EQ with different Q for boost and cut

[Eddododo]

Ok this MAY be a false memory / fever dream, but I 'feel like' I remember seeing in application notes or an article or something a circuit where the boost and cut of an EQ ( or bandpass or
Something) would have different Q's.
As I said I BARELY (or maybe I don't at all ) remember this. If anything rings a bell ( even if it's not an EQ) point me there. Even if I'm wrong about it being EQ or similar, if nothing else I'd love to get this off the tip of my tongue or wherever it is.

[Rob Strand]

It is definitely possible.
An symmetric EQ produces a boost & cut responses which are mirror images on a dB scale.
An asymmetric EQ generally has the same shape boost response as the symmetric but the cut response is different.
Most form allow a narrower notch for cut.  The notch can be infinite, or it can bottom out at a specific cut level.

The idea is easy:
- The boost response is  Boost(w)  =  1 + kboost * BP(w)
- The cut response  is  Cut(w)  =   1 - kcut * BP(w)

BP(w) can be implemented with any bandpass filter;  State variable or a simpler one or two opamp circuit.

kboost is the gain in boost mode.    The actual boost is 1 + kboost*BP(w0)
kcut is the gain in cut mode.           The actual cut is 1 - kcut*BP(w0)

You need to mix the "1" part  with the boost and cut part.

Keep in mind to get a notch the  "1 - kcut * BP(w0)" value will be zero.  For a finite notch 1 - kcut * BP(w0) is a small value.

You can implement the boost/cut control with a single pot using an one opamp connected as a mixer with an inverting an non-inverting input.
The inverting mixer input connects to the BP output.  The input and feedback resistors set kboost.
The non-inverting input (the + input of the opamp) connects to the boost-cut pot.
BP output connects to a trimpot to tweak the cut level,  then to the anticlockwise terminal of the boost/cut pot.
The clockwise terminal of the boost/cut pot connects to ground.

Noise-wise this boost/cut method isn't be the best implementation.  However it is easy to understand.  The main idea being it can control the gain from a positive value to a negative value in order to the the add and subtract required for boost and cut.

There's many ways to do this.   As I recall there was a parametric circuit in Elektor which was asymmetrical; you might find it on Aron's old circuits page on this site.   The TC electronics one might also be asymmetrical; I can't remember.   Check out some of the Rane papers on EQ's.

In general the asymmetrical ones are a bit funky and often have quirks/limitations regarding noise and overloads.

Anyway I don't hang out here often but I'm sure someone will be able to fill in any blanks.

[Rob Strand]

Here's one:
http://audiocircuit.blogspot.com.au/2011/08/parametric-eq.html

[Hatredman]

But I'm curious: why would you want that?

Edit: I mean, of course a parametric is useful, but why would you want different Qs for boost and cut?


--
Scarlett Johansson uses a Burst Box with her Telecaster.

[Eddododo]

But I'm curious: why would you want that?

Edit: I mean, of course a parametric is useful, but why would you want different Qs for boost and cut?


--
Scarlett Johansson uses a Burst Box with her Telecaster.

Well really I wouldn't - mostly I just want to track down the application note that had it. But abstractly I like the sound of it for a semi parametric mid EQ- on bass ( and in general ) I've found that 'boost wide and cut narrow' is a good mantra. Now like most things this application falls in the uncanny valley of 'not quite enough control but not quite well tuned enough' but still, I'd love to have it in my notebook

[Eddododo]

Here's one:
http://audiocircuit.blogspot.com.au/2011/08/parametric-eq.html

cool thanks for the link!
And thanks for the above advice.. I had wrapped my head around a couple of ways to do this, I feel like more than anything I'm hoping to find that application note that may never have exsisted

[teemuk]

You remember correctly.

At least Hughes&Kettner "Quantum" -series bass amplifiers feature an EQ where Q-factor is higher for "cut" and lower for "boost". Basically frequencies are boosted at wide bandwidth and cut at narrow bandwidth.

They explain that the principle of operation iis copied from "high end" mixing consoles and alike.

Manual explains that this is done in order to minimize interaction with adjacent bands, though I'm not sure if manuals are written by same guys who designed the circuit. I can only see this theory in effect in "cut" mode.

Another explanation why they do this could be human hearing's different sensitivity for louder and quieter signals. Perhaps boosting and cutting -with same Q-factor - does not introduce equal effect to perceived loudness.

Interesting topic, nevertheless.

[Rob Strand]

At least Hughes&Kettner "Quantum" -series bass amplifiers feature an EQ where Q-factor is higher for "cut" and lower for "boost". Basically frequencies are boosted at wide bandwidth and cut at narrow bandwidth.

They explain that the principle of operation iis copied from "high end" mixing consoles and alike.

Yes, the HK circuit is another way to do more or less the same thing.   It's not a bad circuit, quite neat.  In that circuit the input filters contain most of the gain.   So that means it has the most chance of overload but it also has the best chance of keeping noise down.

Manual explains that this is done in order to minimize interaction with adjacent bands, though I'm not sure if manuals are written by same guys who designed the circuit. I can only see this theory in effect in "cut" mode.

I've often seen that claim but  even though the paths look separate they interact.   For example look at a phaser: the all-pass path and the direct path are completely separate but the final mixed signal has a stack of notches.

Regardless of the circuit, the bottom line is you don't want *bad* interractions.   At the end of the day when you set the controls you can visualize the shape of the final response.   For professional use I would say equalizers with bad interactions produce a response which is vastly different to the response in your mind.   This is why a lot of professional equipment often uses completely separate bands in cascade.    Some circuit are sooo far off this image it is scary.  Some circuits look OK when the bands are on full but it falls apart when the bands are at half way etc.   The circuits which mix bands suffer from some sort of interaction  this can be good and/or bad.  The thing is it simplifies the circuit and sounds different.  For instrument EQ you chuck out all the technical arguments and you just want it to "sound right".   After all these years I still don't have the perfect recipe!