Skip to main content

ok I'm sorry this is a horrible question..
but can someone kinda explain a compressor to me
ive never been explained what "ratio, attack, release, and "knee?" really do...
ive messed around with them on my pro tools le and use the foactory presests for a lot of things...(vocals, drums)
and i know it levels off my vocals tracks (i.e. when i start screaming he level stays the same *nice!haha)
but i really dont "understand" a compressor, meaning if i hear a track i dont kow what needs to be adjusted...i pretty much just play with the threshhold till i think it sounds better...
can anyone shed some simple light?
please...
thanks
-chris

Comments

Guitarman Thu, 08/14/2003 - 22:43

Hey vess,

Ok...here it goes. I will say this like I am talking to a six year old because this is how it was explained to me when I first started.LOL(And I still may not have it right exactly technicallywise,hehe)

You have a signal like your voice. Like you said when you scream the level stays the same. That is because your compressor takes the signal sees how loud it is going in. Then you set your ratio to say for every 1db(1:1) or 2db(2:1) etc coming in, the compressor increases the signal 1db(hence the :1).

Are you with me so far?

Now that you have decided that you can scream your ass off, you want the compresser to only do it at a certain level of your screaming. That is where you set your "threshold" -10db, -20db etc. SO, now when your voice gets up to the "threshold level" your compression will start to kick in.

How we doin? Still here?

Now that we have that set we want the compressor to either take its time applying the "ratio" you have selected or come in really fast. This is your "Attack".

Hey wake up!

The "Release" is pretty much what it is. It determines how long the compressor will maintain its hold on the signal.

Now the "Knee" is basically like a rubber band. It gives you a little flexiblity between your attack and threshold.(at least that's how I percieve it.)

The "Output" is the overall level of the signal you are processing.

Now considering it does all this and try to keep the original "frequency range" to fit in your controlled signal.

It's applications range from keeping the drums from spiking (limiting, just like when you scream) to keeping a vocal or bass track smooth and level.

That's the best I ca do my friend. I am sure many others here can elaborate much more than I. Hope it helps.

Best wishes,

JD( 0}===;;;

KurtFoster Thu, 08/14/2003 - 22:47

Originally posted by thevessels:
ok im sorry this is a horrible question..
but can someone kinda explain a compressor to me
ive never been explained what "ratio, attack, release, and "knee?" really do...
ive messed around with them on my pro tools le and use the foactory presests for alot of things...(vocals, drums)
and i know it levels off my vocals tracks (i.e. when i start screaming he level stays the same *nice!haha)
but i really dont "understand" a compresser, meaning if i hear a track i dont kow what needs to be adjusted...i pretty much just play with the threshhold till i think it sounds better...
can anyone shed some simple light?
please...
thanks
-chris

Chris,
Ratio is the amount of compression (attenuation) that is applied after you reach the threshold.. A 2 to 1 ratio means every 2 dB over the set threshold level, the audio will be 1 dB louder.. so if you set the threshold at 0dB and the level before it enters the comp is at +2 dB, at the output of the comp it will be +1db..

Attack is how fast or slow the compressor clamps down on the signal..

Release is how fast or slow the compressor lets up on the compression after the initial sound hits the comp.

Knee is how drastic reduction is applied. A sharp or fast knee will engage the gain reduction in a faster manner than a soft knee will. There is also gain reduction.. which is the amount of gain attenuation that is applied overall..

If you want a naturals sounding compression setting a 2 to 1 or 4 to 1 ratio , with a soft knee and 2 to 4 dB of gain reduction is good…

Ethan Winer Fri, 08/15/2003 - 04:33

Chris,

Below is my standard blurb.

--Ethan

=================

A compressor or limiter is an automatic volume control that reduces the volume when the input gets too loud. Originally they were used to prevent AM radio transmitters from distorting if the announcer got too close to the mike. Then some creative folks discovered that a compressor can sound cool as an effect on voices and musical instruments.

The primary controls on a compressor are:

Threshold - also called ceiling - This sets the point at which the automatic volume reduction kicks in. Below that volume the compressor does nothing. When the input gets above that level, the compressor reduces the volume automatically to keep the signal from getting much louder.

Attack time - how quickly the volume is reduced when the input exceeds the threshold. If it's too slow, then a short burst of loud music can get through and possibly cause distortion. So when using a compressor as a tool to prevent overload you generally want a very fast attack time. But when used on an electric bass to get a little more punch, 20-50 milliseconds is often good because that lets a little burst of the attack get through before the volume is reduced. So each note has a little extra "definition" but without the full length of the note being too loud.

Release time - how quickly the volume comes back up once the input is no longer above the threshold. If it's too fast, you'll hear the volume as it goes up and down. That sound is called "pumping" or "breathing." Sometimes this sound is desirable, but often it is not. The best setting depends on whether you're using the compressor as a tool to prevent overloading, or as an effect to create a cool sound or add more sustain to an instrument. If you don't want to hear the compressor work, set the release time fairly long (one second or more).

