[ddots-l] Re: garbage pure garbage
- From: Chris Smart <chris_s@xxxxxxxxxxxx>
- To: ddots-l@xxxxxxxxxxxxx
- Date: Wed, 21 Jul 2010 03:20:10 -0400
At 01:43 AM 7/21/2010, you wrote:
COMPRESSOR! HUH?! YOU GOT ME THERE. WELL DO I JUST LOAD
AN E Q OPTION
Comressor as in one of the effects included with Sonar.
just curious: are you using CakeTalking or JSonar?
Anyway, here's a great primer on dynamic compression, written by
Tom Kingston and originally posted to the midi-mag listserve. Hope
it helps you some.
Compression 101, by Tom Kingston
INTRODUCTION:
Hello folks and thanks for tuning in. This article contains a
comprehensive breakdown of what is meant by the term dynamic
compression and what it can do for your recordings, not to mention
your delicate musician's psyche. But be forewarned! Don't feel
discouraged if dynamic compression initially confuses you. It's one
of those things that seems relatively simple in theory, but can be
overwhelming and complex in practice. Confusion often results from
both the complex nature of sound itself, and how we perceive it,
not to mention the subtle and transparent nature of properly
applied compression. So don't be afraid to read this article
repeatedly while - and this is the most important part - you spend
some serious time in your studio doing compression tests,
evaluating and comparing the results of various compressor
configurations and applications. There are no simple rules, only
basic guidelines to get you going and some specific points to keep
in mind. Beyond that, it's all up to your ears. The bottom line is
this. Proper use of dynamic compression is the dividing line
between the sound of a professional recording and that of a small
project or private home studio. Even if you've got all the high end
gear, and your recordings are full, clean and sparkling, the lack
or improper use of compression is usually the root of that
intangible feeling that there's just something different about your
recording. It just doesn't have the smooth fullness of the
commercial CD you're using as a reference. Don't be discouraged if
your first few attempts deliver a sonic disaster!
DYNAMIC VS. DATA COMPRESSION:
Dynamic compression has nothing to do with the way audio files are
stored on a computer. The compression done on computers, for
archiving, like that done when converting a wave file into an MP3
file is called data compression. Due to the confusing use of the
term compression, some people refer to this as data reduction
instead. Its only purpose in life is to shrink the size of a file;
for example, compressing a 30mb file into a 5mb file. It works
under the premise of discarding or abbreviating the least valuable
bits in the data stream, or rather, the sounds your brain notices
the least.
In contrast, dynamic compression is a critical phase of signal
processing employed at various points in the recording process, to
raise the perceived volume level, or delineate more clearly the
place of an instrument in a mix. There is virtually no such thing
as a non-compressed recording on the market today. Dynamic
compression squashes, squeezes, or compresses (narrows) the dynamic
range of a signal. That is to say, it reduces the range of volume
changes put out by that signal. Visually, this narrows the swing of
audio level meters, while audibly, it reduces the amount of
perceived volume change between the softest valleys and loudest
peaks of a signal.
Think of a dynamic compressor as being a device that automates the
acts of an incredibly nimble imaginary engineer sitting at your
console. He has set an imaginary zone on the meters within which
your volume levels should stay in order to keep your sound smooth.
When you start your power ballad very softly, he will raise the
volume in order to boost your soft sound. But when you slide into
the chorus and crank it up, that engineer will pull the volume down
to make your overall volume changes less startling. Then when you
calm down and slide back into the soft verse, he will again keep
you balanced by sliding the volume back up. But don't get me wrong.
Compression does not replace real-time volume adjustments. It
simply reduces the amount of adjusting we have to do when mixing
and mastering.
If you want to experience a good example of recordings employing
little and lots of compression respectively, all you have to do is
listen to any orchestral recording, then listen to any heavy metal
tune, listening for volume changes,. Anyone who has ever listened
to much orchestral music has no doubt experienced the impulse to
continuously turn the volume up and down throughout the recording.
This is because the classical genre prefers a purist approach where
we hear pretty much exactly what the orchestra plays, wild dynamics
included. Softer parts are very soft, and louder parts are very
loud. In comparison, rock musicians prefer to be loud and compress
the heck out of their recordings. For example, while the tonal and
timbral changes remain intact, this results in very little volume
change regardless of what the singer is doing, whispering or
screaming.
