Compression Fundamentals Course Notes

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Compression Fundamentals

Course Notes

What is a Compressor?
A compressor is an audio effect that makes adjustments to the level of a signal
over time. Before compressors, this would be the job of an engineer ‘riding a
fader’, waiting for moments when a signal got too loud and then turning it down
when that happened. Obviously, this isn’t exactly an ideal task for an engineer,
who could never be expected to respond quickly or consistently enough to ensure
there were never any signal overloads, so the first analogue (hardware)
compressors were born.

Over time, compressors became a standard feature on every channel of a mixing


desk, as they enabled engineers to make sure levels of multiple signals, e.g. all
the mics on a drum kit, guitarists, singers and so on, never exceeded a certain
point during a studio recording.

Today, you can still see compressors within emulation software that replicates this
kind of old hardware, be it complete mixers like that of Reason, or channel strip
(compression and EQ) software of Softube or Slate Digital.

Most commonly though, you’ll find compressors being added liberally as insert or
send effects in DAWs, so they can give producers complete control over levels
and help them to properly fatten up the sound of their mixes. There’s definitely no
shortage of compressor plugins, each with their own blend of new technology and
‘vintage’ processing, with built-in DAW effects and 3rd-party AU/VSTs providing a
wealth of options to the modern producer!

Page 1 of 7
What is a Compressor used for?
The main usage of a compressor is to ‘squash’ or flatten a signal, in order to
reduce its dynamic range, so the difference between its loud and quiet bits. This
might be a vocal or lead sound that gets too loud in certain places, or a drum beat
that has certain drums that are much louder than others. The compressor turns
down the loud bits of the signal, so that the levels are more similar throughout.
Then, once the signal has been flattened accordingly, the level of the whole lot
can be turned up safely, making the overall volume louder and more consistent.

Before Compression

• Large Dynamic
Range
• Big Peaks
• Low RMS Level

• Smaller Dynamic
Range
• Same Peaks
• Higher RMS Level

After Compression

Compressing sounds in this way, to make sure that they are consistently loud,
means that a producer has more control over the balance of a mix, so sounds
don’t disappear off into the distance when they shouldn’t for example. However, it
also means, when heavy compression is applied, that the overall loudness (RMS
level - more on this shortly) can be pushed up much higher, so that sounds
become a lot louder and therefore stand out a lot more. It’s for this reason that
compression has become such an essential tool, particularly in heavier styles of
electronic music, where insufficient amounts of compression will make a track
sound obviously weaker and thinner than other tracks in that style.

Another common usage for a compressor these days, however, is ‘sidechain’


compression. This functionality makes a compressor turn down a signal like
normal, only this time it does so in response to an external signal, rather than the
one being processed. In other words, when a certain signal like a kick drum
sounds, the compressor on a bass track will turn down the bass so that it ‘ducks’
out the way to allow more room for the kick. More on this in the Sidechain
chapter……
Page 2 of 7
How MUCH to Compress

How much you compress a signal has a big impact on the character of a sound,
affecting its natural dynamics and overall loudness. There are two main
compressor controls that define the amount of compression: THRESHOLD and
RATIO.

THRESHOLD

Threshold is the level


at which compression
starts, meaning
whenever the signal
exceeds that point, it
will be turned down.
Therefore, the lower
the threshold, the
more of the signal will
go over it and so, the
more compression will
occur.

RATIO

Ratio is the amount that the signal gets


turned down by when it goes over the
threshold, so the higher the ratio the
more the signal gets ‘squashed’. This is
defined by the input against output,
meaning a 2:1 ratio halves the signal
going over the threshold, whilst a 10:1
ratio turns it down by a factor of 10,
meaning when the input signal goes
over the threshold by 10dB, the
compressor turns it down by 9dB, so
the output is only 1dB over the
threshold. The max ratio setting is
infinity:1, which is a ‘brick wall’
compression or limiter setting, whereby
the output signal never goes over the Graph showing how the peaks in an audio
threshold, so is truly squashed! signal are squashed by compression

Page 3 of 7
How FAST to Compress

The threshold and ratio alone don’t control the amount of compression on the
majority of compressors though, as the speed of the compressor also has a big
part to play. Most compressors have additional timing controls that define how
quickly the compressor responds to the input signal. These controls are: ATTACK
and RELEASE.
signal before
compression Graph showing the attack and release
level times of a compressor

signal after
compression

threshold A
T
T
A
C
K RELEASE
time
ATTACK

The attack time sets how fast a compressor starts acting, meaning how quickly
compression starts after the input signal goes over the threshold. So, even with a
very low threshold and high ratio, which would normally compress a signal very
heavily, no compression will be heard if the attack time is too long for a rapidly-
changing input signal.

Setting a very fast attack is the best way therefore to make sure an entire signal is
compressed. However, the faster the attack, the more the natural transients will be
squashed, which can make a signal lose its punchiness or bite. Compression is
often a trade off between punchiness and fatness! Having a longer attack time
maintains transients but runs the risk of signal overloads if the compressor gain is
too high.

RELEASE

The release time can be a more difficult parameter to get your head round as it
defines how quickly a compressor stops turning a signal down when it goes back
under the threshold again. On a rapid signal like a drum beat, this controls the
‘sustain’ portion of the signal, in other words how fat each drum becomes, with a
short release boosting the sustain more and increasing fatness, whilst a longer
release maintains the natural dynamics more. In general, this is true of the release
time across the board, where a quick release tends to create a louder, more
aggressive sound, with a slower release being smoother and gentler.

Page 4 of 7
Peak and RMS Detection

Most compressors these days have the ability to be toggled between peak and
RMS detection. Peak detection means the compressor follows the maximum
signal level at all times, which makes it precise but not always so musical. RMS
(Root Mean Square) detection means the compressor follows the average signal
level, which is a better representation of its overall loudness. So, RMS detection
can be a little more sluggish but often much nicer-sounding, and will tend to
preserve transients more too.

The reason for transient preservation is that a very fast transient will cause a brief
spike in the peak level but won’t change the RMS level as much as it doesn’t last
as long, so a peak acting compressor will jump on the transient and squash it,
whilst an RMS one is more likely to let it pass through uncompressed.

One example of when you will definitely want to have peak compression is with a
limiter on the master channel, or any compressor that needs to ensure a signal
level doesn’t go over a certain point. This device will need to watch the peak level
closely, to prevent any chance of signal overloads. RMS detection might be the
one to go for if a compressor is squashing a signal too aggressively and you’re
struggling to get a good sound, or the signal is breaking up, as it will allow the
compressor to act in a more relaxed fashion and therefore have less of a drastic
effect on the sound.

Page 5 of 7
Using the Compressor Sidechain

Many compressors feature sidechain controls, which significantly expand the


feature set of the device. The sidechain of the compressor is essentially where the
signal that the compressor is responding to can be adjusted. In its default mode, a
compressor just responds to the signal being compressed, but accessing the
sidechain means you can do things like EQ that signal or even change it to a
different signal altogether, to dramatically change the action of the compressor.

Sidechain EQ

Adding Filters or EQualisation to the sidechain signal of a compressor can be a


great technique for reducing the kind of pumping that you get when compressing
drums, which generally have very loud kick drums. On the one hand, you might
want to compress all your drums together, but on the other hand doing so will
often mean the compressor is dominated by the loud kick, so will turn down
heavily every time the kick hits. However, if a high-pass filter is used on the
sidechain signal to roll off the bass, then the loud bass frequencies won’t trigger
the compressor anymore, which means the higher frequency sounds are able to
do so. This will produce a much smoother effect and won’t totally flatten the kick
drum.

Note: Sidechain EQ only filters the sidechain signal, so the filtering is not heard
outside of the compressor; the compressor still acts on the whole signal, without
any filtering, so even if you remove all of the bass from the sidechain signal, the
signal you hear at the compressor output will still have all its bass intact.

External Signal Triggering - Sidechain Compression/Ducking

Another option you normally find in the compressor sidechain is an input select.
This allows you to route external signals into the compressor, which can then be
used to trigger it rather than the signal that’s being compressed. The classic
example of this is making a bassline duck with a kick drum. This is achieved by
sending the kick drum to the compressor on the bass track, so that every time the
kick hits, the compressor turns down the bass. If wanting ultra-fast and
transparent ducking, just make sure the compressor has a fast attack and release,
with peak detection. Extending out the release, however, will allow a lazier and
more audible pumping effect, which can be desirable in certain styles of music.

Page 6 of 7
Multiband Compression

In certain circumstances, like when compressing more complex signals or when


you want to really maximise things, you may prefer to use multi-band
compression. Whilst using sidechain EQ allows you to have certain frequency
bands in the signal control the compression, the entire frequency range is still
compressed by a single compressor. With multi-band compression though, each
band in the signal can have its own compression settings.

Typically, a multi-band compressor has 3 or 4 compressors within it, allowing you


to compress bass, mids and highs separately. When 4 are available, it gives you
the option of splitting the mids up further into low and high mids, or the bass into in
sub and regular bass.

Although not always necessary, on some tracks a multi-band compressor can get
a much better sound, as it allows you to really get full control of the dynamics of
each area of the spectrum, where you might want to completely squash some
parts, like the bass say, but leave some areas, like the highs for instance, more
natural. A lot of multi-band dynamics processors also feature upwards
compression or expansion, which flip the compressor’s response so that it works
the other way and turns up the quieter parts of the signal to really maximise the
sound, which you can hear to great effect in Xfer’s OTT shareware plugin!

Page 7 of 7

You might also like