• 01.31.13

Finding New Beauty In Old Images, With Algorithms And Code

With the help of open-source code, Adam Ferriss re-sorts pixels from random images into sublime new compositions.

On the blog where Adam Ferriss dumps all his work, you’ll occasionally come across things you recognize–photographs of some plant life, say, recast in technicolor–but for the most part, the endless scroll takes you through a dizzying variety of computer-altered art. Some pieces simply look like the beautiful by-products of busted-up LCD screens; others look like the machines themselves trying to pour out their hearts and souls, if such things existed, in jagged, glitchy form. But all of the pieces, however affecting the final result, are made with the same tools–a series of pixel-sorting programs that transform source images into something else entirely.


Computer-generated art has been around for decades, of course, but the techniques for creating it are constantly evolving. For this particular collection of greyscale patterns, Ferriss used code written in the popular programming language Processing that employed two techniques: pixel sorting and cellular automata. Starting with a photograph of a wave crashing against a craggy shore, Ferriss first used code to sort the pixels from brightest to darkest (his program was a tweaked version of one written by fellow artist Jeff Thompson). Then Ferriss made a greyscale version of that image and sorted its pixels again. At this point in the process, the original ocean scene is totally unrecognizable, having been rearranged pixel by pixel into a bouquet of monochrome diamonds.

That image was then run through a cellular automata program, which randomly places a number of “seed” pixels in the image, which organically affect those around them. It’s the harnessed for the sake of art. “Fewer seeds gives me larger shapes, more seeds generate closer compact shapes,” Ferriss explains, but in the spirit of things, the artist still strives for the chaotic. “I like to feed in values that I know will cause the program to choke in order to generate unexpected results,” he says.

But you don’t have to have any interest in the nitty-gritty technical details to appreciate the images; the combinations of shading and color and line have an aesthetic purity that you don’t get with much “glitch” art. And what you see here are just cropped versions of the full-size images, which you can see in all their flickery beauty on Ferriss’s page. The “wobble” effect, as the artist terms it, was actually an unforeseen product of his particular batch of code, but he came to see it as “a necessary part of viewing the image.” We’ve tried to capture the experience in these GIFs.

Still, while Ferriss considers this group to be “very nice formal and compositional studies,” he actually prefers some of his other work in which the source image isn’t so thoroughly destroyed, like these rearranged but still recognizable alterations of Petrus Christus’s Portrait of a Young Girl.

In a way, these types of altered images have one foot in the realm of the remix. They aren’t things created so much as they are transformed and transmuted, and Ferriss says his interest partially lies in that process of making “new derivative photographs” that “contain all the same pixels as their original, but appear as completely different forms.” But at the same time, there’s something alchemical about those algorithms–something that unlocks new types of patterns and combinations and compositions that we might never quite find otherwise.

At this point, it’s just a matter of who can write the code to get there first. “It feels as if there is still a vast unexplored territory of imagery out there,” Ferriss says, “so part of this for me is a race to discover it.”

See more of Ferriss’s work on his site.