Why? Because our perception of volume is tied more to the average
signal level rather than the actual hills & valleys in that signal.
Plus, at least in contemporary pop music, we perceive volume as a
constant thing, a seemingly quantifiable indicator of clarity and
overall quality. That's why the oldest trick in the book at hi-fi
shops is to play the speakers they most want to sell you at just a
little higher volume. Not much mind you, because it doesn't take a
large change in volume to convince your sub-conscious that those
speakers offer more clarity and definition. This is why
hard-rockers want their CD to be just a little bit louder than
others. If you drop their disk into your multi-disk player and hit
random play, their seemingly louder tune will sound just a little
bit better to many people.
But is this perceived volume increase an actual clarifier? Well,
yes. The reason is that a loudspeaker, with whichever high-tech
components it contains, is still a rather primitive device when
compared with the audio output of the real, or natural, world. No
loudspeaker can come close to reproducing natural sound efficiently
when trying to do so under the extraordinary burden of such extreme
dynamics like those found in nature. While the sound of a gentle
breeze is actually very similar to that of a gale force wind, that
is, when heard at equal loudness, the natural volume change between
them is enormous. No speaker can efficiently reproduce the entire
sweep of this dynamic spectrum. We could design a speaker specific
to each sound that would actually work pretty well, but it's
virtually impossible to design one that can handle such a spread in
volume.
As I said earlier, because our perception of volume has to do with
a sound's average level, dynamically compressing a signal fools our
ears into perceiving a volume boost that seems to solidify and
clarify the sound we're listening to because we're not asking the
speaker to work beyond its physical limitations. Therefore, while a
compressor is actually narrowing the spread of volume changes,
rather than simply cranking everything up, it is perceived as a
volume increase because it raises the average level of a signal.
And if we revisit for a moment the comparison between a rock band
and a symphony orchestra, we can again learn why compression is
more important to a rock band. This is because the typical mix of a
rock band is very congested regarding instruments that crowd the
same tonal frequencies. Guitars, piano and keyboards, and the voice
are all competing for the same space. Conversely, an orchestra by
definition is almost like a frequency map when you listen to each
section. From the lower brass and cello, up to the piccolo and
violins, they each have their own slot in the tonal spectrum.
Classical composers work with this in mind, being careful not to
have similar sections fighting for audible space.
When a mix is crowded, compression can help alleviate one of the
problems this tonal congestion creates, called masking. All masking
means is that when 2 signals are competing for the same frequency
slot, the loudest one will always prevail, seemingly eliminating
the other from the mix. And on very dynamic signals, this can be
very frustrating because the dueling pair will perform a maddening
dance in and out of the mix, depending on which one happens to be a
bit louder at any given point. Masking is why any soloed sound may
sound like it's just what you're looking for, but when you drop it
into the full mix, it seems to go pale and lifeless.
WHAT TO DO?
Stabilizing these signals with a compressor will give us the
ability to make the
adjustments needed to give each signal its own space. This also
involves reshaping the competing sounds with an equalizer,
reconsidering the timing of one, or simply re-writing the part for
one of them. When done properly, everything seems easier to hear,
with nothing buried or obscured.
But wait! there's one more reason why compression becomes more
necessary in a pop mix, the manner in which we record. Don't let
the purists down at symphony hall fool you. Even though they scowl
at compression as if it's the work of the devil, the way in which
they record their beloved orchestra employs natural compression.
Even though they may mic a solo instrument, or mic the orchestra by
section, they rely heavily on ambient mic's. These are mic's placed
at a considerable distance from the orchestra, be it overhead or
scattered out at various points in the auditorium. The purpose of
this approach is to capture the ambient sound you would experience
if you were in that concert hall. But the added benefit of space
between the sound source and the microphone is that it creates a
buffer zone that compresses the sound. Like with any other sound,
distance dulls the dynamics. So this natural compression is the
form they choose to use. But that's not to say that compression is
never used on such recordings either. It's just that engineers, for
the most
part, use very slight compression, just enough to nip the stray
transient peaks in the bud. Or they may reserve it for the final
mastering phase of the album.
