Compressor Release Question

[ITHertz]

Hi All,

I just read through the excellent discussion of how compressor release works here https://www.soundonsound.com/forum/view ... 3&start=40 - thanks Hugh for the clear explanation.

Am I right in thinking that if we have a steady signal that's 6dB above the threshold, and it suddenly changes to 3dB above the threshold then the release setting affects the rate of change even though the signal has never dropped below the threshold?

Otherwise we'd get a more-or-less instantaneous level change which usually isn't a good thing.

Cheers!

[Hugh Robjohns]

Yes, of course. It controls the rate at which the gain reduction is reduced, regardless of the amount of reduction.

[ITHertz]

Thanks, it's surprising how many explanations of this don't make it clear. And with regards to your "automated fader" analogy, I like to think of the attack and release settings as the fader having some "stiffness" that can be different depending on whether it's moving up or down.

Cheers!

[sonics]

You might think about it like this. Above the threshold, the compressor is active and so any changes will indeed be affected by the attack and release times.

The "stiffness" is better thought of as lag, I think.

[Drew Stephenson]

The "stiffness" is better thought of as lag, I think.

Or the knee setting - though arguably that's more of a stiction parallel.

[Hugh Robjohns]

I keep banging on about the importance of the engineering side of 'audio engineering'.... but this is a classic example of where some understanding of basic electronics makes the workings of a compressor so obvious.

In this case, it's the concept of a simple passive RCR filter.

Imagine a capacitor with one end grounded, wired with a resistor across it in parallel and another resistor connected to the top as a feed-in point.

Now, when the audio signal level exceeds the threshold an electronic switch applies its voltage to the open end of the feed-in resistor, forcing a current into the capacitor, charging it up do that the voltage across it rises.

The voltage building across that capacitor is used to control the audio gain-reduction device, so as the capacitor charges up the audio gain reduction is increased.

The feed-in resistor therefore determines the attack time as it limits the rate at which current can get into the capacitor to build that charge.

An even bigger input signal forces more current in, resulting in a higher voltage and thus more gain reduction.

However, if the input signal current falls slightly the capacitor will partially discharge via that parallel resistor to match the new lower input level.

The rate of that discharge is determined by the size of that parallel resistor which therefore sets the release or recovery time.

If the signal falls below the threshold level the current through the feed-in resistor stops completely, and so the capacitor discharges fully, taking the gain reduction back to zero.

Obviously, a practical side-chain circuit involves rather more finessing and complexity, but this RCR filter concept lies at the core of it all.

[Drew Stephenson]

[ITHertz]

I keep banging on about the importance of the engineering side of 'audio engineering'.... but this is a classic example of where some understanding of basic electronics makes the workings of a compressor so obvious.

In this case, it's the concept of a simple passive RCR filter.

Imagine a capacitor with one end grounded, wired with a resistor across it in parallel and another resistor connected to the top as a feed-in point.

Now, when the audio signal level exceeds the threshold an electronic switch applies its voltage to the open end of the feed-in resistor, forcing a current into the capacitor, charging it up do that the voltage across it rises.

The voltage building across that capacitor is used to control the audio gain-reduction device, so as the capacitor charges up the audio gain reduction is increased.

The feed-in resistor therefore determines the attack time as it limits the rate at which current can get into the capacitor to build that charge.

An even bigger input signal forces more current in, resulting in a higher voltage and thus more gain reduction.

However, if the input signal current falls slightly the capacitor will partially discharge via that parallel resistor to match the new lower input level.

The rate of that discharge is determined by the size of that parallel resistor which therefore sets the release or recovery time.

If the signal falls below the threshold level the current through the feed-in resistor stops completely, and so the capacitor discharges fully, taking the gain reduction back to zero.

Obviously, a practical side-chain circuit involves rather more finessing and complexity, but this RCR filter concept lies at the core of it all.

This is a great explanation, and as someone who has a moderate knowledge of electronics, makes it very clear how release works.

I think the reason for the general misunderstanding of compressor release comes back to the people who write the manuals (btw, I'm a technical writer). Nearly every manual I've read explains compressor release as something like:

"Recovery is a parameter that is nowadays usually referred to as Release. In a compressor, it sets the time it takes to "restore" the original dynamics after the audio falls under the threshold level"

or

the time it takes to go from compressed state to original non-compressed signal (Release).

or

Controls the time the signal needs to recover from compression when the level falls below the Threshold.

Cheers!

[ITHertz]

And another one:

The release time controls how quickly the gain reduction returns to zero after the detector drops below the threshold value.

[Hugh Robjohns]

All factually true, and all simplifications that won't affect the user's operation of the compressor in any practical way.

But that's often what's needed in a user manual aimed at the technically illiterate (as most who need the manual usually are).

It's similar to telling someone that pressing the accelerator will make an automatic car go faster.

Explaining how the gearbox might automatically select a different gear as they press the accelerator to allow the engine to operate in the optimum rev range to maximise torque and acceleration won't really help their use of the car...

[ITHertz]

All factually true, and all simplifications that won't affect the user's operation of the compressor in any practical way.

But that's often what's needed in a user manual aimed at the technically illiterate (as most who need the manual usually are).

It's similar to telling someone that pressing the accelerator will make an automatic car go faster.

Explaining how the gearbox might automatically select a different gear as they press the accelerator to allow the engine to operate in the optimum rev range to maximise torque and acceleration won't really help their use of the car...

That's true, and as a sometime technical writer I understand that a simple explanation is all that's needed in most cases. However, with the acceleration/deceleration analogy, most drivers won't care but race car drivers (and their mechanics) will. I think that mix engineers should aspire to be more like race car drivers, shaping the "acceleration of the sound".

And a bit of "forced" education for the technically illiterate can't be a bad thing, lol!

Cheers!

[Drew Stephenson]

To be taken with a pinch of salt...

Found on Reddit:
Threshold is the volume at which your mum comes in to tell you to turn it down.
Ratio is how much you turn it down.
Attack is how quickly you reach for the volume knob.
Hold is how long you wait before you turn it back up.
Release is how slowly you start to nudge it back up.