|
|
|
HOME
RECORDING |
August
1, 2001 "BEST OF" ISSUE |
|
|
|
A Compression Kickstart
|
|
BY KEN LANYON (WITH ADDED CONTENT BY BOB
DENNIS)
|
Originally
published by REQ in January 2000, this article explains the
complex subject of compressionin terms anyone can
understand. This is Ken's first article for REQ - Bob
|
|
|
| This article was originally written by Ken Lanyon and
posted to The Recording Website.
Taken with the article, Bob Dennis wanted to use it as a guest article in REQ and to
provide additional information on the subject. About
the authors |
The Basics
| Compression has to be one of the most confusing and elusive effects out
there. It's easy to know you need it just by watching your meters, but what does each knob
and button really do and how does it all work? This article should answer those questions,
and will touch on the "whens" and "whys" of compression. |
| Let me first start by explaining the basics of dynamic range in
recording. First, we have the noise floor. This is the lowest level, where tape hiss and
electrical hum reside at. Next we have the nominal level, which is the level that is best
for recording your incoming signal in order to minimize distortion and overcome the noise
floor. The distance between the noise floor and the nominal level is called the
signal-to-noise ratio. Next is the maximum level, which is where distortion occurs at when
your incoming level reaches it. This is the highest level in the total dynamic range.
Distortion is something that you definitely want to avoid unless you are versed in the
skills of good tape saturation (sometimes engineers will try to slightly distort the
signal by pushing it over the maximum level because this will give a stronger sound to an
originally weak one). However, in digital recording, any distortion due to overpeaking is
distasteful. Now the difference between the nominal level and the maximum level is
referred to as your headroom. This is your safety zone, and this is needed to account for
some stray peaks here and there without hitting the maximum level. And to wrap this
up...the whole thing, from noise floor to the maximum level is called the dynamic range. |
| Okay, lets cover how compressors work. Imagine a recording scenario where
you are starting to record some tracks on your multitrack recorder. You have set a good
recording level for your instrument which is at or near the nominal level, but you notice
that the incoming signal occasionally jumps up into the red. That is typically going to be
the nature of either the instrument, your playing, or both. So, you don't want those
distortions going to tape and ruining an otherwise fine performance. This is where the
compressor comes in handy. |
| The Alesis Company a while back issued a brochure on how compressors
work, and it gives the analogy of the compressor acting like your own dedicated engineer
for that one track. It will monitor all the incoming signals and then act like it is
pulling down the fader the instant that high volume peak occurs. In a more technical
explanation, what the compressor is actually doing is reading the incoming signals, and
then according to the compression ratio that you set, it knocks the hot signal down by
that ratio. This allows you to keep the level down to one that is manageable and
recordable, without the wild peaks. [See link at end of article -ed.] |
| Compression ratio you ask? Well, let me explain the 5 main controls.
First, we have the threshold. Think of this as the decibel level where the compression
will start working. I visualize the threshold as a line that is lowered onto the offending
noise peak, and the lower the threshold level, the more the incoming signal will be
compressed. This is because more of the noise peak is now ABOVE the threshold level so
there is more to squash. |
| Next we have the ratio settings. This knob has different ratios on it
like 2:1, 3:1, 4:1, and usually any combination in between. Okay, assume you set your
ratio to 3:1. What this does is that for every 3dB your incoming signal goes over your
threshold line, the compressor will allow only 1dB to pass. The level still goes over the
threshold, but assuming that you set the threshold low enough and used an appropriate
ratio, the peak will never reach the maximum level and distort. This is also due to the
amount of headroom you have. Typically, I set my ratio first, and then use the threshold
knob to find the point that the incoming levels are being compressed. This is done while
watching the meters on the mixer, and you will see the offending peaks all falling within
the same lower range which is nearer to the nominal level. Keep in mind that if your
incoming signal is lower than the threshold level, (or the threshold is set too high),
then none of the signal will be affected. |
| Next we have the attack parameter. Think of this as how fast the
compressor acts on the peaks once they pass the threshold. Some instruments have a really
quick attack sound as soon as they are played, and most peaks arise from this attack.
