Question on Compressor Attack and Release times

J

JMac52

New member
I was reading an SOS Session Notes article by Mike Senior and he wrote in one instance (it happened to be a trombone) he used compression with a "medium" release time. I looked at the screen shot for the plug-in he used, and the release time was 400 msec. Now it might be my electrical engineering training, but to me 400 msec seems like a long time.

Since I've seen in other articles and books references to these relative range-type terms, I was wondering - is there a generally accepted definition for the ranges of short/fast, medium, long/slow for attack and release times?

I expect someone to answer "it depends on the source" and that's fine if it does, but wondering if there is a general across-the-board definition.

Thanks,

J
 
Attack and release "times" aren't really times, they are time constants, as in a dB of reduction/recovery per given time. The numbers around the knob can't be taken too literally since the time constant may not be defined the same way for different compressors. Your ears have the final say.
 
JMac52 said:
.. but wondering if there is a general across-the-board definition. ..
Click to expand...
Not that I've noticed. But in general look at 400, for a release it is rather 'medium.
1176- 50ms to about one second.
http://www.uaudio.com/hardware/compressors/1176LN.html

It is a long time' (for our meaning) but when you think about how it sounds it does sit in that area between fast = more up front' / forward in effect, and the much slower end of the scale as in longer term leveling.
 
That's hard to say, because it is always relative. A song with a slow tempo will in some cases require a slower release time than a song with a faster tempo on a given instrument. I personally base fast and slow off of the compressor knob itself. Fast would be from 0-25%, medium from 25-75%, and slow from 75-100%
 
i have had this question always. I always hear people say long attack , slow release and then give a lot of different numbers.

I think large reason why newbies like me struggle is we can't tell the difference when we listen A big problem is the room acoustics. It could be the type of compressor too. If you don't have good room acoustics, you cannot tell subtle differences in compression or anything else.
 
HoldenHarris said:
That's hard to say, because it is always relative. A song with a slow tempo will in some cases require a slower release time than a song with a faster tempo on a given instrument.
Click to expand...

Tempo is definitely worth considering when setting release time.

HoldenHarris said:
I personally base fast and slow off of the compressor knob itself. Fast would be from 0-25%, medium from 25-75%, and slow from 75-100%
Click to expand...

I don't even know what this means. Most of my compressors, hardware and software, have several knobs, none of which are labelled "compressor". I've seen guitar pedals with a "compression" knob which sets the level going into the threshold.
 
JMac52 said:
I was reading an SOS Session Notes article by Mike Senior and he wrote in one instance (it happened to be a trombone) he used compression with a "medium" release time. I looked at the screen shot for the plug-in he used, and the release time was 400 msec. Now it might be my electrical engineering training, but to me 400 msec seems like a long time.
Click to expand...

Another point, release times are typically longer than attack times because signal decays tend to be longer than initial transients. A long attack time might be over 100ms while a long release time might be over a second.
 
bouldersoundguy said:
Attack and release "times" aren't really times, they are time constants, as in a dB of reduction/recovery per given time.
Click to expand...

Interesting. I've never heard/seen it explained that way, but doing some quick research on compressor designs gives me a better understanding in that regard, and I wouldn't expect a setting of 100 ms on one compressor to necessarily be the same thing on another compressor. Still, I wouldn't expect a release time of 50 ms on one compressor to be like 500 ms on another one so the range question should still be fair. But I suppose it's all academic in the end.

J
 
anoopbal said:
i have had this question always. I always hear people say long attack , slow release and then give a lot of different numbers.

I think large reason why newbies like me struggle is we can't tell the difference when we listen A big problem is the room acoustics. It could be the type of compressor too. If you don't have good room acoustics, you cannot tell subtle differences in compression or anything else.
Click to expand...

The more I use compression and the better I understand it's purpose and the goal I have for it, the better I'm able to hear the subtle differences - I'm not just talking about subjective loudness, and I don't have a good room. Kinda why I asked the question, if read a suggestion or how someone else achieved a certain sonic improvement in a specific situation using, say, a long release time, then I'd like to know about where to start and then use my ears to adjust from there.

J
 
Honestly, I hardly pay attention to attack and release numbers, either on the unit or in somebody's description. Besides the fact that the numbers vary among compressors, your audio is different from that of whoever is making attack and release setting suggestions. So you're going to have to do it by ear to make it right for your mix anyway. Getting too hung up on the numbers is a big waste of time.
 
I'd offer that you can to a significant degree relate to the attack and release numbers. I would guess there's several I'm familiar with that for one thing use these sort of generic scales that do respond at least fairly close to 'as expected.
I'd be curious, do others not relate (or do?) for example to the general effect' difference of say 2ms attack, 5ms, 10, 30, on any number of sources- electric or ac guitar? Drums (drum buss?), a full mix?

One thing (I believe) is that there is a lot of action (reaction) coming from a compressor's relation of attack time to what it's seeing. And piled in and on that is the detector style (type?, 'peak or rms?).
I wont pretend to know the inner workings, but where that can be swapped', you can see those initial responses shift (i.e dig deeper, harder) for the same attack time. And if that means the amount of reduction shoots up.. there's the making of a seemingly different reacting compressor. What was' (for a given attack) now the threshold is too low, or the ratio too high.
So there's a big variable that should be part of our thinking. :)

I'd suggest trying a little exercise. Set up a generic comp, very low ratio (2:1), keep the release med-fast (100? Not too critical- looking for 'not so fast you might miss what's going on), no auto' modes! -including makeup gain.
Feed it- a bass or guitar, or a drum buss.
Play a little game with the attack - 1-- 30ms of 'what gets through?'
Things (and reactions) should be particularly interesting in the first 1-5 or ten ms ranges.
(I guess out past the '30+ ranges that'd be more the 'how does it move' game :)
 
Last edited:
anoopbal said:
i have had this question always. I always hear people say long attack , slow release and then give a lot of different numbers.

