Dynamic Range Compression:
Just The Basics

Dynamic range compression will make your songs.....quieter.

What??!!

Those of you familiar with basic audio wisdom tells you the exact opposite, so let me explain.

Dynamic range compression, also known as dynamic compression or audio compression, automates volume changes.

The analogy for compression is always given of the engineer sitting at the console with her hands on the fader. When the audio rises above a certain level she lowers the fader and therefore the volume, a certain amount for every increase in level.

This is why your song will be quieter.

And when the audio drops below that certain level, she moves the faders back to the original position.

Now that you understand that, here's how it can make your song louder.

First, though, here are five main terms associated with dynamic range compression.

Threshold is the point above which the audio volume is decreased. (The engineer pulling down the fader). In the digital world this is expressed as negative number such as -10, or -15 dB. It's amount below the maximum level that digital audio can put out.

Attack is the time it takes to lower the volume (or how long it takes the engineer to pull down the fader). This is expressed in milliseconds (ms).

• Short, or fast, attack times can generally be anywhere from 0 – 25 ms. Long, or slow, attack times are around 75 – 150 ms, or more. Longer attack times are most often used in mastering.

Ratio determines how much the audio is decreased (how the far the engineer pulls down the fader). Figures like 2:1, 4:1 etc. express the ratio.

• What it means is that, with a 2:1 ratio for example, for every 2 dB increase of the audio input above the threshold, there is corresponding 1 dB increase, only, of the audio output.

The behaviour of the compressor at this point is referred to as being non-linear. Below the threshold the compressor acts in a linear fashion, because the audio output is unchanged in comparison to the input.

(It's a whole other story, of course, if the software is also emulating a tube compressor, but for now let's keep it simple!)

Release determines how fast the volume (or the engineer pulling the fader) returns to its original level once it's below the threshold. This is also expressed in milliseconds.

• And again, short or fast release times, can mean anywhere from 0 – 50 ms, while long or slow release times can be 100 – 200 ms, or more. And as well, longer times are generally used more often in mastering.

Make-up Gain is where your song gets to be louder! The amount the audio is decreased above the threshold is often the same amount used to increase the overall make-up gain. All compressors give some kind of visual feedback as to the amount reduction being applied and this is where you get your starting point for setting your make-up gain. Your ears are always the final judge!

Dynamic range compression, if and when used in mastering, should, in most cases, apply no more than around 2 dB reduction.

Having longer attack times (when coupled with the right threshold and release times), lets more of the accents out before compression starts. This can make your chorus, which is usually louder than the verse, more dynamic and strong.

Sometimes a song just needs a little 'glue'! With a low threshold and a still lower ratio setting (like 1.1:1 or 1.01:1), this can add some warmth a song might need.

The terms I mentioned above were just the basics.

Different programmers can create various types of attacks and release; for example an attack can start very quickly and then taper off. The same type of action could be applied to the release but in reverse.

An Opto-compressor, so named after the hardware circuitry it emulates, adjusts its release characteristics depending on the amount of reduction, and time of reduction. This characteristic can be selected in Waves Renaissance, and Linear Phase Multiband Compressors.

You may have heard of the term 'knee' or 'soft-knee'. This simply means that dynamic range compression starts at specified amount below the threshold at a much lower ratio. And as the audio level rises towards the threshold, the ratio gradually increases to the amount finally set by you.

I've used the terms linear and non-linear: below the threshold the compressor is 'linear'; above, it's 'non-linear'. Some compressors, especially mastering compressors, become linear again at higher levels.

This would mean that, at the loudest parts of a song, the original dynamics would be preserved, after the make-up gain is applied; and would have an overall effect of bringing up and warming the mid-volume levels of a song.

Something similar to that can be achieved with parallel compression. This is where you mix in a heavily compressed version of your track with the original. Bob Katz describes the particulars of this approach in his book 'Mastering Audio”. When it's appropriate, it's one of my favourite techniques for achieving that 'mastered sound'!

The methods of dynamic range compression I've described fall the under category referred to as wideband compression. This means that the compression is being to all frequencies when it is in effect.

Multiband compression divides the compressor action to specific frequency regions. For example, a two-band compressor would work separately and differently on frequencies below 100 or 120 Hz, for instance, than on those above that region. This can actually give you some control over specific instruments in a mix, namely the bass and kick drum.

With three-band compression you could get benefits similar to, but not matching, opto-compression. You would set the ratios in the lowest band to accentuate the bass region; the mid-band to 'warm' up the frequencies where the vocal fundamentals generally reside; and leave the highest band and frequencies completely untouched. With the individual make-up gains of each band set properly, you can get a nice warm verse section, with a nice, tight but 'sparkly', chorus.

With multiband techniques, it's like getting two, three, sometimes more, dynamic range compression devices in one.

The possibilities are almost endless.

So where do you start?

You start by answering another question:

What does your song need?

Return from Dynamic Range Compression to
Signal Processing

Return to About Audio Mastering Software