[CM] Finding the duration of a time-varying resampled sound.

[Bill Schottstaedt]

 > (define a '(1 1 0 5))
 > (define b '(1 1 0.5 0.5))
 > (define c '(0.5 0.5 0 0))

These envelopes don't make any sense -- the x axis is
going backwards?


But my original function wasn't right either (of course...);
here's another stab and if it's wrong, I'll go home and
test the next one!

(define (average-inverse x1 y1 x2 y2)
   (/ (* .5 (+ (/ 1.0 y1) (/ 1.0 y2))) (- x2 x1)))

I'm not sure it's the right idea that interpolating a
point on the existing segment should then add to the
original -- since we're actually dealing with the
inverted envelope, the interpolated point probably
changes the slopes of the "true" curve.