To answer the OP's question: I think when someone dives deep in learning synthesis, the more one comes to appreciate how well-designed Zebra actually is. For that reason, I've come to think of Zebra as a synth for the truly dedicated.
I've had Zebra for a few years now. Several times, a more "specialized" synthesizer would become my main... but after learning a lot from a given synth, I'd return to Zebra with a new understanding of synthesis. And each time, I've had an epiphany moment that I can usually get a similar and/or better result with Zebra (not always of course -- sometimes the architecture of a dedicated synth can be pushed to greater extremes which Zebra just can't replicate...which is exceedingly rare...hint: it usually involves samples). You can think of me using other synthesizers as "Research and Development," and then usually using Zebra for the final output, where thoughts have been refined and crystallized.
As a recent example, I've been studying FM synthesis in-depth the last few months. I've demoed a bunch of synths, bought a few, have read several books on the subject, studied a bunch of DX7 patches and factory FM patches for other synths like FM8 and Sytrus, algorithmic vs matrix FM, etc etc... I've come to a number of conclusions about FM synthesis that made me realize just how well-implemented Zebra's FM actually is.
To put things starkly, let's think about a 6-op FM synth for a moment (limiting the operators to sine waves, for clarity). Let's imagine a single carrier being modulated by 2 modulators in series (ie, Mod A -> Mod B -> Carrier). What we are doing is using two simple waveforms to generate a complex waveform, which in turn modulates another simple waveform.
Zebra cuts out the middle man by starting with a complex waveform to modulate a simpler waveform. You don't need a bunch of operators because you can cut straight to the point. And you get to take advantage of Zebra's wavetables for time-varying waveforms, AND you get to take advantage of Zebra's extensive modulation options (which I have found sorely lacking in FM8... bleeding-edge sound design with FM8 is reliant on host automation, IMO). Now, this seems like a "well duh" moment...but once you truly grasp what this means, you begin to understand that Zebra's practical FM capabilities are more extensive than many dedicated FM synths. One trick I've picked up is to see what spectral content a modulation scheme creates, then simply model that spectrum in a Zebra wavetable (research and development, so to speak).
Also, Zebra's FMO modules are better than people realize. For waveforms which are typically useful in a traditionally musical context, the FMO module has things covered for carrier waves: sine, triangle-like, square-like, sawtooth-like. The FMO doesn't fulfill the fantasy of arbitrary waveforms which I'd like...but I've come to realize what a fringe case that actually is. And when I DO want something that fringe via FM, Serum (which I also own) can fulfill that desire (because 2-op is actually all you need in that case).
Now this isn't to say Zebra replaces FM8 or Sytrus. It doesn't. You can get results with those that you can't or would never *think* to get with Zebra, especially without having gone through the experience of programming a matrix FM synth. But again, after having gone through that experience, my understanding of Zebra has increased and I now know new, useful ways to extract sounds from Zebra that I can't get from other synths.
Last edited by KBSoundSmith on Tue Sep 25, 2018 9:42 am, edited 1 time in total.