Breaking Things
on Purpose

Glitch art’s pixel-mixing algorithms

Usually when something electronic has a software glitch, it’s as good as broken. An errant couple of letters or numbers in the code of, say, your computer’s operating system could very well leave you staring at a blank screen.

Often though a glitch can create unexpected and interesting results. It’s uniquely evident in digital images, relatively simple entities in which a misspelled bit of code can drastically change the resulting picture — if it opens at all. Like a mutation of DNA, the results may be hideous or they might be beautiful.

Now, a so-called “glitch aesthetic” is gaining ground as visual artists develop and master ever more ways of creating beautiful images through the corruption of their data.

“You’re trying to find this really fine balance where something doesn’t break fully, but breaks just to the point that you can see it breaking,” says Sabato Visconti, a photographer and visual artist who specializes in the so-called glitch aesthetic.

“You could say it’s like torturing an image.”

Images can be “broken” in countless ways — their colors skewed, whole sections moved and mismatched, grids and lines shattering the picture into jagged blasts of color and line that look exactly like what you’d expect to come from the imagination of a machine.

People seem increasingly to be taking to these distortions, whether from their novelty and unpredictability, their association with the modern digital age, or just nostalgia over the 8-bit graphics of the 80s and 90s.

“It kind of carries on from childhood,” says Visconti of his own work. “Like when you’re a kid and you see snow on an analog TV and you’re fascinated by it. I was that kid.”

There are a lot of ways to get an image to glitch. Sometimes it happens by accident, like a memory card in a camera that mistakenly adds or subtracts 1s or 0s with each save, thereby corrupting every picture. Some glitchers will go directly into the code of an image and rewire the data, either with a specific result in mind or simply to see what kinds of cool effects emerge from its digital dregs. Other techniques, like using audio software to edit images or switching image compression formats are also used to tease out nuanced results.

In recent years though, a wave of glitch applications have hit the scene (and the market). Available as home-brew scripts or full-fledged programs to be bought and downloaded, they use custom algorithms to target and alter parts of an image’s code to disrupt it with more predictable results. These applications offer a chance for people without coding backgrounds to experiment with their own digitally interrupted images, giving some degree of control over how the maligned image turns out.

“You get an idea more or less of what you’re going to get, but you can’t really predict it. And that’s part of the charm — you don’t have control over what you’re doing, you kind of rely on the machine,” Visconti says. “If I wanted to a lot of squares and fractal squares I’ll probably want to do a ‘cash smash’ glitch, because those tend to produce squares. If you want to have a kind of wavy distortion that’s almost like a super warped-out VHS effect, you’re going to want to use an audio program for that … It’s definitely more satisfying sometimes when you get down and dirty with the code and kind of just finesse something out.”

While Visconti likes to get his hands into the nitty gritty of the code, he’s also fond of the applications that are getting wider use. For his Images Adrift series (from which all the images in this post come), he exclusively used a single glitch program called Pixel Drifter to get into the guts of photos he’d taken. The program was first released in February by Russian programmer and sound experimentalist Dmitriy Krotevich, who wrote it from the ground up.

“The algorithm is based on comparisons of pixels to their closest neighbors,” Krotevich told me by email. “Each pixel has a ‘power’ value which is determined by the pixel’s luminosity or position on a screen. Normally, pixels with a higher power value can travel farther, but this behavior may be altered by choosing a certain combination of options.”

The technique is called pixel sorting. Whereas some approaches might target a specific line of code to be altered for a glitch effect in an image, this one imbues each of its thousands of pixels with a set of instructions that give them a self-determining path along the screen. The algorithm scrapes them along the frame to new points based on their color, brightness, position, and other variables, resulting in a completely contorted version of the original picture. There’s room for the user to influence the result, but the overall effect is determined by where the individual pixels decide to go.

“It’s not just like a one-shot deal, sometimes you’ll get an amazing glitch on the first try but usually you’ll run it ten, twenty, thirty times and you’ll have all these candidates,” Visconti says. “The glitching process can be long, and sometimes arduous.”

As much as knowing which effects to use, an artist’s role in using these applications and techniques is largely about knowing when to stop. The images in Visconti’s series retain the traditional sensibilities for composition, form and color that all images are judged by, but what makes them even more interesting is that they were formed from the shattering of other images that already looked fine.

It may seem a bit haphazard — a push of a button, however many times, and the decision of whether to keep the result or keep breaking it. Rinse, repeat. But beyond the shredded images, the artists, hobbyists, and amateurs who are drawn to these glitches are exploring a deeper question, consciously or not.

As we upload our lives in pictures, we often have the sense that we’ve sealed our images into a cloudborne vault. But there’s no guaranteed protection. By tearing apart the digital material that forms image files, Visconti and his glitch-art peers demonstrate that our photos are as error-prone as our memories.

“The truth is that they’re just as impermanent, they disintegrate because digital files disintegrate, they break easily,” he says, “You change one byte from a jpeg and then it’s not really even the same image.”

You can follow Doug Bierend on Twitter at @dougbierend. Subscribe to re:form’s RSS feed, sign up to receive our stories by email, and follow the main page here.

Artwork by Sabato Visconti