What Is Compression?

If we ever take the time to look at the waveform of our audio recording, we'd notice that it is made up several peaks and valleys of varying size - this is why it's given this name. Regardless of whether we've recorded gentle vocals on a RØDE microphone, or pounding, loud drums, there will always be a natural range in volume, from the softest recorded sound to the loudest. This is known as the dynamic range.

Dynamic range is what makes the recorded sound both natural and realistic. If we can manage to record a precise, exact replica of the original sound source, this is, of course, a very good thing, and commendable. Sometimes, though, we may wish for our track to be a little more consistent, with fewer volume changes. For example, a multi-track session is far more difficult to mix when each of those tracks have fluctuating volumes. Additionally, a presentation would be far more difficult to follow, perhaps to the point of distraction, should the speaker's voice volume intermittently change.

The Great Compression

To address this, and to help us achieve a more volume-consistent audio recording, it's necessary to apply a form of dynamic processing known as 'compression.' In simple terms, compression reduces the dynamic range of an audio track by lowering the volume level of the loudest parts, closer to the softer segments of the track. This will result in a track that has fewer variations in volume, from the microphone to the end product.    

At this point, we can raise the overall volume of the newly compressed track, which will lend the impression that the softer parts of the recording have been raised in volume - matching the louder sections.

Compression gives us more control of the dynamics range of our tracks, which can make it easier to mix together with other pieces. Additionally, it gives us more overall consistency within the recording.

How does compression work?

Now that we know what compression does, let's take a look at how it actually works. For example, how does the compressor 'know' when soft sounds begin and loud ones end (and vice-versa)? Perhaps we don't want to compress the whole track, just certain parts of it. How can this be achieved? Read on to find out!

The following parameters are commonly found on most audio compressors, giving complete control over what we are compressing, and how much we are compressing it by:

Threshold

The threshold sets the decibel (dB) level at which the compressor kicks in and starts to reduce the dynamic range. To illustrate, if we were to set the threshold to -15 dB, the compressor would not activate until the audio signal reaches and exceeds that number. Once this level is surpassed, the signal will be compressed by a certain amount.

Ratio

When we speak of 'ratio', we refer to exactly how much we want the audio to be compressed above the threshold level. The higher the ratio, the more compression will be applied. Let's take a look at a few core ratios:

•  The lowest possible ration is 1:1, which mean that no compression is applied whatsoever.
•   A ratio of 2:1 means that for every 2 dB the signal rises above the threshold, the compressor will deliver just 1 dB of output.
•   A 4:1 ratio means just the same, only in higher terms. Hence, for every 4 dB that the signal goes above the threshold, the compressor will give 1 dB of output.

Attack

The 'attack' refers to the amount of time that it takes for the compressor to react to the audio signal, and kick in to the maximum ratio once the threshold has been passed. The attack is generally measured in milliseconds (ms).

If we set a fast attack, we'll ensure that any spikes or transients of the signal are fully compressed in an instant. A slower attack will allow these spikes and transients through, but the audio that follows will be compressed.

Release

The 'release' is the time that it takes for the compressor to stop once the audio signal has fallen below the threshold. A quick release is not something that we should aim for, as it can sound wholly unnatural - the audio signal can drop off very quickly. Therefore, a slower, gradual reduction will perhaps be more pleasing to the ear.

Make-up gain

Once the audio signal has been compressed, the overall volume of the track will be lower. Make-up gain is applied to raise the all-around level of the track, so that it will sit naturally within the mix. Ultimately, the recording will have a smaller dynamic range and a more consistent volume.  

Now that we've learned a basic understanding of the general rules of compression, we can apply them to our recordings as two distinct tools. We can use compression in a correctional manner (to tame volume fluctuation, for example) or a creative device (to bring all the levels of a drum kit closer together, perhaps). 

As with most audio recording techniques, there are no correct or incorrect methods, only guidelines. Why not experiment with your own settings, be creative and use trial and error to find the sound that works best for you and your project?