line to audio_ some basic research

I have a developed a system for making audio with a given waveform in SuperCollider.

I have been considering the issue of the basic waveform in the synthesis. My initial notion was to translate the image of the dancer, in say a still pose, into a waveform which is read by the sound synthesis engine. There was also the notion of using the full image as the frequency (mapped to vertical space of the dancer on stage), panning (where the read head/light line directs) and amplitude (based on the intensity of white, where black is silence). This all assumes a black and white world! Maybe that is worth going toward.

The problem that I am working on now is how to translate the still image into a waveform. What I need is the position (normalized to -1.0 to 1.0) and intensity (scaled 0.0 to 1.0) of the most prominent (the whitest) pixel in each column.These values can be used to write a waveform. I have just cracked the book on Processing and this will be a good means of getting that data. Once I have it, I can fill an array and read that into SuperCollider.

The waveform will be cycled x times per row in order to correspond to the divisions of frequency space settled upon.