Therefore, on instruments like bass and kick drums, you would want to set a quick attack. |
| The release parameter works by setting how fast the compressor lets go of
the incoming signal once it has gone below the threshold level (where the signal doesn't
need to be compressed anymore.) You could set the release to fast and cut off a signal
quickly, or set it to slow which results in a longer sustain. Many guitar players use this
to sustain their notes. |
| The last main function is the output setting. Typically, when you lower
the threshold and the compressor kicks in to squash the signal, your nominal level will be
lowered slightly depending on the amount of compression being used. You can then use the
output knob to bring the input level back up to nominal. Be careful though, because by
raising your signal back to the nominal level, you are also increasing the noise floor due
to added noise from within the compressor itself. You may want to increase the trim on
your channel or master fader so more pure signal is getting to the compressor instead.
Everytime you patch your signal through another pathway (such as another processor), you
are also adding the inherent noise of that pathway. |
| There is one other feature that not all compressors have, and this is the
option to compress with "hard knee" or "soft knee". Hard knee is where
the signal is compressed the moment it goes above the threshold to the full extent of the
ratio that is set. Soft knee is where the compression is applied more softly so as not to
sound so abrupt. This is similar to using the attack knob, and I use hard knee compression
on signals like bass and kickdrum. |
| Hooking up a compressor is a simple procedure involving an insert cable.
This is a Y configuration cable with one 1/4" TRS connector that splits out to two
1/4" connectors. One of these connectors is an RS and the other the TS. (I should
mention here that TRS stands for TIP -RING-SLEEVE, with the tip being the send and the
ring being the return. This way, the TRS connector allows signals to go both ways, and the
TS connector allows on signals to send FROM the compressor to the mixer while the RS
connector returns the signal from the mixer to the compressor.) The TRS end is plugged
into the insert jack on one channel of your mixer, the TS to the compressor send, and RS
to the compressor return. This creates a loop where the original signal leaves the mixer,
goes to the compressor, is then compressed, and finally returns to the mixer. |
| As for using compression, that is a matter of personal preference. I use
it only when needed. Unless I am going for a certain type of effect by heavily compressing
the signal, then I use it only for stray peaks, since putting a signal that isn't peaking
through a compressor will only introduce more noise. Some people think that even though
the signal is peaking out during recording, they can compress the signal in the mix and it
will be the same. I used to think that myself but I realize now that when you put a
distorting signal to tape, the damage is already done to that signal's sound. The track is
already saturated with distortion and no amount of compression during the mix will make it
sound as if it were compressed during tracking. That is why you should definitely fix
stray peaks with the compressor when recording. Also, final mixes may also need a little
compression even if you used it on tracks during recording. This is due to the summation
of all the track signals. |
The following are just suggestions of where to start setting your
parameters for certain instruments. As I mentioned earlier, how YOU want to use
compression is your personal preference.
Bass |
Try starting out with a ratio of 4:1, and a fast attack and medium
release. I usually use the hard-knee type of compression here since bass is such an
attack-oriented instrument. But if you were playing smooth jazz bass, then you may want to
try soft-knee. It depends on the sound you are trying to get. |
Guitar |
This depends on the type of sound you are using, but a good general place
to start is 2:1 on acoustic, and maybe 3:1 on distorted guitar (although you may need 4:1
here.) To get a good sustain, try a 4:1 ratio, use a fast attack and slow release. Then
play the note you want to sustain, and raise the ratio until the sustain is as long as you
want it. |
Drums |
Drum signals are often compressed due to their hard-hitting attack
volumes. If nothing else, compress the snare drum, because each hit will likely peak
higher than other hits. Try starting out with a ratio of 3:1, and use a fast attack and
release. If the signal is still peaking, try using a ratio of 4:1. This method could also
be applied to the toms. As for cymbal hits, try starting with a 2:1 ratio (moving to 3:1
if needed), using a fast attack and a slow release (to preserve the natural decay time of
cymbals). |
Vocals |
As with drums, compression is also commonly used on vocals. The ratio to
start at varies for each singer, since some may be very strong and loud singers, and
others quieter, having a smaller dynamic range. Try starting out with a 2:1 ratio, with a
fast attack, and medium to slow release. Keep increasing your ratio until you get your
peaks under control. |
|
| Compression is not typically something that can be heard. You can hear it
if you really spank all the knobs to full-on, but usually that technique is used more for
an effect, rather than to control the level of the individual signal. Compression should
be applied and monitored by using the peak display meters on your compressor or mixer. As
I mentioned earlier, compression is something of an art, and you will have to play with it
to find your personal preferences, so don't be afraid to tweak all the knobs to find out
how they affect your sounds. Just remember that mastering compression techniques will help
to make all of your recordings sound more professional. |
|
|
"How Much" Compression?
