splines (was:Modulus web page)

[Thomas Hudson]

"Budweiser [FTS]" wrote:
> 
> Hey! That makes a lot of sense.... so that Engeering IT mathematics 2
> subject i took last was worth it! I've used NURBS before in some computer
> graphics programming (and seen other apply it much better than me!)
> But applying it to waveforms never even entered my head! I'm sure
> some fantastically fluid waveforms could be made this way...
> 

Now the next step I never got to (left as an exercise for the reader),
was to use a curve fitting program to take two unrelated waveforms,
and approximate them with two nurbs, each with the same number of
control points. Then, changing weights and moving control points, 
morph between the two waves. Seems like it would produce more
interesting results than the other ways I have seen audio morphing
done.

Paul wrote:
> .well done..
> any brave soul want to do this in analog?? ;-)

Man, it was hard enough in software. My understanding of nurbs is
not that great, and docs are hard to find. In fact, the way I
currently calculate the nurb still has an x and y component (straight
from CAD), so the curve can turn back on itself, which is obviously
not much use for a waveform. So my method could be simplified.

Eventually I plan on creating an opcode for CSound.

Thomas