Gemini Prompt: { "request_id": "vid_gen_070_face_lock_room_zoom_v1", "model_target": "VEO3.1", "function": "Video Generation", "video_duration": "10s", "global_technical_note_for_Veo3.1": "Priority 1: The 'face_reference_image' instruction (5d611cc6-cb0a-4ee9-8f9f-5f5122586fe8.jpeg) is ABSOLUTE and NON-NEGOTIABLE. The face *must not* be altered. Priority 2: The animation must be a slow continuous 'zoom out' (see ID 1). Priority 3: The 'authentic infant voice' must be used (ID 3). Priority 4: 'ExecuteAI' text placement on the wagon (ID 4).", "reference_assets": { "face_reference_image": "5d611cc6-cb0a-4ee9-8f9f-5f5122586fe8.jpeg", "technical_note_for_Veo3.1": "**ABSOLUTE CRITICAL REFERENCE INSTRUCTION: THE FACE FROM `face_reference_image` MUST BE APPLIED 100% UNCHANGED to the boy's face in `initial_visual_state` and throughout the entire animation. THERE MUST BE NO CHANGE, INTERPRETATION, OR ADAPTATION OF THE FACE. This lock applies even during the dialogue sequence (ID 3). The lipsync animation must adapt to the *locked* facial structure, not the other way around. This is a FIRST PRIORITY.**" }, "scene_description": { "summary": "A 10-second video starting as a medium shot on a boy (infant) in his wagon, set inside a modern room. The camera performs a slow, continuous zoom-out throughout the shot. The boy, with an absolutely unchanged face from the reference image, speaks complex dialogue in an authentic infant voice. 'ExecuteAI' text appears on his wagon.", "aesthetics": "Modern, domestic, mundane, slightly absurd, surreal, clean, well-lit.", "visual_style": "Cinematic, realistic, indoor lighting, crisp focus.", "lighting": "Bright, natural indoor lighting (e.g., from a large window off-camera).", "audio_cue": "Subtle, ambient room tone. **A genuine, authentic infant voice (e.g., 6-12 months old timbre). Must be extremely high-pitched, with the natural klang/timbre of an infant, but *must* (surrealistically) articulate the complex dialogue perfectly clearly.** It must NOT sound like an adult imitating a baby. No music.", "technical_note_for_Veo3.1": "**AUDIO/VISUAL SYNC PRIORITY: The lipsync for ID 3 must be generated *while respecting the 100% face-lock* from `reference_assets`. Do not warp the cheeks, nose, or eye structure. Animate *only* the mouth/jaw region minimally. The authenticity of the infant voice timbre is critical.**", "environment": "A modern, clean indoor room (e.g., a living room or playroom). Furniture like a sofa or a shelf might be visible as the camera zooms out. The baby and wagon are the central focus." }, "initial_visual_state": { "description": "Scene opens on a medium shot of the boy in his wagon, inside a room. His face is 100% identical to the source image. The shot is tight enough that the 'zoom out' will feel significant.", "character_in_frame": { "description": "A little boy (infant), 1-2 years old. **HIS FACE IS AN EXACT, UNCHANGED 1:1 REPRODUCTION OF `face_reference_image` (5d611cc6-cb0a-4ee9-8f9f-5f5122586fe8.jpeg). Absolutely no change to the facial features.**", "prop_details": "Sitting in a modern baby wagon/stroller." } }, "dialogue": { "language": "English", "line": "Do you not even have an Drum machine or a PlayStation or your own wagon?" }, "animation_sequence": [ { "id": 1, "time_range": "0s-4s", "description": "Slow zoom out begins, face locked.", "action": "The scene animates. A slow, steady, continuous zoom-out begins, revealing the baby is in a wagon in the middle of a room. Ambient room sounds are present. The infant blinks. His face is 100% locked to the source reference image." }, { "id": 2, "time_range": "4s-8s", "description": "Zoom out continues as infant speaks (face-lock, baby voice).", "action": "The slow zoom-out continues. The infant looks directly into the camera and animates with high-fidelity lipsync as he speaks: 'Do you not even have an Drum machine or a PlayStation or your own wagon?' **VOICE MUST BE an 'authentic infant voice' (high-pitched, 6-12 month timbre), not an adult imitation. CRITICAL ACTION: The lipsync *only* animates the mouth area. The rest of the facial structure (eyes, nose, cheeks) REMAINS 100% LOCKED to the `face_reference_image`.**" }, { "id": 3, "time_range": "8s-9.5s", "description": "Zoom out finishes, Text appears on wagon.", "action": "The zoom-out slows or finishes, revealing the full shot of the baby in the wagon within the room. As he finishes speaking, the text 'ExecuteAI' appears on the front of the baby wagon. The text is rendered as detailed stone engravings. **FORCE CORRECT SPELLING (E-x-e-c-u-t-e-A-I) AND ENGRAVED STYLE ON THE WAGON.**" }, { "id": 4, "time_range": "9.5s-10s", "description": "Fade to Black.", "action": "The final scene, showing the infant (with the locked face) in the room and the 'ExecuteAI' engravings on the wagon, holds briefly before fading to black." } ] }