Compression ratio - 1:1 does nothing. 2:1 means if the input rises to 2 dB. above the threshold, the compressor will reduce the level by 1 dB. so now it's 1 dB. above. 10:1 means you have to get 10 dB. above the threshold for the output to go up by 1 dB.

Besides serving as an automatic volume control, a compressor can also make notes sustain longer. To make a note sustain requires raising the volume of a note as it fades out. That is, making the trailing part of a note louder to counter its natural fadeout is what makes it seem to sustain more.

To do this with a compressor you'll set the threshold low enough that the volume is reduced most of the time. Then as the note fades the compressor reduces the volume less, which is the same thing as raising the volume. For example, when you play a note on an electric bass the compressor immediately reduces the volume by, say, 10 dB. because the start of the note exceeds the threshold by 10 dB. You don't hear the volume be reduced because it happens so quickly. But as the note fades over time, the compressor raises the volume which gives the effect of adding sustain.

That said, I pretty much stopped using compressors a few years ago. Now I use volume envelopes in Sonar as needed to raise soft syllables or lower too-loud stuff. Programming volume changes manually rarely takes longer than finding the right compressor settings, and of course you can change the volume envelopes any time in the future. The big advantage of avoiding a compressor is to not add pumping and breathing sounds. These days the only things I compress - and always after recording, non-destructively - are acoustic guitar and electric bass if they need a little more sustain as an effect.

Doublehelix Fri, 08/15/2003 - 05:40

OK...obviously, there are no true "rules of thumb" because music is an art form, and there are a million ways to skin a cat...but:

I'd like to hear some general guidlines on how to use attack and release settings under "normal" conditions (i.e. not to generate special effects like pumping). This probably also will vary depending on the musical style, but let's start by focusing on rock to start.

For example:

Bass: short attack and release
Vocals: long attack, short release...

Or whatever...(I just made these up).

Any takers?

Guitarman Fri, 08/15/2003 - 07:09

Hey DH,

As far as normal conditions go the variables are many. The main thing you strive for when using compression is to maintain the dynamic range of a particular signal with actually hearing the processor work(pumping).

This requires basically tweeking the unit for any instrument you are using it on. So as you can see it varies quite a bit to just throw out ratios and thresholds. Just experiment with everything and trust your ears.

Best wishes,

JD( o}===;;;

Ethan Winer Fri, 08/15/2003 - 07:09

Falkon,

> what's your primary DAW software? <

Er, um, Sonar. :)

My first DAW was a Pentium MMX with Windows 95 using SAWPlus for audio and Master Tracks Pro for MIDI sequencing. Over the years I looked at many other audio / MIDI programs but rejected them all. When Sonar came out I switched to that, and never looked elsewhere.

--Ethan

Ethan Winer Fri, 08/15/2003 - 07:16

DH,

> I'd like to hear some general guidlines on how to use attack and release settings under "normal" conditions <

I almost always set the attack time as fast as possible, and adjust only the release time. The faster the release, the more aggressive and punchy the sound. The slower the release, the less you'll hear the compressor working. To use a compressor as an automatic volume control - either to prevent clipping on a track or to make a music bed behind an announcer consistent - you usually want a fairly slow release time.

--Ethan

falkon2 Fri, 08/15/2003 - 08:06

Okay... Just thought I'd ask, because everyone seems to consider it more of a toy than a tool. ;)

Nice to see a fellow SONAR user.

Well, back to the topic at hand... and since you use SONAR... do you have the normal or XL version? How does the Timeworks compressor (64-bit processing? :confused: ) rank up against other plugins?

Doublehelix Fri, 08/15/2003 - 08:40

This requires basically tweeking the unit for any instrument you are using it on. So as you can see it varies quite a bit to just throw out ratios and thresholds. Just experiment with everything and trust your ears.

This is how I have been doing it as well... I just tweak the attack and release times until I find a sound that fits the instrument and the song. I have started establishing some starting points for certain instruments however, and was hoping to see if anyone had some guidelines.

Ethan Winer Fri, 08/15/2003 - 09:06

Falkon,

> everyone seems to consider it more of a toy than a tool. <

While Sonar is not "eye candy" in the same way as Cubase, it's a far better choice for me. That it works reliably and does everything I want is the main reason I chose it. That it's not copy protected makes it all the better. I never buy copy protected software. Ever.

> do you have the normal or XL version? How does the Timeworks compressor (64-bit processing? :confused: ) rank up against other plugins? <

I do have the XL version, but I never installed the Timeworks compressor because I'm so satisfied with the UltraFunkplug-ins They do everything I want perfectly well, so I never felt a need to look further.

