[sdiy] Alternative Q control in SVF (was Re: 'Reverse' connected 2164 approach for HP filter stage purpose..)

[Tom Wiltshire]

On 11 Mar 2014, at 22:24, rburnett at richieburnett.co.uk wrote:

>>> The only difference with controlling the Q by altering the integrator gains independently is that it buggers up the level matching between the LP, BP and HP outputs if you are going to switch between them in an audio application.
>> It saves you a VCA for controlling the resonance, but aside from that,
>> I don't see much advantage.
>> Do all the output levels alter, or just the ones with integrators?
> 
> Only the BP output changes in amplitude when compared to the traditional way in which a standard SVF is controlled with three variable gain cells.  The best way to describe the effect of changing Q by independently varying the integrator gains is to say that it makes the peak of the bandpass response always be at the same amplitude, (e.g. unity, 0dB).  So, where the BP peak gain increases along with the LP and HP gain in the conventional SVF, the 2 gain-cell variant exhibits an overall decrease in amplitude at the BP output as the Q increases such that the PEAK GAIN IS CONSTANT.  I can understand why they showed this circuit in the paper you referenced, because this is the type of behaviour that you might want for a parametric equaliser.
> 
> -Richie,
> 
> PS. Tom, see attached GIF of frequency responses, which will probably get stripped off the list copy.<svf.gif>

Ok, so aside from a different BP behaviour, it works the same. Richie's GIF shows what you imagine - his description is good.

It would potentially allow JP to build his variable-width 4-pole bandpass filter with two SVFs (one high pass, one lowpass) and then to add voltage-controlled resonance without needing any extra VCAs, just a few more op-amps. That's not a bad trick to know.

Thanks,
Tom