But what the heck does that have to do with why pop music needs
more compression? Quite a lot, really. The way in which pop tunes
are recorded is the complete opposite of the orchestral scenario I
just outlined. People record everything direct or close up, so
there is no room for natural compression to have any effect. We
work in mere inches when placing mic's in front of a singer,
acoustic instruments, or guitar amps. We'll even mount microphones
literally inside of drums or pianos. And if it's not a mic we're
using, we plug the keyboard, drum-machine, or guitar amp directly
into our recorder or console. If it's a sound-card and recording
program you're using, you again have the same basic setup where
your sound-card outputs are plugged (routed) directly into your
virtual console. And most sampled sounds, be they from a keyboard,
sound module, or sound-card, are sampled (recorded) in a very
direct manner. So it's all the same no matter how you slice it. And
regardless of how you choose to look at this methodology, the
bottom line is that the dynamics and tonal characteristics are
amplified to the nth degree, due to something called the "proximity
effect". That is, because the mic is right near the sound source,
every little change in volume is tracked with much more precision
than when the mic is placed at a greater distance, hence the
natural compression effect.
Stepping into a studio environment offers 2 advantages. The first
is that because the singer is using headphones, the engineer can
crank up the mic gain and allow the singer to back off the mic,
expanding the pickup of the mic and smoothing out or lessening the
exaggerated dynamics. But even though this studio approach to
vocals is indeed an improvement over the live environment, it still
offers nowhere near the smooth response or natural compression
inherent in a mic placed 30 feet away. So our only choice is
compression. Oh yes, and EQ as well, but that's a whole other
discussion.
WHERE, WHEN, AND HOW DO WE APPLY DYNAMIC COMPRESSION?:
Wow! what a loaded question! And not an easy one to answer, but
here goes. Just don't forget, all of what follows is no more than a
set of basic guide-posts to get you going. There are no hard & fast
rules for compression. Trust me. If you had the priceless
opportunity to sit down and chat with 20 of the most sought after
engineers in the industry, and asked them each to outline their
compression technique and philosophy, I guarantee you would hear 20
very different answers. This is why you, when comparing 2 albums,
may prefer the overall sound of the first, while you actually like
the tunes more on the second. Naturally, this goes beyond
compression and into EQ and recording technique as well,
but compression is a big part of the overall feel of any music.
But that's enough with the psycho-acoustics for now! Let's move
on to the technical heart of the matter.
WHEN TO COMPRESS:
The basic rule of thumb is to employ light to moderate compression
on signals as they are being recorded. You just want enough to give
the recorder a good strong signal to work with and reduce the
chance of your soft signals flirting with the noise floor. But,
having said that, the more accustomed you become to compression,
the more you will discover how much you need and on which signals.
Over time you may be able to be more aggressive with compression at
the recording phase and reduce, or at least make easier, the next
phase, mixing. The closer you can get the incoming signals to their
final state, the quicker and easier your mix will fall into form.
And this will make your mixing chores considerably easier.
Which leads us to our next layer of compression, Mix-down.
Here we can do more compression to fit each track into the mix the
way we want it. Let's say that your vocals need just a little more
steadying, or the bass doesn't quite fill the groove the way you
want it to, or you'd like to just take the edge off of the guitar a
bit, or, maybe you want to push the drums a bit further back in the
mix. No problem. Compression can work all these wonders. So you
fiddle to no end and get everything sounding just the way you want
it. Now you mix down your entire tune.
But wait! We're not done yet. The next use of compression is part
of what's called
mastering. This, though usually done by a mastering house, can be
done to a lesser degree if you've got the tools to do it with. All
it amounts to, and that's the understatement of the century, is the
final phase of very discrete overall compression and EQ applied to
the entire mix. It usually employs what's known as Split Frequency
Compression. All this means is that the sound is broken up into
frequency blocks, like highs mids and lows, each of which is
compressed separately, because these different frequency blocks
react somewhat differently to compression, which we'll get to in a
bit. But suffice it to say that these frequency zones of your
material are compressed just a little bit more and then reassembled
back into the whole on the final master tracks to be used for
duplication.
OK, so that's a basic overview of what compression is, and when
compression is applied. Now let's get to the nuts & bolts of
actually applying it. First, I'll give an extensive breakdown and
description of the typical controls on a compressor, sometimes
referred to as a compressor/limiter. I'll then give some generic
examples of compressor settings for various signals.
