AI-Assisted Production — AI-assisted animation
Still to Motion
A self-taught experiment in AI animation: finished stills brought to life — wind in the planting, animals moving, atmosphere drifting — without touching the accuracy of the frame underneath.
- Role
- Self-initiated, in my own time and at my own cost — learned the tools and models, directed each clip from its still and cut anything that drifted from the real product
- Deliverables
- Short product films created from finished stills · controlled camera movement and atmosphere · a tested sense of which models suit which scene
Brief
This one runs in the same spirit as the scene revivals: my own initiative, my own time and my own budget, spent learning what AI animation can add to a finished image. Full 3D animation is the gold standard, but not every product or campaign carries its production time — so I wanted to know whether an accurate still could be the starting point instead of an empty timeline, and set about learning the tools, the platforms and which models work best on which material.
Approach
Each clip begins from a finished still. Camera movement and atmosphere are generated over it in short, directed passes, and anything that drifts from the real product is cut. What survives is the point of the experiment: the scene comes to life. In the pair below the camera is locked off and nothing about the fence has changed — but the wind moves through the grass, the sheep graze behind the posts, and a frame that was already accurate picks up life and character.
Chestnut sheep fence — the finished still (left) and the clip generated from it (right): same fence, same posts, same paddock, with the camera locked off and only the wind and the sheep behind it moving.