The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Compressor Release.. has a Threshold?
Old 6th July 2007
  #1
Here for the gear
 
ubik's Avatar
 

Compressor Release.. has a Threshold?

Hello.

Ive been noticing that when a compressor goes into the Release stage or "recovery stage" and even though it falls below the Threshold, at the very beginning of the release stage its still compressing somewhat in-order to return to unity gain at the end of the Release time.

Im wondering now.. this compression that takes place at the onset of the Release stage, what determines how deep it cuts into the audio? I understand that the Threshold and Ratio affect it.. but is there a fixed amount or is it some type of formula or something to figure out how much compression is taking place in the Release stage in relation to how much compression is taking place above the Threshold?

this question has baffled me since the beginning of time.. lol, (since i first learned about compressors)

your help is appreciated.
Old 6th July 2007
  #2
Lives for gear
 

Think of the amount of potential compression as first being functions of input level vs threshold and ratio.
But then there is the time greater than attack time above threshold that would be the second factor determining the amount of compression.
Now release only delays recovery, not how much reduction came before. So instead of 'how deep' in this phase, rather how long.
Old 6th July 2007
  #3
Here for the gear
 
ubik's Avatar
 

Hey Wayne, thanks for replying.

hmm.. ive read your response quite a few times trying to comprehend it fully..

so does the Release, then, use a Threshold and Ratio of it's own? im thinking perhaps somehow relative to the amount of gain reduction that is occuring?
Old 7th July 2007
  #4
Lives for gear
 

Quote:
Originally Posted by ubik View Post
...so does the Release, then, use a Threshold and Ratio of it's own? im thinking perhaps somehow relative to the amount of gain reduction that is occuring?
I'm just passing on what I've picked up along the way but I've not run into anything to indicate anything other than time being the factor in release. Where does the idea of a separate ratio or threshold come from?
Old 7th July 2007
  #5
Here for the gear
 
ubik's Avatar
 

Quote:
Originally Posted by Wayne View Post
I'm just passing on what I've picked up along the way but I've not run into anything to indicate anything other than time being the factor in release. Where does the idea of a separate ratio or threshold come from?
Well here's the thing.

I understand that the Release is the amount of time for the compressor to go from full compression back to no compression. But what I dont understand is that...

If one has to set a specific Threshold to tell the compressor to compress everything above that mark.... then when the signal drops below the Threshold, and goes into the Release stage... how is it possible that it is still compressing the audio if the signal is below the Threshold?

Though I accept this as being part of the compressor now, but my question is.. ok, it's in the Release stage and its compressing... so the Threshold must have dropped when the signal dropped did it not?

here's an image to point out exactly what im talking about, the compression that takes place during the Release stage.

Old 7th July 2007
  #6
Lives for gear
 

Quote:
Originally Posted by ubik View Post
... I understand that the Release is the amount of time for the compressor to go from full compression back to no compression.

... how is it possible that it is still compressing the audio if the signal is below the Threshold?
I'm guessing that the image is of a repetitive pattern whose level is being stepped up and down into the comp?
I believe your two statements may be at the crux of it. Just as the signal spends some amount of time above the threshold going into compression as the signal is stepped up (time to ramp gain down) it also has to honor the release time to ramp the gain back up on the other end.
Under this line of reasoning, at both ends of the process the signal spends some amount of time at these other levels.
Old 7th July 2007
  #7
Deleted User
Guest
I don't even understand what you mean.

If the threshold is hit at lets say -2db, the small dwarf in the compressor pulls out a stopwatch called "attack". After the attack time is reached, he gets his second stopwatch out called "release". He starts to pull the gain down and immediatly starts to push it back up slowly until he has reached unity again at the exact time his "release" stopwatch hits zero.
If during that another peak is reached, another little dwarf pulls out his stopwatch and does the same routine.
So release really has nothing to do with "threshold". If the threshold is reached than this exact routine starts every time with a set release time that will always be the same and there. If there is a continous peak, it's the same as just lowering gain.

I think you didn't understand how the compressor works?
Dunno.
Old 7th July 2007
  #8
Here for the gear
 
ubik's Avatar
 

Quote:
Originally Posted by zwaps View Post
I don't even understand what you mean.

If the threshold is hit at lets say -2db, the small dwarf in the compressor pulls out a stopwatch called "attack". After the attack time is reached, he gets his second stopwatch out called "release". He starts to pull the gain down and immediatly starts to push it back up slowly until he has reached unity again at the exact time his "release" stopwatch hits zero.
If during that another peak is reached, another little dwarf pulls out his stopwatch and does the same routine.
So release really has nothing to do with "threshold". If the threshold is reached than this exact routine starts every time with a set release time that will always be the same and there. If there is a continous peak, it's the same as just lowering gain.

I think you didn't understand how the compressor works?
Dunno.

