I’ve mostly completed the main part of the Synchrodyne deep dive. It didn’t take that long after all. I had some confusion about the behavior of CLK IN in the PLL section, but it turns out it was an undocumented change — instead of overriding the internal VCO, it is a logical OR with the output of the PLL going into the filter, which the designer thought was more interesting. It’s available on the expander module, but since they stopped building expanders before the last batch of Synchdrodynes he went ahead and put it on the main module. There are a couple of things I want to try with that.

Overall I’m much more in favor of Synchrodyne than I was before. The key thing I understand now is its tracking behavior — how to make it stable, and how to carefully make it unstable and take advantage of that instability for added texture and even FM bell-like pings. But also the usefulness of the main folder, and several patch ideas for the filter as well as the PLL and VCO (e.g. clocking Drezno in ways that add noise to a signal running through it).

With the official release of Serge GTO, I thought I’d better get going on that Shapeshifter deep dive and make my decision about keeping or selling it, or risk them selling out of the first batch.

My doubts in Shapeshifter were based mainly on the fact I was using so few of its features, and maybe a different design of complex oscillator would work — or just rely on my other oscillators to handle TZFM duties.

So one night I did some tests and confirmed my biases: the wavetables are mostly garbage (*), the tilt sounds harsh and unpleasant, the sync sounds harsh and unpleasant most of the time, the bets combo modes are the ones I could do with Mystic Circuits Ana, the chord mode isn’t as nice as other modules, and so on. The wavefolder is nice. The sines are SUPER pure and clean and the TZFM is by far the smoothest and cleanest of everything I have, but that’s more of a neutral fact than an advantage to be honest.

(*) I prefer wavetable banks to have a progression of shapes or spectra from one end to the other, not an unoragnized collection of cycles with little relation so that scanning through it sounds random. The latter style of bank is very common, but leads to a tedious hunt for sweet spots. If you have a simple wavetable oscillator like Kermit with one bank like that it can be fine (if the oscillator otherwise has good character); if you have 128 wavetable banks and 103 of them have no organization and 50 of those have not one single pleasant cycle among them, it’s tedious.

I was thinking about the new wave of TZFM-capable oscillators that use the very capable newish SSI2130 VCO chip. There are a variety of them, with different designs and features built around the core, and generally they sound pretty great. Perhaps the flagship of those is the Cosmotronic Vortex, a dual oscillator that features two variations of the oscillator with different wavefolders, two different wavefolders, ringmod, a nice FM bus for both exponential and linear FM, and a smart and flexible overall design, and is exactly the size of Shapeshifter.

Or, I could go with a Rossum Morpheus z-plane morphing filter, which makes for a pretty fantastic resonator according to a few adventurous souls, and still have room for one of the smaller 2130 VCOs such as Shakmat Banshee Reach or Befaco Pony, or some other thing instead. But I decided I’d (A) play with Instruo Cs-L in VCV Rack, and (B) sleep on it.

I woke up early because it was a bit too warm, and decided to fire up Shapeshifter again… and stumbled into a couple of amazing things fairly quickly.

  • the Harmo3 wave bank consists of smooth, clean sines advancing through higher and harmonics in a smooth, clean way. This is really nice for things like FM or combo modes.
  • the delay! I really have not played with this enough. The delay is clocked by osc2’s frequency, is mixed into output 1, and the feedback level also serves as a mix level. If you set osc2 to LFO rates it’s a “normal” rate delay, but at audio rates it acts as a resonator…
  • Osc2 still works as an oscillator with the delay engaged, and phase and waveshape do not affect the delay. So you can phase modulate osc2 from out1 (which includes the delay repeats!) — whether Osc2 is running at audio or LFO rates — or bypass the oscillator and use out1 for a wavetable lookup, and patch that in stereo.
  • You can also FM osc1 from out2, including the phase modulation you’ve applied… and if you’re running osc2 at audio rates, the resonator feedback will smooth out the usual weird harshness of that kind of cross-modulation. You can also use combo modes, chords etc. all to pretty great advantage. And that Harmo3 wavebank is really nice for sweeping the harmonics that are fed back through a resonator…

I need to experiment more in this area, but what a difference these discoveries can make. I’m sure Vortex is cool but I will be keeping the Shapeshifter, thanks.

Exploration of Shapeshifter is going to take a bit of a different format from Synchrodyne; the manual for Shapeshifter is more complete (and up to date) and it’s easier to understand both technically and intuitively what’s happening with it. It’s just that there’s so much possibility there, I wound up leaving a lot of it inadequately explored.