deal – understanding the parameters of a compression unit:

compressor parameters

Threshold

It’s a pre-defined level at which the compressor comes into action, the point where the gain reduction starts to happen. In order to duck the transients automatically, we must define a level (threshold) by first looking at the metre, setting that value, and then tweak while listening to it.

ALWAYS REMEMBER - the metre is merely a tool for you to assess a reading but the listener is not going to judge it by that. That's why you need to follow a thumb rule - always make your last decision based on what you hear and not what you see.

Ratio

After setting the threshold, you decide the ratio – intensity of gain reduction. Suppose your signal’s transient peaks at 4 dBFS and you’ve set your threshold at 0 dBFS, a difference of 4 dB.

Now if you set the ratio at 3:1, the signal will reduce by 3 dB, i.e. it’ll peak at 1 dB instead of 4 dB.

Similarly, if you set the ratio at 2:1, the signal will reduce by 2 dB, i.e. it’ll peak at 2 dB instead of 4 dB.

Attack and Release

Now that we’ve set the threshold and ratio, we need to define how fast or slow is the gain reduction going to take place. That is set by attack and release, it’s mostly set in milliseconds.

Attack is how fast or slow the compressor kicks in, as soon as the threshold value is touched.
Release is how fast or slow the compressor shuts off as soon as the signal goes below the threshold.

These values are decided according to the transient and the sustain of the sample, also known as the envelope or the ADSR curve.

Make-up gain

We’ve controlled the peaks by now but the sustain still needs to go up. We have to create a balance, remember? Here’s where Make-up gain comes into the picture.

The gain that got reduced in order to control the transients will again be fed back so that the section with the lower volume can also be heard now.

This is the primary purpose of Make-up gain.

Conclusion and Quick Tips

  • Set the values by the metre but take a final call with the help of your ears.
  • If you’re compressing a long sample and the gain reduction is massive during a part or two, it’s better to manually adjust the gain of those parts and go for subtle gain reduction.
  • A faster attack on drums can deliver that punch but keep it too fast and it’ll totally take the transient away. Be wary of that because that’s a major reason for a dull mix.
  • Don’t set the threshold too low because that will make the release parameter useless. If the signal never goes below the threshold then the compressor is always at work and no matter what the release value be, it’s of no importance.

Check out my Ultimate Guide on How to Become a Music Producer in India to know more about a career in Music Production.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top