But wait, doesn't the Release time only begin once the signal drops below the Threshold?

no trust me.. ive read like every compression explanation out there probably.. even tried looking at vst opensource code to see the functions of the release time.. but i still dont see it...

ive just settled on this final explanation then, please correct me if im wrong:

Quote:
The Attack stage compression lasts until the audio drops below the Threshold again. Once the Attack stage has ended and the signal has dropped below the Threshold, the Release or "recovery" stage begins. Even though the signal drops below the Threshold where no compression is supposed to take place, the Release function is a special one where, once the signal drops below the Threshold into the Release stage, it is still compressing a little bit of the signal below the Threshold. The Release function that you adjust on a compressor is the amount of time you want the compressor to take, in milliseconds or seconds, to go from that little bit of compression back up to the original, uncompressed, audio level.
Old 8th July 2007
  #9
Gear Maniac
 

Quote:
Originally Posted by ubik View Post
But wait, doesn't the Release time only begin once the signal drops below the Threshold?

no trust me.. ive read like every compression explanation out there probably.. even tried looking at vst opensource code to see the functions of the release time.. but i still dont see it...

ive just settled on this final explanation then, please correct me if im wrong:
"The Release function that you adjust on a compressor is the amount of time you want the compressor to take, in milliseconds or seconds, to go from that little bit of compression back up to the original, uncompressed, audio level." is exactly right. Saying "the signal is below threshold, there's supposed to be no compression" is wrong because of that very definition of the release stage.

If the signal never reaches the threshold to begin with, yes there will be no compression. Once there IS compression, you're in a different state. If the gain reduction dropped immediately back to 0 as soon as the signal went below threshold, you would get what sounds like distortion from the level being adjusted so quickly.

So basically there IS supposed to be compression on the signal once it initially returns below the threshold. The release lets the compressor know how long before that compression returns to zero.
Old 8th July 2007
  #10
Here for the gear
 
ubik's Avatar
 

Quote:
Originally Posted by beachhunt View Post
"The Release function that you adjust on a compressor is the amount of time you want the compressor to take, in milliseconds or seconds, to go from that little bit of compression back up to the original, uncompressed, audio level." is exactly right. Saying "the signal is below threshold, there's supposed to be no compression" is wrong because of that very definition of the release stage.

If the signal never reaches the threshold to begin with, yes there will be no compression. Once there IS compression, you're in a different state. If the gain reduction dropped immediately back to 0 as soon as the signal went below threshold, you would get what sounds like distortion from the level being adjusted so quickly.

So basically there IS supposed to be compression on the signal once it initially returns below the threshold. The release lets the compressor know how long before that compression returns to zero.
thank you, thats probably one of the best explanations i've come across yet. I appreciate everyone's help here.
Old 13th July 2007
  #11
Gear Head
 
miditus's Avatar
 

it also contribute to the overall sound of a certain because this curve is unique for every compressor.
Old 13th July 2007
  #12
Beachhunt gave a good explanation.

The confusion is happening because you're imagining a compressor as a switch. It's not on/off.

Set up a mic and a compressor and clap once and watch the meters (use a medium attack, slow release and a very low threshold). The meters will dip down and move back up and it will all take after the sound of the clap as ended.

Recovery time is expressed in terms of the number of seconds it takes to get back to zero, but in a lot of cases, that's not necessarily true. Well, I haven't gotten a certian answer, but I've been asking around about this a bit lately.

Think of a compressor as a door closing and there's a door man pulling it back to fully open (zero). Think of the release time as inches per second or feet per second. It's rate at which the door travels.

You may have a signal that goes back and forth, above and below the threshold faster than the compressor can completely recover, so that door is swining back and forth between 1/3rd open and 2/3rds open.

It's also possible that the signal is alwasy above the threshold, but changes by how much and changes back and forth. So the door may swing back and forth between 2/3rds closed and fully closed and never get back to open until the songs ends. Or, soeimtes you get a dead stop in the middle of the song, and hear a giant sucking sound as the noise and ambience is brought up.

Hopefully this helps.


Here's my question:

If a compressor squashes 2dB with a 100ms release, the release time can be expressed as 50ms/dB.

If that some copmressor squashes 10dB, will it open more slowly or still over 100ms? For it to open over 100ms, the release time must change to 10ms/dB or if it's linear, then it would take 500ms to open. I think the latter is the definition of program depenent release.

I'm sure there's also the possiblity that 100ms a 2dB translates into 115ms or 130ms at 10dB as well and that the real answer is somewhere in between, it's it's design dependent.

I'd love to know how it actually works.
Old 13th July 2007
  #13
Gear Maniac
 

Quote:
Originally Posted by Mike Caffrey View Post
Here's my question:

If a compressor squashes 2dB with a 100ms release, the release time can be expressed as 50ms/dB.

