[sdiy] Shruthi 4PM (was re something else...)

[Tom Wiltshire]

Hi Jay,

Something similar had occurred to me too - use a four pole core, with the outputs from each stage mixable via a VCA. My own spin was that this would enable crossfading from one filter type to another. Then I discovered that Doepfer had beaten me to it!

http://www.doepfer.de/a107.htm

If you read carefully about this module, you'll discover that Doepfer came up against the problem you mention too - that first stage that needs to be switched out. There are certain filter types that require it to be there, and the others that don't. Since it is either switched in or out, you can only morph between filters which are in the same group. I haven't yet seen a morphing filter that gets around this, although I don't see why it shouldn't be possible, if you could control the cutoff of the first stage independently and push it 6 or 8 octaves up out of the way, as Olivier did. Still, it's all getting a bit complicated.

The other question about this approach is back to the "sensitivity" issue. Now, it seems to me that we're still waiting for someone to do the work required to actually specify exactly how sensitive the different responses are to component or CV tolerances. I suspect that "_very_ sensitive to tolerances" is how it seemed to Oberheim engineers back in the day. After all, they had to specify 1% resistors in their circuit to make it work right! Blimey!

Nowadays I build all my circuits with 1% resistors as standard, and I expect most of you do too, so I'd think of  "requires 1% resistors" as 
"needs decent parts", and probably not as "very sensitive". Still, Oberheim have given us a guideline there - 1% or better. The repeatability of your CVs and VCAs is going to have to manage that. I think it could be an application where a multiplying DAC would be ideal - repeatable CV and highly linear VCA in one. In fact, Oberheim themselves used one for the FM on the Xpander for exactly that reason.

I hope these comments help you to do rather than put you off. I think it's a worthwhile idea, but it needs some work to make it usable.

Regards,
Tom


On 28 Nov 2013, at 23:14, Jay Schwichtenberg <jays at aracnet.com> wrote:

> I've been thinking about some sort of 'multiple identity filter' for a
> while. I've looked at Olivier's, the Xpander's and the Craig Anderton's MIF
> filter schematics and glanced at the math.
> 
> There is an experiment that I want to do to see if there might be a 'generic
> approch' to implementing a filter with synth modules vs. having a deadicated
> filter. I'll get around to building my TH/TP Mankato filter next year which
> I would use as the filter core so it'll be a while before I can do this
> (maybe someone else has, ideas, comments?). The other module that would come
> into play would be the Blacet VCA2930 Super VCA (quad VCA, mixer, panner,
> morphing) (http://blacet.com/BlacetUserPage/VCA2930user.pdf). The experiment
> would be hook up the outputs from the Mankato (using inverted and
> non-inverted as needed) to the Super VCA doing a mix for the filter output.
> You could statically set the filter 'coefficients' or voltage control them
> to generate the different types of filters.
> 
> Only thing so far I've seen that might be a problem is what Olivier
> mentioned about the first filter stage being bypassed.
> 
> Ideas, comments?
> 
> Jay S.
> 
> 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy