Wan 3.0 Video Prime Cinematic Language Visual Notes: Three Camera Movement Demonstrations

Using wan3.0-video-prime's reference-driven capability as the creative path, create a "cinematic language visual note": use a reference image of the same subject, paired with text prompts that separately describe three common camera movements—dolly-in, lateral tracking, and orbit—and observe how the model changes camera position and movement direction while maintaining subject consistency. In terms of composition, base carries only non-media parameters—the prompt is submitted via the schema's prompt field, seed is fixed for comparison, duration is set to 10 seconds, resolution 720P, aspect ratio 16:9, and synchronized audio track enabled; the reference asset is generated via assets and then bound to reference_image_urls[0] to lock the subject's appearance. The prompt refers to the reference image as Image1 following the schema's reference mode. Use cases: demonstrate to developers the model's camera control effects under reference images + text prompts, as a reference sample for storyboard previsualization, shot style alignment, and prompt writing in batch video workflows. Note that the maximum duration per generation is 30 seconds; longer shot sequences need to be generated in segments.