I've been having fun lately with
Andre Michelle's Tonematrix. The site offers a blank 16 x 16 grid representing a musical score on which the
x axis is time and the
y axis shows ascending notes in three five-note octaves. By filling in squares one can create simple repeating songs that are quite mesmerizing. The rest of this post will make much more sense if you take a moment now to
visit the site. Click on a few squares and watch what happens (making sure to have your computer sound ON).
I've found that it's virtually impossible to create an ugly song. You can enter a whole slew of notes completely at random and wind up with a nice mix of chords, interplay, and syncopation. I've entered designs, blocks, letters, figures from
Conway's Life. Each one has a unique song that's surprisingly euphonious (and, perhaps unsurprisingly, monotonous).
Curious about the photographic implications, I began playing around with b/w square photographs as the root of my designs. The first was Nancy Rexroth's
A Woman's Bed, Logan, Ohio, 1970:

I converted the image into a 16 x 16 pixel array...

Saved it as a two color gif...

Then copied the pattern into Tonematrix...

Here's what it sounds like:
Next up was Arbus'
Young Brooklyn Family Going for a Sunday Outing, 1966:
Harry Callahan's
Eleanor, 1949:
Irving Penn's
Portrait of Truman Capote, 1965:
As you can hear, each image generates its own song which isn't quite the same as any other. I think they're a bit like audio fingerprints or whale songs uniquely keyed to the individual.
Beyond that fact I'm not sure what significance they may have. I thought about creating a quiz in which I'd have people guess an image after hearing its song but quickly realized it would be impossible to solve. For color photographs I can imagine other possibilities like a 3D graph with color representing multiple instruments across the
z axis.
I think perhaps these audio clips are best appreciated as bizarre little experiments just weird enough to be of general interest.