--Ethan

Ethan Winer Fri, 08/15/2003 - 09:57

Flood,

> I prefer the dongle on Nuendo over having to call for an authorization code like you do with Windows XP or Sound Forge <

I agree that a dongle is the lesser of the two evils. If you have to call every time you install a program, the day the company goes out of business is the last day you'll be able to buy a new computer or even get a new hard drive.

I will point out that as of a few months ago, anyway, SoundForge does not have to be registered to work. That's true if you download it, but if you get it on a CD it stops pestering you to register after a few times.

Likewise, I never had to register Windows XP. It came preinstalled on my Dell, and the first thing I did was format the hard drive, partition it, and install XP from scratch. It never once asked me to register. But I realize this is probably the exception, and perhaps it searches for a Dell signature in the BIOS or something. Of course, if Microsoft goes out of business we all have much bigger problems than not being able to install Nuendo! :D

--Ethan

anonymous Sun, 08/17/2003 - 19:21

Forgive me if I interject a clarification about the definition of knee, which is the only aspect of the question that wasn't pretty much covered by the other answers by the real experts.

In my primitive understanding, knee only affects the signals that are being processed right around the threshold level. On a hard knee setting, anything below the threshold will not be compressed at all, and as soon as the signal crosses the threshold the gain will be reduced by exactly the amount that the ratio dictates.

If there is a soft knee setting, it works a bit differently. As the signal level approaches the threshold, it starts to be compressed by a slight amount, gradually increasing until the level crosses the threshold, and may not actually get to the full value of the ratio until it is a bit above the threshold setting.

Viewed as a graphic, a hard knee would look like a rising line at a 45 degree angle, that when it hits the threshold value would take a sharp turn and continue rising at a smaller angle.

The soft knee graphic would look pretty muchthe same, except instead of the line taking an instant turn at the threshold point, there would be a curve that would start just before the threshold and continue until just after.

I don't know if that makes it more clear or just confused everyone!

As for Ethan's comment, I would never argue with an expert. But within my limited technical skills I adjust the attack time quite often. If I'm trying to limit a track, I will use a fast attack. But if i'm trying to keep something very punchy, i might let all the attack transients through unscathed.

anonymous Mon, 08/18/2003 - 06:42

Wow! (blush)

All this positive feedback encourages me to add one more point:

The original question also included "output". On many compressors this is known as "make-up gain" which is perhaps a better description.

While many times a compressor is used to make a signal more audible in a mix, or essentially "louder", the irony is that compression itself actually makes a track softer! But by shrinking the dynamic range of the track, it enables us to use the "make-up gain" to bring the overall level of the track up in volume after it has been compressed without worrying about clipping.

So while the very loudest transients may not, in fact, be any louder (they may even be softer!) the average or perceived level of the track is louder.

Once again, easier to demonstrate than describe.

anonymous Mon, 08/18/2003 - 10:55

Something people do is purposefully make the compressor "breathe" or "pump" in time with the music. This sorta adds some movement to the music. What you ussually do is set the compression's release time so that on something like a snare track with steady quarter notes, the compression dips down on the snare hit and then it releases just in time for the next snare beat.

Another way people use compression is so that it gives drums or other more rythmic instruments more "punch" or "bite" (though, i hate using those terms) This is done by setting the attack time so that the compression happens just after the snare or kick hit. yeah it's been mentioned that it can help bass guitar have more "attack". but on bass guitar, people use compression in a more "conventional" way because listeners like the sound of a constent volume of bass at all times. Most bass guitar players can't play notes consistently from one note to the next, and that's due to many reasons including the acoustics of the room they are playing in. But yeah, the technique of the performer has alot to do with it too. There are only a few bass players in this world who some think don't need any compression becuase they are just so talented!!

Anyways...
If your having trouble hearing the way compression affects the sound try loading up a single source file into a wav editor. Load in a drum beat (for example) and apply some compression and see how it shapes the waveform. Just use outragous amounts of each setting and see what happens... Of course, you'll get used to the sound and purpose of compression after time but it does take time to learn to listen for it and know whether or not you have too much or too little, or too fast of attack or release. It can easily take years and years to master, and hours and hours of tinkering about - Alot of extremely experienced engineers i know (yeah I know a couple) say that, while they know how to use it to get great sounds, they still don't have complete control or understanding of it.

no, it's not a stupid you question you ask. not at all!

[ August 18, 2003, 02:33 PM: Message edited by: by ]

Ethan Winer Mon, 08/18/2003 - 14:10

Swan,

> and may not actually get to the full value of the ratio until it is a bit above the threshold setting. <

Great explanation. I hope you don't mind that I added it to my "standard blurb" since this is asked all the time.

> I would never argue with an expert. But within my limited technical skills I adjust the attack time quite often. <

No argument there! I even made the point in my post above that allowing a short burst through can be useful.

--Ethan