What do these words mean on my compressor?

G

guitarguy101

New member
I want to know what raising and lowering the following things on my compressor will accomplish. Specifically, I'm using it on a drum kit, but I'd like to know what they mean in general.
The words are:
Threshold
Ratio
Attack
Gain
I already know what gain is, but what are the other 3?
 
guitarguy101 said:
I want to know what raising and lowering the following things on my compressor will accomplish. Specifically, I'm using it on a drum kit, but I'd like to know what they mean in general.
The words are:
Threshold
Ratio
Attack
Gain
I already know what gain is, but what are the other 3?
Click to expand...


Should read as - make up gain.

Here book mark THIS for future reference.
 
guitarguy101 said:
I want to know what raising and lowering the following things on my compressor will accomplish. Specifically, I'm using it on a drum kit, but I'd like to know what they mean in general.
The words are:
Threshold
Click to expand...

This is the dividing line between the low level stuff which isn't compressed and the high level stuff which is compressed.

guitarguy101 said:
Ratio
Click to expand...

How much to compress the signal once it's over the threshold. At 4:1 the output goes up only 1dB for every 4dB the input rises past the threshold.

guitarguy101 said:
Attack
Click to expand...

How fast the compressor reacts to a rising signal level, like right when you hit a drum.

guitarguy101 said:
Release
Click to expand...

How fast the compressor reacts to a falling signal level, like right after you hit a drum.

guitarguy101 said:
Gain
I already know what gain is, but what are the other 3?
Click to expand...

Since the effect of the compressor itself is always to lower the level most are equipped with a make up gain control to compensate.
 
This thread actually cleared a few things up. I wasn't quite sure on attack and release. This helped for sure, thanks!
 
bouldersoundguy said:
How fast the compressor reacts to a rising signal level, like right when you hit a drum.



How fast the compressor reacts to a falling signal level, like right after you hit a drum.
Click to expand...

Just to expand a bit on these two definitions...

Attack is how fast the compressor reacts to a rising signal ONCE IT GOES ABOVE THE THRESHOLD. Normally you'd want a very short/fast attack time.

Release is how fast the compressor reacts to a falling signal, ONCE IT FALLS BELOW THE THRESHOLD. For rock-style music this tends to be a fairly fast setting...but usually a bit slower than the attack. Just as you tend to do a quick fade up or down when using a mixer rather than throwing a switch which would cause an unpleasant change, you want the compressor to automatically do a quick fade in or out to disguise it's operation.
 
Bobbsy said:
Just to expand a bit on these two definitions...

Attack is how fast the compressor reacts to a rising signal ONCE IT GOES ABOVE THE THRESHOLD. Normally you'd want a very short/fast attack time.

Release is how fast the compressor reacts to a falling signal, ONCE IT FALLS BELOW THE THRESHOLD. For rock-style music this tends to be a fairly fast setting...but usually a bit slower than the attack. Just as you tend to do a quick fade up or down when using a mixer rather than throwing a switch which would cause an unpleasant change, you want the compressor to automatically do a quick fade in or out to disguise it's operation.
Click to expand...



And let's not forget about the compressors that have *AUTO attack and release*! ;)
 
Release time is how long it takes the compressor to return to a non compressing state, after the audio falls back below the threshold.
 
Bobbsy said:
That would be a neat trick but I've never found a compressor that could respond to "direction of travel". Normally the attack and release phases are just triggered by crossing the threshold.

I had a look for a diagram and this one (from Wiki of all places) seems to show it as clearly any anything else I could find:

http://upload.wikimedia.org/wikipedia/commons/6/6c/Audio_Compression_Attack_and_Release-2.svg
Click to expand...

