[sdiy] Oberheim Xpander VCF

[Dave Manley]

Tom Wiltshire wrote:
> Yes, it is more advanced, but it is also more complicated.
> 
> Having four LPF stages and a mixer is pretty straightforward, but having 
> four stages that are switchable between LPF, HPF and APF is quite a bit 
> more involved and also removes the possibility of morphing between 
> responses.
> 
> WRT Tim's point about the Doepfer filter, I'm not at all surprised that 
> many of the filter settings in the A107 don't sound that different - 
> with that many (36), I'd have thought they're going to, no? Perhaps it 
> only needs 8 or 10 presets, or maybe 15 like the Xpander is enough - the 
> point is that that is already many times more flexibility than your 
> average analogue filter (24db lowpass or...errrr...well, 12dB lowpass if 
> you're lucky). There are only a few synths (some, but not many) that go 
> beyond this.
> 
> Still, I think this is a fertile area for experimentation, in both the 
> digital and analogue domains. We should all get designing all kinds of 
> interesting multimode filters!
> 
> T.

I haven't followed this thread closely, but thought about doing this a 
few years back, and the expander has different gain coefficients for the 
different modes, and my recall is the filter response is quite sensitive 
to these gain settings.  It would be interesting to rerun the curves at:

http://www.vellocet.com/dsp/CascadedFilterResponses/

with errors applied to the gain coefficients and check the response. The 
error amount depends on the method chosen to mix the outputs.  I'm not 
arguing that it wouldn't be a useful filter even with the errors, just 
stating the given response curves may not be easily achievable without 
precise control and/or a well calibrated circuit.

-Dave