| Remember that "the compressor [is] acting like your own dedicated
engineer for that one track. It will monitor all the incoming signals and then act like it
is pulling down the fader the instant that high volume peak occurs." In other
words it reduces gain on the high-level peaks. "How much" compression you
are using is determined by how much the gain is reduced on the peaks. After you have
set the ratio, attack time and release time, the threshold setting determines
"how much" compression you are using. |
| There is usually a series of lights (or possibly a meter) that tells you
how much the gain is being reduced. Having this meter read "-6dB" on the
louder peaks means the peaks are being brought down by this amount. This amount of
compression (6dB) is usually a good guideline. In more severe cases, or if you are
trying to get an extra-even track level, you could adjust your threshold to read
"-10dB" on the peaks. Beyond this point the instrument begins to sound
lifeless and you can really hear the compressor work. This is fine if you want that
effect, but generally I avoid this much compression. |
| Stereo Compression |
| Probably well over 90% of the CDs released have the entire stereo mix
compressed. A stereo compressor is two compressors working together, one for the
left signal and one for the right signal. But you really can't just hook your stereo
mix signal though two different compressors, because they have to "work
together." The stereo compressor always reduces gain on both the left
& right signals by the same amount, even when only one side has the uncontrolled peak.
If the two (left & right) compressors operated independently, it would sound
like someone was playing with the balance control. A stereo compressor will usually have a
"stereo" mode or a "link" button that will need to be activated.
"Link" simply means that both channels will be linked to work together,
rather than separately. If you are using the stereo compressor as two independent
compressors (like for two individual instruments), take the unit out of the stereo mode or
turn off the link. |
| The settings most often used for stereo compression of the entire mix
include a low ratio (2:1 or 2.5:1), fast attack, medium release and the threshold
control adjusted for about 6dB of gain reduction on the loudest peaks. Overall compression
of the stereo mix is usually sounds best with a soft knee (if available). |
| Compression of the mix makes the average level louder. You
apply gain reduction and then re-adjust your output level control for a good level,
causing everything to sound louder to the ear. Also the ear itself compresses really
loud sounds, so adding compression "sounds louder." |
| The Down Side |
| Compression sounds wonderful - right? Well, there can be
side-effects you should watch out for. A lot of these side-effects have to do with
background noise. When you adjust the threshold to cause compression and then boost
the output level to restore normal signal strength, you effectively boost the strength of
low-level sounds. Some of these sounds we don't want - such as shuffling of
the feet, air-conditioning noise or electronic hiss. If you are recording more than
one instrument at a time, the leakage
of the wrong instrument sound into the microphone is also an undesirable sound. If
you add a lot of compression for the super-even effect it causes, these background noises
can get really loud. Sometimes you can hear the background noise "pump,"
meaning you hear background noise slowly rise between the vocal lines (for example).
|
| These background noises getting loud is usually just a
sign that you are using too much compression and the solution is to use less.
Sometimes a bit of pumping can be corrected by adjusting the release time (usually making
it faster). Another solution is to use a noise gate after the compressor which
turns off the signal when the instrument isn't playing. Some (more expensive)
compressors have noise gates built in for this reason. |
| De-esser |
| When a singer sings a word with an "S" there is a blast of air
that rushes through the teeth, causing a high-frequency peak. These "S"
blasts can easily distort analog tape and also becomes annoying in digital
recording. Many compressors have a "Deessing" setting that can be used to
correct this. When a compressor is in the deessing mode, it becomes more sensitive
to the higher frequencies; in other words it will reduce gain more for loud
high-frequency peaks than it will for other peaks. |
| You say you don't have a dessing feature on your compressor? -
Never fear - how to "make one" out of a stereo compressor is listed in the
"Advanced Techniques & Tricks" given below. |
| Advanced Information, Techniques & "Tricks"
with Compressors |
| So you feel good about the basic operation of a compressor? You
want more? Click on the links below! |
|
|
Alesis Corp posting on the basics of compression is
available at http://www.alesis.com/support/notes/Signal_Processing/compbasc.html |
Copyright © 1999 by Ken Laynon & Robert
Dennis - ALL RIGHTS RESERVED |
USE OF THIS ARTICLE SUBJECT TO USER AGREEMENT |
"Alexander" is a service mark owned by Robert
Dennis for advanced web training |
|
Published by REQ with author permission |
|
|
|
|
|