But before we dive into this, there's one more thing I must first
clarify. And that's the metering system being used in my examples.
The digital console and recorder I use employs the digital metering
system which is a bit different from the VU metering of a standard
analog deck. It operates on a negative numbering system where 0 Db
indicates the clipping threshold of a signal. All this means is
that in practical terms, a digital
signal cannot go over that 0 Db peak limit. Unlike on an analog
deck, there's no such thing as soft distortion, the act of pushing
the signals so that they distort just enough to conjure up some
nice fuzzy warmth, or even a smooth audible distortion, common on
guitars. Digital systems are not capable of creating this type of
harmonic distortion in this manner. It's all a numbers game in the
digital environment, and the 0 Db peak limit is a non-flexible
absolute. Beyond this point, the actual wave form has its peak
clipped off as if trying to go higher but running into a wall. And
if you clip a large enough portion of a signal, it will scream at
you in pain. So this is the level I'm referencing when I speak of
any metering, such as, threshold and peak levels.
COMPRESSOR CONTROLS:
1. Threshold: Typical range: -60 to 0 Db.
This sets the level at which compression kicks in. When a signal
crosses over the threshold, the compressor takes notice and makes
its move in accordance with the other settings you've configured.
Average setting: -20 Db.
Tips: A higher threshold (-16 to -12) is used when all you want to
do is smooth out a signal, such as an entire mix. An even higher
(-12 to -8) threshold may be used when you want to simply grab a
signal, limit it at that point, and boost it up to a constant
level, sometimes used for bass guitar or vocals. A lower threshold
(-28 to -38) allows you to compress the entirety of a signal. This
for example has a lot to do with how breathy a vocal is, or how
loud fingering noise is, because the lower the threshold, the more
the lower or softer parts will also be compressed.
2. Attack time: Range: 0 to 250 milliseconds. Average setting: 10.
This determines how long a signal must stay over the threshold
before compression actually begins. The reason for this delay is
that we often want the leading edge, or peak transient of a sound
to have a chance to make its point before we call in the troops and
beat it back down. A slower attack time will allow sound elements
such as the initial strike of a drum or piano, the pluck of a
guitar string, or the emphasis of the voice to lead, or articulate,
the sound before it's compressed.
Tips: The lower the threshold, the more critical this setting
becomes because you're then working at the start level of a sound.
If you compress it before it has time to ramp up, you're going to
dull the heck out of that sound. And this may be exactly what you
want on a fat bass guitar or drum sound. But it's probably not what
you want on an acoustic guitar because it will dull the overall
clarity of the sound. This also has a much more apparent effect on
high frequency sounds, or the high frequency components of
otherwise low frequency sounds, such as the finger or fret noise of
a bass. Because high frequency components are usually the leading
edge of most sounds, subtle adjustments of the attack time can have
enormous effect on the perceived placement and sharpness of a sound.
For example, I was just recently playing around with the entire
setup for my snare drum, trying to capture more of the shell sound
out of the drum. And when I started playing around with the
compressor, I discovered a 5 millisecond window of complete control
over that drum. When I quickened the attack time, the compressor
held back just enough of the initial strike snap to allow more of
the body tone to come through. But if I really slammed down a quick
attack, I heard the drum move to the back of the studio as if it
had been pushed 10 feet back from its mic. So always play around
with the attack time.
On the other hand, longer (30 to 50 millisecond) attack times are
used when doing overall compression of a mix because it lets the
tune swell and sway within reason, just pulling in the reins when a
sustained peak comes, such as building to a chorus or finale..
For the most part, this is the one setting that should really be
played around with for each signal. It can easily make or break
your sound.
3. Release: Range: 5 to 2000 milliseconds. Average: 100. This
determines how long a signal must fall and stay beneath the
threshold before the compressor actually lets go of it.
Tips: Having too short a release time on a signal can cause a
pumping or breathing effect because the compressor is trying like
mad to chase a punchy signal, or it's letting go during slight
breaks in the music, bringing the noise floor up, then letting it
fall back down. . Too short a release can also cause distortion in
low frequency signals, such as bass guitar and drums. But the good
side of a shorter release is that it can keep the punch in the
music. A longer release, on the other hand, smoothes out the
overall flow of the piece, but at the expense of some definition.
