Car photo and motion reference explained

One upload defines the car

The upload slot accepts one car photo. It supplies the appearance reference, including body shape, paint, wheels and recognizable features. There is no second character slot or role assignment in this generator.

The video defines the sequence

The built-in reference supplies the ramp, flight path, camera framing and timing. You can watch it before generating. It is shown as a motion reference, not as proof of a result generated with your photo.

Keep the two sources distinct

The prompt asks the AI to preserve the uploaded car while transferring the reference motion. It also asks for original sound effects and no copied publisher marks. Model output can still depart from either source.

Check identity through motion

Watch the vehicle across the whole clip, especially where the camera angle changes. A clear first frame alone does not establish consistency. Note any mismatch before deciding whether a different photo is worth another paid attempt.

Which source controls each part of the result?

The car image supplies evidence for the subject: its shape, paint, wheels and visible accessories. The fixed video supplies the scene sequence and motion target. Uploading a landscape car photo does not request a landscape output, and uploading a parked car does not request a stationary result. The generator attempts to combine the two sources, but it does not perform a frame-for-frame cutout replacement.

If the body looks different, examine whether the input clearly shows the missing feature. If the car turns or lands differently, compare that section with the motion reference instead. These are different failure modes and a new crop will not necessarily solve both. The result may reconstruct surfaces that were never visible in the single image; details on the far side of the car should not be treated as verified replicas.