Here’s something a lot of people don’t realize: Generative Fill AI also works on video (sort of; we’ll get to that). Wait, Photoshop has video? Yes, it has for a long time, even in CS6. Remember the Extended version? Because Generative Fill creates its own layer, it has no issues sampling a video frame. Because each generated image is unique, it can’t replace a moving object in a video, because only one frame would match, so you can’t add that duck driving your car yet. If, however, you have a video where there’s no movement on one of the sides, you can use the new General Expand feature of the Crop tool in the beta to extend, expand, or change the shape of your video on that side.
First click on the video group folder in the Layers panel to make it active because you want the new Generative Layer to appear above the video group. Then use the Crop tool (C) to make the canvas larger on the side where there’s no movement. In the Contextual Task Bar, type in what you want, or leave the prompt field blank, and click Generate. Now you can see that you’re able to extend your video with the AI pixels. This is basically a cinemagraph.
This tip originally published in Colin Smith’s “Photoshop Tips” column in the August, 2023 issue of Photoshop User magazine.
