 |
Fundamentals of Compression
Well, it’s time to get all scholastic on you gentle readers of this column.
Yes, the following textual verbosity may be a tad drier than the moist,
wholesome entertainment you’ve become (perhaps unintentionally) accustomed
to; but listening to some of the demos and albums I’ve been sent for
mastering and consultations as of late, I think it’s high time to address
the old “sonic squeeze”.
Next to eq, compression may well be the most “go to” of all the available
processing choices, in the effort to make things sound “mo better”. The fact
that the results vary so widely depending the application, varying
parameters and the characteristics of the source material only serves to
underscore the necessity of utilizing it with a sure and practiced hand. You
could say it’s like cooking with salt; the right amount is the right amount.
Too little is still fixable, but too much is potentially catastrophic.
So what is compression and how does it work? Essentially, it’s a process
wherein the loudest parts of a signal are attenuated (a fancy word for
“automatically turned down”) to bring them closer in perceived volume to the
softer passages. Using the example of a bass guitar track, one might find
that while the overall performance is consistent, some notes are “jumping
out” at you intermittently. Judicious use of a compressor can tame those
peaks, helping the part to “sit” better within the mix.
The judicious part comes from knowing the various parameters of the
compression stage and how they affect the signal. Of course, all compressors
are not made equally. Many omit some features entirely, while others add
some esoteric ones. What follows is the most typical feature set.
Input Gain: This controls how loud the signal you’re working on will be.
You’re not really trying to do much here but make things loud enough for the
compressor to see, which is why many don’t even have this feature.
Threshold: Okay, now we start to “get it on”. The threshold is the point at
which the compressor starts to well…compress. Think of it like the height
line at amusement park attractions: everything above this line gets on the
ride, everything below it remains unaffected.
Ratio: This is how much you attenuate the signal, once it rises above the
threshold. A ratio of 2:1 means that for every 2dB of volume increase above
the threshold, the perceived volume increase is only 1dB. A ratio of 4:1
signifies an even more drastic reduction in volume, because for every 4dB
jump in volume, it’s being heard as a 1dB jump. Considering that 3dB is
essentially twice as loud as 1dB, a ratio of 3:1 renders your peaks half as
loud as they would be without compression. Compression ratios of 10:1 or
more (up to infinity:1) are actually called “limiting” and can be quite
drastic by comparison.
Attack: The time it takes for the compression to kick in. More accurately,
it’s the amount of transients you’re allowing to pass though before
attenuating the signal. On our bass track, we may want to adjust this so the
initial “pluck” is still heard before the fundamental note is subsequently
attenuated. This is what is often referred to as “punch”. Other times, we
may be trying to tame down some nasty “clickiness” in the note’s attack, in
which case we would go for the fastest attack possible, which can kind of
smooth over those transients.
Release: This controls how quickly the compressor “lets go” of the signal,
returning it to its normal, unattenuated state. You may have heard of
compressors exhibiting “pumping” or “breathing”. This refers to the volume
constantly rising up at the end of notes, as the compressor releases the
signal. Slower release times make this anomaly less noticeable, but a fast
release is sometimes used as a special effect as well.
Knee: There are two basic types of compression “knees”, which are labeled
simply enough “soft” and “hard”. Hard knees are easier to explain, in that
they work pretty much as I’ve described to this point; everything above the
threshold gets squashed the amount you’ve set your ratio to. Unfortunately,
this can be less “musical” sounding than soft knee compression slope. Soft
knee means that even though you’ve set your ratio to 4:1, signals barely
above the threshold may only be attenuated at 2:1, as signals rise further
above the threshold, the intensity of the compression gradually increases.
This makes for a less noticeable effect upon the signal. Used on vocals, the
net result might be described as more “natural” sounding. On bass drum
however, a more consistent leveling may be desired, so hard knee would be
the choice.
Output (or Makeup) Gain: Lots of resources refer to compression as “making
the soft parts louder”. Well, sort of. While it does make the louder parts
closer in volume to the softer ones, the bottom line is that all that
attenuation is going to make for a signal that is overall not as loud as the
unaffected signal. That’s where this comes in.
The wonderful irony of all this gain reduction is that it allows us to
increase a track’s level overall. In the case of our bass line, the notes
jumping out at the listener causes us to lower the level of the bass in the
mix. This makes us say “where’s the bass?” until the loud notes jump out, at
which point we go “Oh, there it is.” This can result in days in frustration
over the bass not being exactly right in our mixes.
Let’s say the peaks of those loud notes are at -1dB and the rest of the
track is hovering around -8dB. Set your threshold at -12dB, with an attack
of about 5 milliseconds. Set the release to around180 milliseconds. Try a
ratio of about 4:1 to start, with a soft knee and bring the output to where
the volume in bypass is close to the effected signal. This should put you in
the ballpark.
Look at your Gain Reduction meter. The average attenuation should be around
2 or 3 dB, with severe peaks at like 6dB. If you’re seeing it crunching
things 9dB or more, you may want to back off the input or raise the
threshold.
Similar settings can be used for other longer note-type instruments like
vocals, keys or strings but always, always, always use your ears and good
musical taste.
I personally use compression most consistently on lead vocals (5:1), backing
vocals (8:1), bass (4:1) and bass drum (6:1). Sometimes I compress banjos,
dobros and acoustic guitars a little (2:1). I almost never compress snare
drums (blindly compressing them should be outlawed) and less rarely,
electric guitars (the amp compresses it a lot already). In mastering I often
compress at a ratio of 1.65:1 with perhaps an additional 1.5dB of
post-compression limiting.
The ultimate goal is to tame these all of these elements so that they can be
utilized to their fullest musical potential within the context of your work.
Think of it as giving your tracks a nice warm hug. Who doesn’t like that?
Sven-Erik Seaholm is an award-winning independent producer with well over
200 recording credits, as well as a
singer and
songwriter. |
|
