Toxic threads
Printing color guard T-shirts with an ‘infected’ look
By Lon Winters
Images courtesy of Lon Winters
There’s something about a small town. At Elizabeth High School in Elizabeth, Colorado, the Winter Guard performances share a bit of magic. Not polished, but great, nonetheless. Watching a guard performance in a town like this, for a moment, we forget we’re sitting on cold metal bleachers. That was the backdrop for this project: Not a retail brand or a corporate rollout, but a handful of shirts for the guard. Just 10 or 20 by the time the parents got theirs. Amaya was a first-year choreographer for the EHS Winter Guard. She grew up here. Now she was back from college and running the program. There’s pressure in that, whether she admits it or not. Amaya is a part-timer here at our shop, Graphic Elephants, while her mother has been with us for a few years and has become somewhat of the house mom.
The theme and name of the show was “Infection.” Not subtle, but a kind of visual storytelling. The shirts would build identity and energy and give the team something to rally around. We didn’t get much to go on, just a couple low-res JPEGs, some thoughts, and a timeline that never seems to do us any favors. What was different was the tool we leaned on. We needed to start somewhere, and instead of opening up Adobe Illustrator and staring at a blank artboard, we started with a conversation. What does “spooky and hazy but still clean” actually mean? How do you take “infected” and make it school appropriate without losing the edge? These were the directives from Amaya. So, we did something a little unconventional for us. We asked ChatGPT what prompts we might use in this case.
The quality of what you might get out of any generative AI system is directly tied to what you put in. We didn’t just type “zombie gas mask” and hit enter. Instead, we used phrases like “diffused glow,” “toxic atmosphere,” “high contrast subject,” and “clean silhouette.” We realized pretty quickly that prompting isn’t all that different from art direction. The early generations came back reasonably close. Not perfect, but close enough to get a reaction. Amaya knew what she wanted, so back and forth we went. Tweak the mask. Adjust the posture. Push the glow. Working with AI was frustrating. We couldn’t push a line or adjust a curve. We had to regenerate. Each was a new roll of the dice. We chased our tail for a while to get changes without destroying everything else. Eventually, we landed on something that clicked. A figure in shadow surrounded by a toxic haze. We chose neon green to symbolize that toxic environment surrounding our figure. It told the story.
The problem? It wasn’t usable for print. Not yet. Raster based, inconsistent edges, unpredictable color. AI got us a look, but it didn’t get us a separation-ready file. Not even close. So, we did what we always do. We rebuilt it in Adobe Illustrator. We did some tracing but, like most auto-traced vectors, it was pretty rough. There were 27 colors, give or take. Gradients and subtle transitions in hard steps. We started by isolating color groups using the Recolor Artwork panel as it lets us see the entire palette to consolidate similar colors, hues, and tones. Colors that varied slightly were merged and grouped. The gray structure that defined the mask and background got its own layer. From there, it became a manual exercise. Pathfinder to unify shapes. Direct Selection to clean up. Expand Appearance to flatten.
We landed on five colors. The white printer or base plate, Flesh, Brown, Sick Green, and Gray. That palette didn’t come from the art, but more from the inks on press. By printing colors over white and directly onto the black fabric, we doubled our tonal range without adding screens. We created spot colors for each ink, then built separate layers to represent on-base and off-base areas. An overall choke with some gutters on the white printer would keep everything clean on press with minimal flashing.
The mesh selection included an N-166 for the white printer for enough opacity, N-272s for all the other screens for minimal ink deposit, all at a tension of 45 N/cm2. Gray was first and established structure without needing a base. Then the white printer or base, a flash, smooth, and into the wet-on-wet sequence with the balance. A second flash locked it, and finally, an inline transfer machine to smooth everything out.
With the shirt in hand, everything we tried to anticipate became visible. The green radiated for the glow. The hooded figure’s folds of the robe used the same ink but had two values. Almost like we planned it like that.
The back print should have been easy. Sponsor logos, simple type, one color. We didn’t think much about it. Until it fell out of registration. The top edge had a white halo. The bottom edge had a dark green shadow. Before we could fix it, Amaya saw it and loved it. We spend so much effort chasing perfect registration that we forget sometimes the human eye likes a little imperfection. It added energy, movement, and it echoed the theme of the show. We call that a “happy accident,” a phrase we stole from Mark Coudray [and Bob Ross] years ago.
At the end of the belt, we had a short stack of shirts with a design that felt bigger than the run. A figure that carried the story. It was part of the production.
Amaya summed it up. “After working with Graphic Elephants, I knew and trusted that they would be able to help me make what I wanted. In the end, I think they did a great job helping us express our show through our shirts.”
Lon Winters is founder and president of Colorado-based Graphic Elephants, an apparel decorating studio and international consulting firm. He was inducted into the Academy of Screen and Digital Printing Technologies in 2013 and is widely recognized for his contribution to the graphic printing industry. Learn more at graphicelephants.com.
