Question

A media team uses Azure OpenAI Sora to generate a 10-second product-demo video from a text prompt. The stakeholders approve the overall scene but ask you to change only the color of the product's packaging in the final clip while keeping the camera motion, lighting, and all other objects visually identical across frames. You must adjust one generated segment without regenerating the entire video from scratch. Which approach best satisfies this requirement?

Accepted Answer

Use a video-editing (remix/edit) operation that takes the existing generated video plus an edit prompt describing the packaging color change, so the model modifies the targeted attribute while preserving temporal consistency of the rest of the scene

Answer

Submit a brand-new text-to-video generation request with a longer prompt that mentions the new packaging color, then manually pick the take that looks closest to the original

Answer

Extract individual frames, apply image inpainting to each frame separately, and reassemble them into a video file

Answer

Lower the temperature parameter and resubmit the original prompt so the output becomes deterministic and the color naturally shifts