So it all depends on the tune, or even the section of the tune
you're working on. For example, ballads usually use a longer
release while punchy tunes have a shorter overall release.
4. Ratio: Range: 1 to 1 up to 100 to 1 infinite. This determines
how much the signal is compressed once it hangs out over the
threshold for the duration of the attack time. The actual ratios
available may go something like this. 1 to 1, 1.5 to 1, 2 to 1, 2.5
to 1, 3 to 1, 4 to 1, 5 to 1, 6 to 1, 7 to 1, 8 to 1, 10 to 1, 20
to 1, 30 to 1, 40 to 1, 50 to 1, 100 to 1, infinite. For example,
If a 4 to 1 ratio is being used, for every 4 db your signal moves
over the threshold, the compressor will only allow it to move 1 db.
So if it peaks at 8 db over the threshold, the compressor will only
allow it to peak 2 db over the threshold.
Tips: In general, the lower the threshold, the lower the ratio, and
the higher the threshold, the higher the ratio. This is because
with a low threshold, you're compressing more material, some of
which may be noisy elements. A good example of this is the
breathiness of vocals and even the crackling of lips. While the
breath tones may be just what you're looking for, you'll soon
discover that it's all or nothing. Therefore, the unwanted pops and
even your movement in front of the mic will also be much more
apparent. So this usually requires considerable playing around with
the settings in order to find the best compromise.
Conversely, you'll want to hit a high threshold harder because you
have less room to play with on the meters. But all rules are made
to be broken. You may just want to soften up the peaks just a
little bit, so you use a gentle 1.5 to 1 ratio on them at a high -6
level.
But let's now look at the other extreme of ratio settings. In
practical terms, a 10 to 1 or higher ratio is considered a limiter.
In other words, it is limiting that signal, stopping it dead in its
tracks at the threshold level. This, as I said earlier, may be used
to grab a signal and crank it up to a rock steady level, by
limiting it let's say at -10 Db, then boosting it up 4 db to sit it
right there at -6 Db. This method may also be used to add just a
slight overall boost to a finished mix by placing the limiter at -3
db and boosting it up 2 db. Limiting however is a tough call. Some
people like it, and some think it strips the signal of its life.
The decision is yours.
Limiting can also be used as a protective type of compression. It
can be used on signals just to stop any stray peaks from ruining an
otherwise good recording, or, it may be used as an overload
protector for your speakers in a live environment. In either case,
you're simply placing a limit at the point beyond which you want no
signal to go.
And finally, back to the other end of the spectrum and the 1 to 1
ratio. This is a bypass ratio that makes the compressor do nothing.
It may be used for quick comparisons while working out your
settings, or, it may be used when all you want to use is another
feature of the compressor, such as, a noise gate or downward expander.
5. Gain: Range: 0 to 12 or 15 Db. This is how you compensate for
the gain reduction being done by the compressor. In general, it's a
simple corollary to the ratio you're using. For example, if using a
4 to 1 ratio, adding 4 Db of output gain will pull the compressed
signal back up to the correct level. Sometimes, particularly on
vocals, and when using a low threshold and ratio, even more gain
may be used to power up the vocals. Sometimes, this is done with 2
compressors in series where the second unit is simply a high
threshold limiter waiting to keep those power-house vocals in check.
But the most important use of the gain knob is simply to compensate
for the compression. No signal likes to be squashed, so forgetting
to compensate correctly will result in significant degradation of
signal quality.
OPTIONAL CONTROLS:
That's it for the basic controls found on many compressors. A few
additional switches may also be available.
1. Bypass: pretty self explanatory.
2. Stereo link: This allows you to control both channels from one
set of controllers. This is very important when compressing stereo
signals. Lack of in sync settings can make sounds wander.
3. Noise gate or Downward expander: Consists of at least a
threshold knob, and perhaps attack and release knobs as well. This
tries to cut low signal noise during silent parts. When a signal
drops beneath the threshold, a noise gate acts as a simple gate and
closes, basically cutting the signal off. Then when the signal
comes back above the threshold, the gate reopens and lets it come
through. The attack and release time work just like on the
compressor. The signal has to stay below the threshold for the
duration of the attack time before the gate will close and it has
to come back up and stay above the threshold for the duration of
the release time before it will actually be let through.
