1176 Sidechain analysis

  Hello, I was looking at the 1176 schematic http://www.gyraf.dk/gy_pd/1176/1176sch.gif
And I was minding, when the attack is set to fastest, the CV should be full of AC, does this cause distortion in the audio chain?

And also the attack is supposed to be 20-800µs, is that real? I mean how is that possible for the "slowest" setting with a 25k 220n RC network?

All dynamics processors suffer from some level of cross modulation caused by side chain related gain changes.

The faster the gain change the more obvious distortion, slower gain changes can be too slow to be effective.  This is the "art" of side chain design and there have been many variants developed over the years.

JR

Sure, but isn't the 1176 slow attack setting something more like 5 or 10ms ?

saint gillis said:

when the attack is set to fastest, the CV should be full of AC, does this cause distortion in the audio chain?

Click to expand...

Absolutely. Even vaguely fast compressors (2ms = period of 500Hz which is the center of music frequencies) are basically wave-shaping slow clippers. So if you think about that, it's not possible to control gain faster and not have it wave-shape.

saint gillis said:

And also the attack is supposed to be 20-800µs, is that real? I mean how is that possible for the "slowest" setting with a 25k 220n RC network?

Click to expand...

That's a good point: 25k / 220n is 25Hz = 40ms. Of course that's one full cycle. If the CV suddenly changes, the time it takes for it to actually cause attenuation is going to be a fraction of that. But it doesn't really explain the 800us number. I'm not sure where they came up with those numbers. I don't think you can really put numbers to the settings at all because of how the CV reacts to all of the different settings. It's all terribly imprecise actually. I have a Quant Asylum QA400 audio analyzer which puts out an impulse response which is a pulse every 1s that has all frequencies present (so it sounds like a thud but with a click to it). So that short pulse every 1s is actually somewhat useful for visualizing faster compressors because I can watch the CV bouncing in reaction to it. When you look at that on a scope, you can see how the settings actually work and it's probably not what you would expect.

squarewave said:

Absolutely. Even vaguely fast compressors (2ms = period of 500Hz which is the center of music frequencies) are basically wave-shaping slow clippers. So if you think about that, it's not possible to control gain faster and not have it wave-shape.

Click to expand...

It is important to remember that it is the overall envelope of the side chain CR that determines waveshape, not just the attack. If the release is 1 second there will be no wave shaping except on the initial transient that causes compression to occur. If the release is also 2mS then the sidechain will be able to follow a 250Hz signal. This is why release time really should be limited to no les than 50mS (20Hz)

Cheers

Ian

This BBC monograph is a great read from 1967, it uses a LC pulse shaping thingy and delay line. https://downloads.bbc.co.uk/rd/pubs/archive/pdffiles/monographs/bbc_monograph_70.pdf

Price 5 shillings, worth every groat, postage free

ruffrecords said:

It is important to remember that it is the overall envelope of the side chain CR that determines waveshape, not just the attack. If the release is 1 second there will be no wave shaping except on the initial transient that causes compression to occur. If the release is also 2mS then the sidechain will be able to follow a 250Hz signal. This is why release time really should be limited to no les than 50mS (20Hz)

Cheers

Ian

Click to expand...

Yes good point

ruffrecords said:

It is important to remember that it is the overall envelope of the side chain CR that determines waveshape, not just the attack. If the release is 1 second there will be no wave shaping except on the initial transient that causes compression to occur. If the release is also 2mS then the sidechain will be able to follow a 250Hz signal. This is why release time really should be limited to no les than 50mS (20Hz)

Cheers

Ian

Click to expand...

That is objectively true for compressing steady state sine waves (perhaps audible compressing a bass guitar line ) but subjectively more complicated than that, and many brain cells have been killed pursuing clever side chain tricks to reduce audible artifacts over the decades.

JR

Btw why are the 200p caps in parallel with the 2.2M resistors ?

Counteracting gate capacitance?

JohnRoberts said:

That is objectively true for compressing steady state sine waves (perhaps audible compressing a bass guitar line ) but subjectively more complicated than that, and many brain cells have been killed pursuing clever side chain tricks to reduce audible artifacts over the decades.

JR

Click to expand...

What about using some fet switching to engage the release resistor to discharge the cap only when the signal drops to a level close to zero ?
Something like this :