I think large reason why newbies like me struggle is we can't tell the difference when we listen A big problem is the room acoustics. It could be the type of compressor too. If you don't have good room acoustics, you cannot tell subtle differences in compression or anything else.
Click to expand...

Room acoustics are a factor to a point. Anything you can do to help set the space up to avoid problems from bass traps or whatever to positioning your monitors correctly will help.

Even as such, learning to hear what compression sounds like is important. I'd suggest something similar to what Mixsit posted above. Only with a higher ratio. Maybe take a source and set your attack and release times to 12 o'clock. Set maybe a 4:1 ratio or so and turn the threshold down until you're getting 6 to 10 dB gain reduction. As the source is playing, slowly turn the attack control down and listen carefully. At some point the compressor will probably kill the signal. The point where that happens can change depending on the source. The idea is to get some kind of exaggerated compression going on and play with it. Once you get used to how all the controls can react with the source you'll be able to hear it a lot easier at more subtle, musical sounding settings. The attack and release controls are critical for getting a source to behave naturally with the compressor. Do the same kind of thing with the release control.

Ratio and threshold control how much compression you get, but you can also change the shape of the compression. Gain reduction indicates how much compression you get but you can apply the same amount at a variety of different ratios by changing the threshold setting. Get 6 dB or so of gain reduction with 4 or 5 different ratios to hear the difference. Individual settings can only be meaningful when they provide the sound you want.
 
snow lizard said:
.. Get 6 dB or so of gain reduction with 4 or 5 different ratios to hear the difference. Individual settings can only be meaningful when they provide the sound you want.
Click to expand...
Ya know just off the top here, I think for me ratio changes would actually be tougher to pick out, than say attacks and releases.
Don't know if that's normal or not.. ;)


(..but I just got to use my Low Ratio Limiter again [patent pending :D
 
My usual attack is like 10 ms and release 280-305 ms. But my songs are bollywood with a lot of reverb and long tails.

I feel like I could tell the difference between all these times when I cleaned up my room acoustics a bit.
 
Tell me if this makes sense, I set my attack time based on the attack of the instrument playing and the same for the release. A pad for example that has a slow attack, will not peak for a certain length of time. Release depending on how much sustain is on the instrument. A drum hit I would think would require a fast attack and, well, quick release since it is more of a pulse. Cymbal would depend on the hit, hut it would be a "faster attack" but slower release (depending how it is to be controlled). Then if you are doing make up, vocals, those instruments that can vary, then it is a pie times thumb thing.

Then there is the other use, if you're just catching peaks, and want to "ease it" back in to normal. I understand the general use of compressors, but how to artfully use them I am still learning.
 
I set my attack time based on how I want it to sound. If something has a strong initial transient I may or may not want to suppress it. If I do want to suppress that initial transient then I turn the knob counterclockwise until it's suppressed, and if I don't want to suppress it I turn the knob clockwise until it's not suppressed. Same basic process with release. I really don't care what number the knob ends up pointing to because it's the sound that matters.

You can't rely on the numbers because threshold, knee and the signal itself all contribute to how the compressor actually sounds. If your attack is set to 24ms, for example, and you change another parameter then the sound will change. In the end all that matters is how it sounds.
 
Last edited:
anoopbal said:
My usual attack is like 10 ms and release 280-305 ms. But my songs are bollywood with a lot of reverb and long tails.

I feel like I could tell the difference between all these times when I cleaned up my room acoustics a bit.
Click to expand...
You're referring to compression for your mixes I take it?
 
bouldersoundguy said:
I set my attack time based on how I want it to sound. If something has a strong initial transient I may or may not want to suppress it. If I do want to suppress that initial transient then I turn the knob counterclockwise until it's suppressed, and if I don't want to suppress it I turn the knob clockwise until it's suppressed. Same basic process with release. I really don't care what number the knob ends up pointing to because it's the sound that matters.

You can't rely on the numbers because threshold, knee and the signal itself all contribute to how the compressor actually sounds. If your attack is set to 24ms, for example, and you change another parameter then the sound will change. In the end all that matters is how it sounds.
Click to expand...
There. (that :D
So - assuming one does even want to get into thinking about what you're feeding a compressor, and thus maybe what kind of set up' might get you where you think you want to go..There is the opportunity to apply a bit of this logic(? best guess?) ..up front.

I suggested the little 'numbers game for that. Basically it's an exercise - to look into the shapes, levels, or the mix of things in the case buss/song that's going to hit your compressor.
And (again assuming one even likes this approach.. many won't perhaps, one may also consider what is it you want to do with this 'information?
A two-mix comp makes for a great example. You might like how the drums (a big potential source of 'front end action) and the mix are sitting- and decide to stay well out side those transients, or pick a bit of them off, etc, etc.

But I do agree the numbers can not be relied upon for some of those good reasons you mentioned. Another reason.. hell we're working off seat of the pants best guess' mode in either case.
 
You must log in or register to reply here.
Back
Top