You can hear the effect of attack and release settings even when the signal stays over the threshold, so they must be acting on rising/falling levels rather than threshold crossing. The Wiki illustration is wrong. Here's my research data:

I generated a tone alternating every second between -12dBFS peak and -3dBFS peak. That's the first pic. Then I applied a compressor with an ∞:1 ratio, -16dBFS threshold, 90ms attack and 250ms release. Note that the levels ramp up or down at each change in level even though the signal is above the threshold the whole time. This compressor (the stock one in Sound Forge 6) is reacting to rising and falling levels above the threshold.

before.jpg


after.jpg
 
Maybe I'm seeing something different than you but (by putting a straight edge between your two graphics) it appears to me that the curves indicating the attacks and releases are only trigger when the level crosses the threshold level.

It's midday on New Years day for me so perhaps I'm being a bit slow...but how could a compressor work properly if it detected the "direction of travel" of signals--you could have lots of "downward moving" levels on a signal that were still all above the threshold level and the system would be pumping like crazy since every music signal is continually rising and falling.

Edited to Add: I just had another thought. Does the compression in Soundforge have a "look ahead" setting like Audition has? If so, this could be a factor here. In Audition I can set a look ahead time (I usually use about 3ms) so I CAN make the release appear to start prior to crossing the threshold setting. In reality, it's still the "crossing the threshold" that's triggering the release--but, because of the lookahead, it starts releasing a bit earlier.

However, this is a setting more applicable to DAWS than to outboard gear used live. Working with live music, the only way a lookahead could be achieved is by introducing a delay slightly longer than the lookahead period--otherwise the compressor manufacturer as invented time travel!

Interesting discussion nonetheless!
 
Last edited:
I don't understand that.
From your images, it nearly looks like the release time is brought into play with the output level crosses below the threshold.
Am i just looking at this wrong? It is annual mental hangover day afterall.
 
I see the same as you, Steenamaroo--as mentioned I even lined them up with a straight edge.

I did do some experiments here, trying to use a signal with a slope to it rather than an instant change...it's harder to read but I got the same result with the release only happening when it crosses threshold. The one exception to this though was if I use the "look ahead" feature which did seem to let the release start prior to the signal crossing the threshold. However, as I said above, this isn't an option in hardware compressors, especially in a live scenario, because of the latency that would be required to let the gear look ahead.
 
The clip is above the threshold the whole time, aside from the start and end. The slopes illustrate the attack and release settings acting on signal without any crossing of the threshold.

Level/threshold/ratio determine the target gain reduction. When the level rises the target gain reduction increases. Attack controls how quickly the gain is actually reduced towards the target. When the level drops the target gain reduction decreases. Release controls how quickly gain is actually increased towards the target. This is happening all the time while the signal is above the threshold.

After a bit of experimentation I'll say that the Wiki illustration is correct but that it's still consistent with what I'm saying. There's nothing about it that says that attack and release respond to threshold crossing.
 
I still don't get this. Your diagram seems to prove otherwise, but every resource i can find seems to say release only comes into play once the signal has passed below the threshold again.

The only way both of these things can be right is if the 'signal' we're talking about is actually the output signal.
I don't imagine that's how it works, but it makes sense in relation to your picture.
 
Bobbsy said:
It's midday on New Years day for me so perhaps I'm being a bit slow...but how could a compressor work properly if it detected the "direction of travel" of signals--you could have lots of "downward moving" levels on a signal that were still all above the threshold level and the system would be pumping like crazy since every music signal is continually rising and falling.
Click to expand...

There is the possibility with peak level sensing, a very fast release time and low frequency signal that you can get distortion on the note decay as the release is not much slower than the slope of the waveform itself. The signal acts like a LF oscillator modulating its own amplitude. Slow the release down just a little and it won't do that.

Imagine a constant signal (like a synth pad) 6dB over the threshold with a 2:1 ratio. The output will be reduced 3dB. Now imagine an intermittent note on top of that synth pad that brings the signal to 8dB over the threshold. That 2dB change in input calculates to an additional 1dB of reduction. The attack setting controls how long it take to get from 3dB of reduction to 4dB of reduction. When the intermittent note stops the release setting controls how long it takes to go from 4dB to 3dB of reduction. All this happens above the threshold, without crossing it.
 
You must log in or register to reply here.
Back
Top