Flutes are a famous instrument to model in a modular synthesizer, I believe because the flexibility is required.
https://www.phys.unsw.edu.au/music/flut ... nB/A4.html
A quadratic rather than sigmoid waveshaper mode could be applied to achieve variations in timbre (compare the first example timbres versus the final "soft" timbre), but even given that the spectra could be reproduced perfectly (I have no doubt a great approximation is trivial) the breath timbre would be more difficult with a single voice.
So, ultimately there are a very large number of small changes and new features that could allow Xhip to produce better flutes and many other timbres. The original 1999 "Acid Tracker" synthesizer used a quadratic shaping function x*(2-|x|) rather than the sigmoid (1+d)*x/(|x|+d), so I've wanted to add that much softer function for quite some time. The old "sine" waveform applied a tweaked version of that quadratic function with bias to the triangle waveform and had a much more interesting flute-like spectrum than the current sine function.
It's definitely an issue I'll need to apply some focused study toward solving, and so for now it's likely best to just "let it be". Adding too many things to the to do list makes it never grow any shorter

