Hey, thanks for your questions!
Correct, the SVGs generates the initial sound, very much alike a pure oscillator wave. The more complex the shape, the more complex the timbre. The modular is simply applying complex modulation and wave folding to the initial signal, in addition to applying a delay to the half that goes out as sound.
Not sure if I understand your second question but yes, you could compose the music first, either as a curated sequence of SVG forms, or as a simple oscillator waveforms from, say, a modular. That would remove the need of Oscistudio generating SVGs.
There’s no way of turning the signal back into an SVG as far as I know. Interesting idea. It would simply display the sound on the XY scope. That would go into the direction of more traditional oscilloscope art.
From my point of view, the power and intrigue of this method comes from exposing very simple SVG forms as sound, then deconstructing and evolving those shapes into something else, all the time giving audible feedback of what’s happening.