Detectors in compressors and Envelop filters.

[POTL]

Let's discuss them.
I studied the principle of operation of the detectors and began to analyze the circuit, however I asked a number of questions.
1) Why do they use only half-wave rectifier in Envelop filters? Why no one uses a full wave rectifier?
2) Compressors use different types of rectifiers and half-wave and full-wave rectifiers. Why do primitive and classical circuits like Dyna Comp or Diamond use a full-wave rectifier, and modern circuits like THAT use only a half-wave rectifier? Is this somehow related to Peak / RMS compressor types?
3) Classical detector layout in OTA / OPTO compressors, how relevant is it today? Or is it worth looking at a precission full wave rectifier?

[Kipper4]

Hi Potl

1 might be an economy of parts on the makers part. Just might be there's no need.
   Maybe the vca is expecting a certain type of rectification.


Here's what Rod Elliot has written about, there may be some stuff of intrest to you.

http://sound.whsites.net/appnotes/an001.htm

I still have many questions about the many uses of rectifiers.
It's worth looking at some synths stuff too.

[Mark Hammer]

Rich/Kipper (welcome back!) is partly correct.

One of the things to consider with any sidechain is the properties of the control element.  Another is what the sonic function of the sidechain is.  Finally, another is what the time constants are to be employed.

For instance, if the control element is something that responds very quickly, like a JFET, OTA, or a photocell with remarkably fast response time, envelope ripply may be noticeable.  If one uses a sluggish photocell, the slower response time can smudge over any residual envelope ripple, making use of half-wave rectification perfectly acceptable. 

If the sonic outcome is something that takes place very quickly (e.g., peak limiting), then half-wave may also be acceptable, because what the sidechain is accomplishing is over in moments, before any ripple would be detectable.

One of the contexts in which one hears concerns over half-wave envelope ripple is with autowahs.  Such ripple is at its greatest during the decay of the strings, where those less familiar with ripple can easily perceive it as "distortion".  Similarly, when compressors or noise-gates use half-wave and also employ a medium-long gain-recovery or  release time, ripple will become apparent.  One solution is to use bigger-value averaging caps (e.g., 33uf instead of 4u7 or 10uf) that smooth out the ripple.  The cost to that is that quick attack and decay times become less available.

[POTL]

Hi Potl

1 might be an economy of parts on the makers part. Just might be there's no need.
   Maybe the vca is expecting a certain type of rectification.


Here's what Rod Elliot has written about, there may be some stuff of intrest to you.

http://sound.whsites.net/appnotes/an001.htm


Thanks, I read it, but I did not find answers to my questions there.
I still have many questions about the many uses of rectifiers.
It's worth looking at some synths stuff too.

[PRR]

> like THAT use only a half-wave rectifier?

The THAT "RMS" chip IS a full-wave rectifier!

What is the RMS of +1V? 1V. What is the RMS of -1V? Also 1V!

They may have futher "rectifiers" to modify the control voltage, but since the RMS is single-polarity they do not have to handle both ways.

Don't believe in THAT? Look at another RMS converter, say Analog Devices AD637. Note that the signal path starts with "Absolute Value".

[POTL]

> like THAT use only a half-wave rectifier?

The THAT "RMS" chip IS a full-wave rectifier!

What is the RMS of +1V? 1V. What is the RMS of -1V? Also 1V!

They may have futher "rectifiers" to modify the control voltage, but since the RMS is single-polarity they do not have to handle both ways.

Don't believe in THAT? Look at another RMS converter, say Analog Devices AD637. Note that the signal path starts with "Absolute Value".

Yes, I re-read the datasheets and other materials on the THAT website, if I understood correctly, the RMS is a full-wave detector, and the next one is full-wave and it is responsible for the operation of the threshold, when the voltage exceeds the threshold, the compressor starts to act.

[POTL]

And what can you tell about the detector in dynacomp / diamond - this detector is full wave or half wave?
In the description of his work and in simulations - he is not like one.


I also noticed an interesting detector on the site madbean that you can tell about it (the diagram is attached).

http://www.madbeanpedals.com/projects/_folders/1590A/schematics/Afterlife.gif

[Rob Strand]

dynacomp

The Dynacomp is full-wave. 

The textbook rectifiers are linear in that the output doubles when the input doubles.  Diodes without opamps act as thresholds and also create non-linearity.   For a feedback type compressor/limiter the non-linearity lets you get higher compression ratios than 2:1.

The Dynacomp rectifier has roughly an exponential non-linearity.  When you use that on a feedback type compressor/limiter it gives you higher compression ratios, almost approaching a limiter.    With a feedforward side chain you can get high compression ratios with a linear rectifier.

[noisette]

dynacomp

The Dynacomp is full-wave. 

The textbook rectifiers are linear in that the output doubles when the input doubles.  Diodes without opamps act as thresholds and also create non-linearity.   For a feedback type compressor/limiter the non-linearity lets you get higher compression ratios than 2:1.

