{ "schema_version": 1, "concept_key": "eye", "video_count": 3, "scripts": [ { "video_key": "eye_context_01", "purpose": "model", "duration_seconds": 8, "format": "16:9 landscape", "frame_rate": "25 or 30 fps constant", "resolution": "1920x1080 minimum", "reference_setup_image": "eye_context_01_setup_01.png", "reference_concept_image": "../images/eye_isolated_01.png", "story_summary": "A 5-8 second slow reveal and natural interaction with eye, keeping the target in frame.", "setting": "plain warm-white studio setting", "people": "one AI-generated child body part", "target_depiction": "South Asian child's right eye looking forward", "timeline": [ { "timecode": "00:00.000-00:01.500", "picture": "Begin with the target partly occluded or resting naturally while its context is visible.", "performance": "Settle naturally; no one looks at camera unless the communicative meaning requires partner gaze.", "camera": "Locked establishing view; level horizon; no handheld shake.", "audio": "Capture clean room tone only.", "edit": "Start from a clean stable frame; no title card." }, { "timecode": "00:01.500-00:05.500", "picture": "Reveal or interact with the target slowly and naturally: South Asian child's right eye looking forward. Keep the whole target readable.", "performance": "Reveal and interact with the target once: South Asian child's right eye looking forward. Keep its identity, scale and appearance constant.", "camera": "Medium-close teaching view; keep the target, relevant hands, reference object and face in focus.", "audio": "Natural production sound at low level; no spoken label is required.", "edit": "One continuous take preferred; one motivated straight cut permitted." }, { "timecode": "00:05.500-00:08.000", "picture": "Hold the target fully visible for recognition and optional pause-frame use.", "performance": "Hold the completed meaning without pointing, celebration or repeated action.", "camera": "Locked result frame with at least 5 percent safe margin around the target.", "audio": "Continue matching room tone; fade neither picture nor sound before 8 seconds.", "edit": "End on a stable frame suitable for pausing." } ], "camera_direction": { "lens_equivalent": "35-50 mm natural perspective; 70-85 mm for isolated face detail", "height": "target or seated eye level", "movement": "locked-off; optional imperceptibly slow push-in", "focus": "sufficient depth of field to keep semantic evidence sharp", "exposure": "natural skin tones and no clipped highlights", "white_balance": "locked throughout", "continuity_locks": [ "participant identity", "clothing", "target instance", "background layout", "lighting direction" ] }, "performance_direction": { "primary": "Reveal and interact with the target once: South Asian child's right eye looking forward. Keep its identity, scale and appearance constant.", "pace": "natural and readable, never slow-motion acting", "gaze": "natural task or partner gaze; never require eye contact", "repetitions": "one complete target event", "safeguarding": "stop on discomfort; no forced compliance or distress performance" }, "spoken_script": { "required": false, "dialogue": [], "localized_voiceover_optional": "Use the approved localized semantic label or instruction for eye only after SLP and native review.", "record_separately": true }, "caption_script": { "language": "en", "picture_only_policy": "Do not caption visual action as sound. Leave captions off when the final mix contains no speech or meaningful sound.", "cues": [], "localized_caption_requirement": "Create from the locked final audio, not by translating this draft blindly." }, "audio_description": { "language": "en", "script": "In plain warm-white studio setting, South Asian child's right eye looking forward.", "delivery": "neutral concise present tense; place in a natural pause or provide as separate accessible track", "review": "SLP and blind/low-vision accessibility review required" }, "accessibility_notes": [ "Meaning must be understandable with audio muted.", "Do not use speech, music, colour, camera movement or a transient cue as the only signal.", "Keep target visible for at least two seconds after the event.", "No flash, flicker, rapid cut, whip pan or unexpected loud sound.", "Provide player pause, replay, captions toggle and audio-description track where supported.", "Check target contrast and visibility at 320-pixel-wide playback." ], "generation_prompt": "Create one 8-second landscape 16:9 realistic clinical-educational video for the language-independent concept 'Eye' (key