From writing shader code to creating a jellyfish dataset, Adam Ferriss discusses his experimental digital practice

You very well may have come across the hypnotic, morphing works of Adam Ferriss’ before. His work has been featured on the likes of The New Yorker and The New York Times on several occasions, while his prowess in creative tech graces the renowned animation studio Buck, too. Originally from Richmond, Virginia, Adam has been residing in sunny LA for near on ten years, where he creates his mesmerising experiments. But it was originally an interest in the manipulation of photography which first led Adam down the road to coding.

Seeking a way to distort images outside the dark room and without using Photoshop, Adam taught himself how to code using Processing during the “slow times” at his job in a photo lab. As his skills progressed, he graduated onto openFrameworks where he experimented with real-time video manipulations, then, he decided to embark on a Master’s degree in Design Media Arts at UCLA, studying under the digital artist, Casey Reas. Upon graduating, Adam started to “poke around the edges of new technologies,” experimenting with the input and output of non-traditional ways of working or methods.

These days however, you’re more likely to find to find the creative writing shader code for realtime graphics – shader being a computer program used for shading in 3D scenes to craft the intended rendered image. He predominantly works in the realms of augmented reality and face filters, and, he tells It’s Nice That: “I’ve done a little bit of exploring GAN’s and some machine learning type things as well” – GAN being a generative adversarial network, in other words, a machine learning system. But for Adam, “it’s much harder to own the work you create with [those methods] because you’re mostly running an architecture that’s already been made.” He adds, “The art is in selecting the dataset that you train the AI on.”

When it comes to his personal work, Adam is most interested in exploring a specific look or visual outcome. He aesthetically plays with the concept of “AI Spring” in one experiment for example, while in another, he creates a jellyfish dataset and a code for your phone’s pocket feedback. “Often, I’ll just daisy chain a number of small different ideas together and keep parameters until something catches my eye,” he says of his instinctive creative process. It’s a similar, intuition-led exploration that pushes his machine learning process forward. “I just like to try out the new white papers and see what happens,” adds Adam, who then goes on to explain that machine learning projects are notorious for having “tons of tweak able parameters”. They even have their own industry term for it “hyperparameters”, which he keeps himself busy with in a creative operation which seems endlessly intricate.

In another example of his technological triumphs, Adam worked in Spark AR to create a face filter for Buck. Though the software is in Adam’s eyes, “not the friendliest”, it gave the multi-disciplinary creative the chance to use ray marching; a technique known for its “extra nice, gloopy look.” Spark was not set up for this however, and Adam went through a multitude of trials and errors to get the software to accurately track a face. But for Adam, the challenge was exciting, explaining, “I love working on projects that really pushes the limits of what’s possible in a given piece of software.” And even though this problem was “way beyond what Spark was intended to do,” the task at hand presented Adam with something fun to solve, while looking great at the same time.

Looking to the future, Adam is excited to delve further into the realms of webXR and native AR in the browser. Amongst these endeavours, which by no means are an easy feat, he’s also looking to try out some new things in ARkit while also being open to any new interactions that may come his way.