The Dynacomp rectifier has roughly an exponential non-linearity.  When you use that on a feedback type compressor/limiter it gives you higher compression ratios, almost approaching a limiter.    With a feedforward side chain you can get high compression ratios with a linear rectifier.

Very interesting, have never thought about feedback ´slope´.
The Mad Professor Forest Green Compressor uses the same scheme, also some other...
http://1.bp.blogspot.com/-lFaW7baQFHY/TxgL2Kh3zLI/AAAAAAAAAoA/iH0mVjiMO84/s1600/FGCC.png
The diodes are always used as shunting neg. current away, not in series with signal (excuse yucky terminology). Is there a reason or are they all copying one circuit?

[Rob Strand]

The diodes are always used as shunting neg. current away, not in series with signal (excuse yucky terminology).

The shunt diodes kind of reset the input caps and make the rectifier act more like you expect.  Without the diodes it has some complex behaviour (read: details that make your head spin).  There was a thread on this forum about the non diode case about two years ago.

[POTL]

Without diodes DOD280
Perhaps the absence of diodes due to the unstable volume with strong compression

[POTL]

OK, we discussed the compressors and it gave me a lot of useful information and understanding.
Let's go back to envelop filters.
Why in all popular circuits - Q-Tron / Mu-Tron 3 & 5 / DOD fx25 & 440 / Mad Professor SWAW and many others we see the detector for only one half period, regardless of whether the circuit works from 9V or from bipolar power supply?
Is this somehow related to the mode of operation of the filter up and down? Will the filter capture both half-periods and give strange sounds or is there simply no need for a full-wave detector?

[merlinb]

Why in all popular circuits and many others we see the detector for only one half period, regardless of whether the circuit works from 9V or from bipolar power supply? ... is there simply no need for a full-wave detector?

Because a full-wave precision rectifier needs an extra opamp. Opamps cost money*, power, and space, so why bother when half-wave is 'good enough'?

*Especially in the 1970s when some of those were designed!

[POTL]

Why in all popular circuits and many others we see the detector for only one half period, regardless of whether the circuit works from 9V or from bipolar power supply? ... is there simply no need for a full-wave detector?

Because a full-wave precision rectifier needs an extra opamp. Opamps cost money*, power, and space, so why bother when half-wave is 'good enough'?

*Especially in the 1970s when some of those were designed!

Hmm saving a couple of cents?
And from the point of view of the filter envelope there will be some advantages in the work? Better tracking or sound?

[rankot]

I'm gonna follow this

[Kipper4]

All good stuff.
Thanks for the welcome Mr Hammer.
It's been a busy year again.
Happy new year to you all too.

Potl is this a purely discussion thread or are you thinking ahead with a project in mind?

A filter will respond to the changes you make to it. Filters can be tuned.

Tracking in the envelope detector can be tuned too. Biasing. Timing cap (commonly referred to as C5). Gain of said detector.

In my signature is a link to a paper I wrote on Optical Enveloping.
There's some stuff on envelope detectors in there.

Rich.

Keeping my glass half full.

[noisette]

In my signature is a link to a paper I wrote on Optical Enveloping.
There's some stuff on envelope detectors in there.

Rich.

Keeping my glass half full.

Thanks that´s a wonderful writeup.
I think there´s really not an universal functional difference between compressors/limiter/gate type gadgets and envelope modulation of filter/delay/fx etc.
You (could possibly) want utter tweakability, attack, release, amount, even a treshold -or an extra hp/bp/lp-filter in the sidechain- could be interesting with a filter.
Only thing that´s different, dynamic processors are (traditionally) meant to be unnoticeable, while you wouldn´t want that of your pedalboard
So THAT RMS-detector is maybe not optimal for fx use?

[POTL]

All good stuff.
Thanks for the welcome Mr Hammer.
It's been a busy year again.
Happy new year to you all too.

Potl is this a purely discussion thread or are you thinking ahead with a project in mind?

A filter will respond to the changes you make to it. Filters can be tuned.

Tracking in the envelope detector can be tuned too. Biasing. Timing cap (commonly referred to as C5). Gain of said detector.

In my signature is a link to a paper I wrote on Optical Enveloping.
There's some stuff on envelope detectors in there.

Rich.

Keeping my glass half full.

This is a discussion topic.
I want to close some points in understanding the work of some parts of the schemes, the practice will be somewhat later =)

[POTL]

And finally. Let's compare the schemes of a full-wave rectifier based on an operational amplifier (it is possible in studio compressors or some envelope filters), and a full-wave rectifier from the Dynacomp scheme, which has already become standard and is used in most compressor pedals, regardless of type - OTA / VCA / Opto.
What are the advantages of each scheme?
What is more relevant in our time?

[rankot]

Opamp version has far less parts.