eye), purpose model. Use the approved reference image eye_isolated_01.png for semantic content and visual continuity, not for identity preservation. Scene: South Asian child's right eye looking forward in plain warm-white studio setting; participants: one AI-generated child body part. Begin with a stable establishing view, show one slow natural target action or reveal, then hold the completed meaning for at least two seconds. Keep the target continuously visible, anatomically and physically plausible, safely framed and naturally lit. Use at most one simple camera cut; no fast zoom, montage or dramatic acting. Language-neutral: no necessary speech and no embedded writing. Preserve culturally respectful, disability-inclusive representation.", "negative_prompt": "no captions, labels, readable text, letters, numbers, logos, brands or watermark; no jump cuts, time lapse, fast camera movement, flicker, strobe, looping discontinuity or motion smear; no extra fingers, fused limbs, changing identity, changing clothing, morphing objects or disappearing props; no unsafe, frightening, coercive, humiliating or medically distressing behavior; no target hidden by hands, crop, foreground objects or shallow focus at the teaching moment; no music as a meaning cue and no speech required to understand the concept", "generation_settings": { "aspect_ratio": "16:9", "duration_seconds": 8, "camera_motion": "locked", "seed": "record exact provider seed when available", "reference_strength": "preserve semantics and composition; do not clone a real identity", "generate_audio": false }, "continuity_review": [ "same participant identity and anatomy", "same clothing and assistive devices", "same target object and quantity", "hands remain anatomically plausible", "background objects do not appear or disappear", "action begins and ends logically" ], "delivery": { "video_filename": "eye_context_01.mp4", "caption_filename": "eye_context_01.captions..vtt", "audio_description_filename": "eye_context_01.audio_description..wav", "poster_filename": "eye_context_01.poster.png" }, "status": "script_ready_review_required" }, { "video_key": "eye_context_02", "purpose": "functional_context", "duration_seconds": 8, "format": "16:9 landscape", "frame_rate": "25 or 30 fps constant", "resolution": "1920x1080 minimum", "reference_setup_image": "eye_context_02_setup_01.png", "reference_concept_image": "../images/eye_natural_context_01.png", "story_summary": "A 6-10 second functional everyday use of eye by a person, without unnecessary spoken language.", "setting": "simple softly blurred garden", "people": "one AI-generated child", "target_depiction": "Black child uses eyes to watch an orange butterfly", "timeline": [ { "timecode": "00:00.000-00:01.500", "picture": "Begin with the target partly occluded or resting naturally while its context is visible.", "performance": "Settle naturally; no one looks at camera unless the communicative meaning requires partner gaze.", "camera": "Locked establishing view; level horizon; no handheld shake.", "audio": "Capture clean room tone only.", "edit": "Start from a clean stable frame; no title card." }, { "timecode": "00:01.500-00:05.500", "picture": "Reveal or interact with the target slowly and naturally: Black child uses eyes to watch an orange butterfly. Keep the whole target readable.", "performance": "Reveal and interact with the target once: Black child uses eyes to watch an orange butterfly. Keep its identity, scale and appearance constant.", "camera": "Medium-close teaching view; keep the target, relevant hands, reference object and face in focus.", "audio": "Natural production sound at low level; no spoken label is required.", "edit": "One continuous take preferred; one motivated straight cut permitted." }, { "timecode": "00:05.500-00:08.000", "picture": "Hold the target fully visible for recognition and optional pause-frame use.", "performance": "Hold the completed meaning without pointing, celebration or repeated action.", "camera": "Locked result frame with at least 5 percent safe margin around the target.", "audio": "Continue matching room tone; fade neither picture nor sound before 8 seconds.", "edit": "End on a stable frame suitable for pausing." } ], "camera_direction": { "lens_equivalent": "35-50 mm natural perspective; 70-85 mm for isolated face detail", "height": "target or seated eye level", "movement": "locked-off; optional imperceptibly slow push-in", "focus": "sufficient depth of field to keep semantic evidence sharp", "exposure": "natural skin tones and no