# A frame from your video isn't a print yet > The still that stopped you scrolling makes a weak shirt, and the reason isn't resolution. A cinematic frame works because of everything around the subject — a print has none of that. August 17, 2026 · https://justog.club/stories/frame-from-an-ai-video Generated video got good, and one of the side effects is that people now have folders of stills they like better than anything they meant to make. A frame from second eleven, the light doing something it wasn't asked to do. The instinct is to crop it and print it. That mostly disappoints, and the usual explanation — it's too small, it's compressed — is the wrong one. ## Can you print a frame from a video? Yes, and the file questions are the least interesting part of it. Resolution, compression artefacts, the fact that a video frame is a lossy still pulled out of a lossy stream — all real, all handled on this side. That's pipeline work, not a decision you have to make, and it's a poor use of the attention this actually deserves. The interesting problem is compositional, and it survives any amount of cleanup. ## Why does a great frame make a weak shirt? Because a cinematic frame works with its whole rectangle, and a print doesn't get one. Look at what a still like that is doing: a small figure, a huge sky, fog holding the middle distance, the weight of the shot living in the empty two thirds. It's beautiful *as a rectangle*. Cut the rectangle away and what's left is a small dark shape. Which leaves the two bad options people actually take. Keep the rectangle, and you've printed a photograph on a shirt with [a visible border around it](/stories/what-makes-a-graphic-tee-look-cheap) — the most reliable way to make a garment look like merchandise. Cut around the figure, and you've thrown away the part that made the frame worth keeping. ## So what do you do with it instead? Hand it over as a **reference** rather than as the artwork. A single image can play three different roles, and which one it plays is decided by what you say about it: - **The image is the design** — adapt it faithfully into a print, keeping its values and palette. Right for something already composed as a graphic. Almost never right for a film still. - **The image carries the subject** — the figure, the pose, the silhouette are preserved and everything else is free to change. This is the one you usually want. - **The image carries the style** — its palette, grade and texture get applied to a different subject entirely. The shirt in the picture above is the second one: same figure, same stance, same ridge, redrawn as a single-colour engraving that has an outline of its own. ## Which frame should I pick? Whichever one has the thing you recognise in it. Sharpness isn't the criterion, and that's not a consolation. The frame that reads best is often not the prettiest — a blurred pass where the coat shape is unmistakable beats a crisp one where the figure is ambiguous, because what a redraw needs from a reference is **information, not fidelity**. It's the same inversion that applies to [a damaged photograph](/stories/restore-an-old-photo): the reference has to carry the identity, not the quality. Motion blur and compression are properties of the recording. They aren't inherited by something that was drawn from it. ## What survives the trip, and what doesn't? Structure survives. Atmosphere doesn't, unless it gets rebuilt as structure. Silhouette, pose, the relationship between figure and horizon, the shape of a coat in wind — those are geometry, and geometry travels. Volumetric fog, lens flare, a shallow depth of field, the particular grey of overcast light: those are lighting on air, and there is no air in a print. Which is why the fix is usually a translation rather than a rescue. A misted valley becomes engraved hatching. A glow becomes a halo of negative space. The mood arrives by a different route because the original route doesn't exist on cotton. ## Does the whole scene have to go? No — scenery is fine as long as it's *inside* the graphic rather than behind it. The distinction that holds: a scene may live within the artwork's own shape — a vignette, a silhouette, the inside of a panel — but never as the canvas the design sits on. The ridge line and the mountains in the shirt above are still there. They stopped being a background and became part of the object's outline. That's usually the move that makes a cinematic still work. Not deleting the world, but letting the world become the shape. ## How do I get one made? Export the frame, bring it to [JustOG](/) with a sentence about what it is to you, and the directions come back as a set. Pick one, drag the crop frame, see it composited on the real garment, and it's made to order and shipped. Designs other people have published are in the [shop](/shop). The frame was a moment. A shirt has to be an object.