A little psycho-acoustic experiment

Let's now try something that will put in evidence that the distinction between melody and noise is only perceptual.

Look at the snippet (and listen the result)

#import Math
#import Modules

#define NCYCLES 2

1 :=> var int compensate;
second :=> var dur d;
const SinOsc s ~> dac;

while(d >= samp) {
  repeat(NCYCLES * compensate) {
    Math.rand2f(220, 880) => s.freq;
    d => now;
  }
  d / 2. :=> d;
  2 *=> compensate;
}

Here we loop until our base duration is one sample Until that we play NCYCLES * compensate notes of our duration. Notes frequencies are randomly chosen between a low and a high bound (here 220 and 880, aka A3 and A5).

At first we hear some boring slow melody. It then gets faster and faster, up to the point we only hear noise. But if you listen carefully, you'll hear that those noises have different granularities. Near the very end, it's almost like if the noise is whisteling, and the thing is that the frequency at which it whistles is the median between the low and the bounds.

Note: line 4, we use a preprocessor define to avoid magic numbers in our code
(this helps with configurability too)