clipped highlights", "white_balance": "locked throughout", "continuity_locks": [ "participant identity", "clothing", "target instance", "background layout", "lighting direction" ] }, "performance_direction": { "primary": "Reveal and interact with the target once: Black child uses eyes to watch an orange butterfly. Keep its identity, scale and appearance constant.", "pace": "natural and readable, never slow-motion acting", "gaze": "natural task or partner gaze; never require eye contact", "repetitions": "one complete target event", "safeguarding": "stop on discomfort; no forced compliance or distress performance" }, "spoken_script": { "required": false, "dialogue": [], "localized_voiceover_optional": "Use the approved localized semantic label or instruction for eye only after SLP and native review.", "record_separately": true }, "caption_script": { "language": "en", "picture_only_policy": "Do not caption visual action as sound. Leave captions off when the final mix contains no speech or meaningful sound.", "cues": [], "localized_caption_requirement": "Create from the locked final audio, not by translating this draft blindly." }, "audio_description": { "language": "en", "script": "In simple softly blurred garden, Black child uses eyes to watch an orange butterfly.", "delivery": "neutral concise present tense; place in a natural pause or provide as separate accessible track", "review": "SLP and blind/low-vision accessibility review required" }, "accessibility_notes": [ "Meaning must be understandable with audio muted.", "Do not use speech, music, colour, camera movement or a transient cue as the only signal.", "Keep target visible for at least two seconds after the event.", "No flash, flicker, rapid cut, whip pan or unexpected loud sound.", "Provide player pause, replay, captions toggle and audio-description track where supported.", "Check target contrast and visibility at 320-pixel-wide playback." ], "generation_prompt": "Create one 8-second landscape 16:9 realistic clinical-educational video for the language-independent concept 'Eye' (key eye), purpose functional_context. Use the approved reference image eye_natural_context_01.png for semantic content and visual continuity, not for identity preservation. Scene: Black child uses eyes to watch an orange butterfly in simple softly blurred garden; participants: one AI-generated child. Begin with a stable establishing view, show one slow natural target action or reveal, then hold the completed meaning for at least two seconds. Keep the target continuously visible, anatomically and physically plausible, safely framed and naturally lit. Use at most one simple camera cut; no fast zoom, montage or dramatic acting. Language-neutral: no necessary speech and no embedded writing. Preserve culturally respectful, disability-inclusive representation.", "negative_prompt": "no captions, labels, readable text, letters, numbers, logos, brands or watermark; no jump cuts, time lapse, fast camera movement, flicker, strobe, looping discontinuity or motion smear; no extra fingers, fused limbs, changing identity, changing clothing, morphing objects or disappearing props; no unsafe, frightening, coercive, humiliating or medically distressing behavior; no target hidden by hands, crop, foreground objects or shallow focus at the teaching moment; no music as a meaning cue and no speech required to understand the concept", "generation_settings": { "aspect_ratio": "16:9", "duration_seconds": 8, "camera_motion": "locked", "seed": "record exact provider seed when available", "reference_strength": "preserve semantics and composition; do not clone a real identity", "generate_audio": false }, "continuity_review": [ "same participant identity and anatomy", "same clothing and assistive devices", "same target object and quantity", "hands remain anatomically plausible", "background objects do not appear or disappear", "action begins and ends logically" ], "delivery": { "video_filename": "eye_context_02.mp4", "caption_filename": "eye_context_02.captions..vtt", "audio_description_filename": "eye_context_02.audio_description..wav", "poster_filename": "eye_context_02.poster.png" }, "status": "script_ready_review_required" }, { "video_key": "eye_context_03", "purpose": "generalisation", "duration_seconds": 8, "format": "16:9 landscape", "frame_rate": "25 or 30 fps constant", "resolution": "1920x1080 minimum", "reference_setup_image": "eye_context_03_setup_01.png", "reference_concept_image": "../images/eye_unfamiliar_01.png", "story_summary": "A 6-10 second novel