Image-to-video
Image-to-video describes producing a moving scene from a supplied still. The image fixes composition, light and figure, the model adds the movement. For us this is the normal case in production, because the result can be approved beforehand.
Why this is the more reliable route
With text-to-video the model decides everything not stated in the description, and that is a lot. With image-to-video the appearance is already fixed. The remaining uncertainty concerns only the movement, and that can be narrowed further with a second image as the end point, see keyframe.
Hence our process: the approved styleframe becomes the opening frame of the movement. What you signed off in the still is what appears in the film. That is the difference between a production with approvals and an experiment with an open outcome.
How your own material comes in
An opening frame does not have to be generated. Product photography, brand images or a shot of your team can serve as the basis. That is exactly how product films are made: from real photographs of the product, so that shape, colour and packaging are right rather than invented.
Frequently asked questions
What is image-to-video?
Image-to-video describes methods where a generative model produces a moving scene from a still image. The image provides composition, light and content, the model adds movement across a few seconds.
Can you use our own product photos?
Yes, and that is the recommended route when the product has to be exact. The better cut out and lit the photo, the more usable the result. Corrections to labels are done with inpainting.
How much movement is possible?
Camera moves, gentle movement in frame and material effects such as steam or liquid work well. Large changes of place or complex action inside one shot do not; for those you cut to a new shot.
Related terms
Read on: All terms in the glossary • Pricing
You have product photos?
Then you have the starting point. Show them to us in the first call.
Book a first call