I can't really say whether you'ld be able to reproduce the sherman filter sound with my new filter. As Suburb_Animal mentioned, the sherman filter is not just only the filter, but also all the other elements around it.
You could perfectly rebuild the routing and functions of the sherman, but even if the filter modules themselves might sound the same, it might still sound totally different at the output, because of all the different stages (you'ld also need to model the used VCA's and all other components that influence the sound).
That said, as there is a kind of compression inside the filter loop, this also alters the resonance and maybe you could make the resonance more or less respond like the sherman's filter response.
All the filter respond-types are all there at the same time (which are just different combinations of the sine/cosine pair and the original audio input), so also the morphing between filter types could be done.
But for this the code should be rewritten to either add all the combinations as outputs (easiest, just alter the code to output the types directly instead of switching between them, though the output limiter would have to be re-used for each output and this will add quite some cpu use). Or I could add another version that's able to morph between "several selectable" or "all" filtertypes.
As for the input gain stage, Suburb_Animal seems to say the distortion responds differently between CCW and CW. I'm not sure about this, I own a sherman filter2 and it seems to me that the distortion is also depending on the input level of the original audio level. If this input audio already goes over the max level, this will distort the audio before the gain stage (which mainly happen at the transients, thus adding distortion to these), even if the gain of the sherman itself is very low. If you lower the input volume before going into the sherman, this distortion might not be happening. Then when adding more gain, the transients might start clipping after which it will just push everything towards a compression type distortion.
Anyways, I think it will be réally hard to make an exact copy of the sherman, bút, making a filter with the same kind of functionalities should be pretty easy and you could even add some more functionality.
So for the input stage, you might want to try out a soft distortion/clipper or compressor with the pre-gain stage.
Then you'll have the gate for the envelope, which can either be triggered by the audio or an external trigger. Here you could use an envelope follower that creates a gate-high when going over a treshold (factory/math/>c), use a logic OR to combine with the trigger input, and send this gate to an envelope generator.
Now you can add something extra with the envelope. I don't know the exact curve modes of the sherman filter (linear/non-linear/combo?), but I have made several envelopes that have customisable curve modes, so you could dial in the exact amount you'ld like, whether it would be logarithmic, inversed logarithmic or just plain linear. Also, you might combine this envelope output with the envelope follower's output, to keep it from retriggering as long as it's above a certain level or only after a certain stage has finished.
The LFO on the sherman only has triangle and saw, but personally I would have loved a bit more options.. so why not just use a LFO with more waveform modification options? And besides, as it's hardware, for the sherman they had to decide what routings and options to use and this puts a limitation on how they could use the available area on the front panel. But we don't have this limitation, so I'ld just go for a way that allows to do the same as sherman, but also allows a lót more. Eg. more waveforms, waveshape capabilities, multiple trigger inputs with multiple LFO's/envelopes, delayed copies of the envelope (thus a k-rate delay with feedback, repeating the envelope while fading out).
As for the filter routing, I always missed the option of where to place the second filter. The first filter's output or the main input? Using it as a nice stereo filter was not really an option as the second output was always dependent on the audio coming from the first filter's output (I would have loved an extra input for the second filter, which would be normalised to the first input if nothing is plugged in, and an output for the gain stage, so I could choose which input to use for the second filter).
Long story short..
if I were you I would just start a small, simple patch with the basic elements and functionality that the sherman uses, even if the modules don't really sound the same.
From here, I'ld make some first changes when it comes to routings (eg. making true stereo filtering an option)
Then, I'ld start changing modules, firstly the filters, because these might give you more options then the sherman filter has, allowing more routing/modulation options. Then the other stages like distortion and envelope.
Then re-do the routings, so you can make use of all the added options.