One photo has to carry the whole Bad Times AI body
There is a single slot. The pet in that photo is the only performer from the first frame to the last. Include the face, fur, body, and paws, and any collar or clothing that should remain. A crop of only the eyes makes the later paw gestures guess at a body you never showed. A photo of two pets makes the intended performer ambiguous. If you want a cat instead of a dog, the cat still has to occupy this one role for the entire 30 seconds. The source motion was built around one animal facing the camera, so a very different body can bend ears, legs, or a tail in ways the reference never had. That is a reason to review the full take, not a reason to upload a second picture of the same pet from another angle. Replace the photo if you want a different animal. Do not stack alternates in the hope that the edit will pick one.