If that some copmressor squashes 10dB, will it open more slowly or still over 100ms?
Definitely design dependent, like you said at the end. For example, many optical circuits have a faster-than-displayed release when initially coming down from high compression, then the release slows as the amount of compression reduces. Other circuits act completely opposite of the above, still others have a more consistent release rate throughout the curve.
Old 14th July 2007
  #14
Lives for gear
 
Darius van H's Avatar
 

Verified Member
I dunno, it seems to me you're over-thinking it somewhat.........each compressor has it's own sound or "action"...it's just a question of yanking the knobs until it give you what you want.
Old 14th July 2007
  #15
Quote:
Originally Posted by beachhunt View Post
Definitely design dependent, like you said at the end. For example, many optical circuits have a faster-than-displayed release when initially coming down from high compression, then the release slows as the amount of compression reduces. Other circuits act completely opposite of the above, still others have a more consistent release rate throughout the curve.
And there's also multi-stage release.

How do you find out which compressors work which way?
Old 15th July 2007
  #16
Lives for gear
 

Quote:
Originally Posted by ubik View Post

so does the Release, then, use a Threshold and Ratio of it's own?

Well, I can tell your thinking about this stuff!! IF a compressor is emulating , or just replacing a person riding a fader, then the interactive controls of threshold, knee shape, attack, ratio and threshold all tell the thing when to clamp down ( threshold) , how fast to reduce the amplitude (attack), how much to turn it down (ratio) and in addition the knee shape also tells it how fast to turn down , either quickly as in a normal knee or , slower , and gentler , as in a soft knee.

After all that , the release phase is just a reversing of the previous clamping down and the release tells it how fast to do that.
Old 15th July 2007
  #17
Lives for gear
 

also you seem to keep transposing the actions that occur above or below threshold. the compressor will clamp down when your signal goes above the threshold. release will occur when the signal has stopped being above the threshold.
Old 15th July 2007
  #18
Actually that's not true. It is realtive to zero, but the realease will begin as the signal drops down toward the threshold. The further above, the more compression and both the signal and compression level can fluctuate above the threshold.

Think of the release time as a timer for how long the compression sustains after it's no longer necessary. Think of it like a reverb tail, hanging after the signal has ended.
Old 15th July 2007
  #19
Lives for gear
 

O.K. MC. fair enough

I like to think that the interactivity of attack and ratio determine the nature of the "clamping " down part of the compressor show....
So (and I know there are diffferences between boxes) the release should be sloping back up in a similar , symetrical, mirrored ( the same way the turning down part was molded by the attack and ratio) , but further modified by the "timer "aspect of release control!

use your ears and then try to figure out why it worked so well; remember you should only be modifying the overall amplitude envelope and not pumping or breathing with every crest of the music ( unless you want those special effects)
Old 15th July 2007
  #20
There's also the knee.

The knee of the attack and release are not always symmetrical, although I usually think of them that way too.
Old 19th July 2007
  #21
Lives for gear
 
drakewire's Avatar
 

Okay uh what

I think you guys have too much time on your hands... LOL

Maybe so do i
Compression is the mixing gnome. He pulls the volume up when its needed and pulls it down when you dont. The gnome doesnt like high sharp sounds so he moves the knob down, and the low up. The gnome's come in several different colors and sizes and can cost you a lot of doing what you can truly do by riding your faders. However, I wouldnt recommend leaving home without a gnome on the other side of the glass...

LOL
Old 19th July 2007
  #22
There's also the Bedrock Compressor with a little teradactyl inside riding a fader.
Old 20th July 2007
  #23
Lives for gear
 
Coyoteous's Avatar
 

Quote:
Originally Posted by Mike Caffrey View Post
There's also the knee.

The knee of the attack and release are not always symmetrical, although I usually think of them that way too.
Okay, I'm just barely following here, but the common usage of "knee" has to do with ratio, right? Are you talking about the linearity/non-linearity of the attack and release?
Old 20th July 2007
  #24
I guess some people could see it that way.


When a signal crosses the threshold, how long until the compression begins?

The answer is the attack time.


Suppose there's 4dB of gain redcution. The gain reduction doesn't go zero, then 4dB. It goes zero to 4dB and through all the amounts in the middle. It's not a instant on off swtich.

How long does it take to get from zero to 4dB of gain reduction?

The answer is attack time + knee, which is not usually measure in time, but the knee determines how the compressor moves from the moment gain reduction starts to when it hit's the target amont of gain reduction.


Ratio is going to determine what that target amount is.
Old 21st July 2007
  #25
Lives for gear
 
Coyoteous's Avatar
 

Man, I thought I understood compression before I hit this thread!

(either going to reread everthing or move on - but thanks)
Old 21st July 2007
  #26
Lives for gear
 



Mr. Transfer plot is your friend!
Old 21st July 2007
  #27
Lives for gear
 
Coyoteous's Avatar
 

Quote:
Originally Posted by flatfinger View Post


Mr. Transfer plot is your friend!
Well, maybe if it were big enough to read! heh

I can tell the change in slope is a "hard knee," though.

(added) actually, here's the one for grown-ups:

http://www.soundonsound.com/sos/jun9...s/mixcomp1.gif
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
ubik / Mastering Forum
7
beesting / So Much Gear, So Little Time
5
smalltownjon / Geekslutz Forum
5
crew14 / Low End Theory
7

Forum Jump
Forum Jump