The difference between a gate and a downward expander is that an
expander provides a more smooth processing of the signal. The gate
in effect is just a simple switching device. It's either on or off.
But the expander employs a gradual slope to signal reduction. It
also has a ratio setting, so it works like a compressor in reverse.
This alleviates the potential audibility of a noise gate that
simply flips open or closed.
PRACTICAL EXAMPLES:
Generic vocals:
-28 threshold, 4 to 1 ratio, 0 attack, 500 release, +4 gain.
Power vocals:
-38 threshold, 2.5 to 1 ratio, 12 attack, 180 release, +5 gain.
Note that compressing vocals too much can be difficult for the
singer because they will not be able to hear their louder parts.
The voice in their head will drown out what they hear in their
monitors. So it is sometimes necessary to use minimal compression
while recording and add more during mix-down.
Guitar:
-20 threshold, 3 to 1 ratio, 20 attack, 200 release, +3 gain.
Bass guitar:
-24 threshold, 4 to 1 ratio, 10 attack, 250 release, +4 gain.
Piano:
-36 threshold, 1.8 to 1 ratio, 3 attack, 100 release, +2 gain. Note
that piano compression in particular seems to be very noticeable.
Obviously, some people don't mind this because I can hear it in
many commercial releases. I hear the strike, then it's as if it is
yanked back by the compressor. So that's why I use a
low threshold and ratio, with a quick attack. This lessens the
audibility of the compressor.
Drums: Very tricky! It would be easy to write an entire book on
drum compression. The problem is that the typical drum set includes
too many very different types of sounds, from the bass to cymbals,
all that react very differently to compression. And this is why
drums really need individual compressors. Bass drums need a longer
(200) release in order to prevent flutter distortion, but cymbals
need a shorter release (100) in order to prevent wavering of the
cymbal decay. Plus cymbals usually require the quickest attack,
while toms and the bass can tolerate flexibility in this
department, depending on what you're looking for. And the snare is
a whole other game, particularly when setting the attack time. This
can completely change the sound of that drum. And all this changes
depending on how much or less you're trying to compress the drums.
But for the sake of this discussion, I would assume that most of
you are probably not working compression on a drum-by-drum basis
anyway. So the best thing I can say is to just start your overall
drum compression with the standard -20 threshold, 4 to 1 ratio, 10
attack, 100 release, and +4 gain. Then spend some serious time
playing around with the attack time as this will have the most
profound effect. Then fiddle a bit with the release time, paying
particular attention to the low end of the groove. Then you might
even try lowering the ratio a bit while leaving the gain up. This
can sometimes pump up the sound a bit.
As far as full mix compression goes, this is a tough call as well
because it has a lot to do with the type of tune you're working
with. But here's a generic configuration to try.
-16 threshold, 3 to 1 ratio, 30 attack, 50 release, +3 gain. Note
that the low frequency sounds are not as susceptable to the
distortion problems caused by a short release when the threshold is
higher. But you should still keep an ear on this.
In conclusion, all I can do is remind you to take all of those
suggested settings with a grain of salt. The simple fact of
personal preference and opinion in sound makes it impossible to
define what the right settings are. Plus, there are many different
compressor designs in use these days, so they all tend to vary to
some extent on how they react under the same conditions. And the
final thing is that there is a line that has to be drawn depending
on the quality of compressor you're using. A high-end $2,000
compressor is going to be virtually transparent no matter what you
ask it to do. So the sky is the limit if that's what you have to
work with. But if it's a $99 compressor you're working with, be
aware of this, and be conservative on what you ask it to do. Heavy
compression with a low quality compressor can suck the life right
out of your otherwise great recording. So be careful and listen
hard. Do a lot of comparing and make sure that there is indeed a
sustainable improvement in your compressed material.
Good luck,
PLEASE READ THIS FOOTER AT LEAST ONCE!
To leave the list, click on the immediately following link:
ddots-l-request@xxxxxxxxxxxxx?subject=unsubscribe
If this link doesn't work then send a message to:
ddots-l-request@xxxxxxxxxxxxx
and in the Subject line type
unsubscribe
For other list commands such as vacation mode,
click on the immediately following link:
ddots-l-request@xxxxxxxxxxxxx?subject=faq or
send a message, to
ddots-l-request@xxxxxxxxxxxxx
and in the Subject line type
faq
Other related posts:
