hey all so i was doing a comparison between a few compressors and noticed afew things. 

i set the release at 250 ms on both, and after the signal dropped below the threshold, the compressor looks like it actually took more like 500 or more ms to release... 

anybody else experienced this?

nobody at all?

What were the attack times?


Hold time is a great mystery on most compressors.



Some compressors track a parallel input signal and respond to that "original" or reference signal while others are in series and respond to the sound after the compressor has started effecting it with the first level stage. What you consider as below the threshold may or may not be below the threshold in the circuit where the detection is happening.

You really have to dig in to the circuit flow to figure out what to expect.

With make believe analog dsp compressors it seems like it's very hard to understand the circuit because there seems to be so little documentation provided.



Also there are release "curves"... they are rarely documented and they can be linear or complicated... if the curves are different, the timing on the level response will seem different even if a time input parameter such as release time seems identical.




Good luck.


best regards,
mike

so what you are saying is that even when a plugin does not have a hold time knob, it could still have a hold time? 

this doesn't seem correct to me, I did the test by having a signal that was below the threshold after the main signal was stopped. I didn't see any "hold" all I saw was the instant, yet gradual release.

around 500 ms rather than 250. 

compression is a precision tool...

There are so many other variables that go into effect when using and setting up a compressor. What you are seeing is  normal, depending on the other settings 

Contrary to popular belief, threshold has very little to do with release times. Most compressors take their release time starting point from the point at which the output signal falls below the input signal which may or may not be below the threshold, depending on material. As CJ says lots of variables.

For full details grab a copy of "Mixing Audio" by Roey Izhaki where he explains the concept.

my reference:

http://www.amazon.ca/Mast...-Science/dp/0240808371
He states that release is a defined time. 

the compressor works with the "compressed signal". This signal is increased over the timespan of 'R' ms.

you say a compressor is complicated to set up. ok. but when i have 1 saw wave with no modulation. just 2 blocks of signal. one 10db above the threshold, one 10db under. the application could not be more basic. 

  Some compressors do have a release time and it is set by the user. It is also independent of the attack time. I'm not sure what is meant by "release is a defined time" unless he his alluding to it being a set by the user function. If a compressor does have a release variable than it is most like fixed. But that fixed time could be very different from compressor to compressor. 

The below is for the PC Channel Compressor PC76 U-Type from its help file.
 

Release. Adjusts the time the compressor takes to go back to an inactive state after the input level has fallen below the threshold value. Short release times will make the compression more flexible and able to adapt to the input signal, but can cause fast changes in gain that may sound displeasing to the ears. Longer release times produce a signal with a more even level and less of a “pumping” effect, but make it harder to maximize the overall compression because changes in the input level cannot be compensated for quickly enough to make the output volume totally consistent. The valid range is 50.0 milliseconds to 1200 milliseconds, and the default value is 510 milliseconds.





This agrees with my manual for a hardware compressor I have.