[brainstorm] Sound deconstructor for AmigaKlang and similar - technically feasible?

[saimon69]

Ok, this topic is growing in my head so will expand later on

I wonder if fractal compression is so advanced that a sample could be deconstructed in a mathematical expression and then rebuilt via AmigaKlang before runtime to save disk RAM - ok i guess might take several days to re-render the sound but a lesser detailed version?

[edit] Am aware that we are talking of a technique similar to the fractal compression of mp3, however there are cases in which a deconstruction of a waveform coming from, in example, an OPL2 or OPL3 chip could be read, reverse-engineered and stored to be re-rendered at runtime same way as the rest of an AmigaKlang engine is working now, to avoid darn imput of all parameters (we could even import instruments from Deflemask in example)