Modèle de prompt vidéo IA : Prompt AI motion Blur image

Analyze the uploaded image and infer the real motion relationship between the main subject, the surrounding environment, and the camera. Do not apply blur based only on whether something is moving. Instead, determine motion blur from relative motion and photographic behavior. First identify the focal subject, the moving regions, the static regions, and the most likely camera behavior, such as a fixed camera, camera panning, subject tracking, or slight camera shake. Then apply realistic motion blur only where it would naturally appear in a real photograph. If the image shows a stationary subject while another object moves past, keep the stationary subject sharp and apply directional blur only to the moving object. If the image shows a moving subject that the camera is tracking, keep the main subject relatively sharp while applying stronger directional blur to the surrounding environment, road, lights, or background. If the image shows a moving subject captured by a fixed camera, allow stronger blur on the subject itself. If both subject and camera move partially together, preserve the subject core while allowing natural blur on edges, limbs, wheels, and nearby background regions. Scale blur intensity, trail length, and edge streaking according to the implied speed and relative motion. Faster relative motion should create stronger and longer blur; slower relative motion should create lighter and shorter blur. Respect scene depth, occlusion, focus, and spatial layering so the result feels physically plausible. Preserve identity, facial features, body proportions, object structure, composition, perspective, colors, lighting, materials, and overall scene content. The final image must look like a real photograph captured with authentic motion blur from real movement or camera exposure, not a generic digital blur filter. Avoid full-image blur, random smearing, ghosting, double edges, fake speed lines, melted structures, inconsistent blur directions, or loss of key subject recognizability.