setting or instance of eye for generalisation.", "setting": "plain soft-beige studio background", "people": "one AI-generated older adult body part", "target_depiction": "older Black woman's right eye looking toward camera", "timeline": [ { "timecode": "00:00.000-00:01.500", "picture": "Begin with the target partly occluded or resting naturally while its context is visible.", "performance": "Settle naturally; no one looks at camera unless the communicative meaning requires partner gaze.", "camera": "Locked establishing view; level horizon; no handheld shake.", "audio": "Capture clean room tone only.", "edit": "Start from a clean stable frame; no title card." }, { "timecode": "00:01.500-00:05.500", "picture": "Reveal or interact with the target slowly and naturally: older Black woman's right eye looking toward camera. Keep the whole target readable.", "performance": "Reveal and interact with the target once: older Black woman's right eye looking toward camera. Keep its identity, scale and appearance constant.", "camera": "Medium-close teaching view; keep the target, relevant hands, reference object and face in focus.", "audio": "Natural production sound at low level; no spoken label is required.", "edit": "One continuous take preferred; one motivated straight cut permitted." }, { "timecode": "00:05.500-00:08.000", "picture": "Hold the target fully visible for recognition and optional pause-frame use.", "performance": "Hold the completed meaning without pointing, celebration or repeated action.", "camera": "Locked result frame with at least 5 percent safe margin around the target.", "audio": "Continue matching room tone; fade neither picture nor sound before 8 seconds.", "edit": "End on a stable frame suitable for pausing." } ], "camera_direction": { "lens_equivalent": "35-50 mm natural perspective; 70-85 mm for isolated face detail", "height": "target or seated eye level", "movement": "locked-off; optional imperceptibly slow push-in", "focus": "sufficient depth of field to keep semantic evidence sharp", "exposure": "natural skin tones and no clipped highlights", "white_balance": "locked throughout", "continuity_locks": [ "participant identity", "clothing", "target instance", "background layout", "lighting direction" ] }, "performance_direction": { "primary": "Reveal and interact with the target once: older Black woman's right eye looking toward camera. Keep its identity, scale and appearance constant.", "pace": "natural and readable, never slow-motion acting", "gaze": "natural task or partner gaze; never require eye contact", "repetitions": "one complete target event", "safeguarding": "stop on discomfort; no forced compliance or distress performance" }, "spoken_script": { "required": false, "dialogue": [], "localized_voiceover_optional": "Use the approved localized semantic label or instruction for eye only after SLP and native review.", "record_separately": true }, "caption_script": { "language": "en", "picture_only_policy": "Do not caption visual action as sound. Leave captions off when the final mix contains no speech or meaningful sound.", "cues": [], "localized_caption_requirement": "Create from the locked final audio, not by translating this draft blindly." }, "audio_description": { "language": "en", "script": "In plain soft-beige studio background, older Black woman's right eye looking toward camera.", "delivery": "neutral concise present tense; place in a natural pause or provide as separate accessible track", "review": "SLP and blind/low-vision accessibility review required" }, "accessibility_notes": [ "Meaning must be understandable with audio muted.", "Do not use speech, music, colour, camera movement or a transient cue as the only signal.", "Keep target visible for at least two seconds after the event.", "No flash, flicker, rapid cut, whip pan or unexpected loud sound.", "Provide player pause, replay, captions toggle and audio-description track where supported.", "Check target contrast and visibility at 320-pixel-wide playback." ], "generation_prompt": "Create one 8-second landscape 16:9 realistic clinical-educational video for the language-independent concept 'Eye' (key eye), purpose generalisation. Use the approved reference image eye_unfamiliar_01.png for semantic content and visual continuity, not for identity preservation. Scene: older Black woman's right eye looking toward camera in plain soft-beige studio background; participants: one AI-generated older adult body part. Begin with a stable establishing view, show one slow natural target action or reveal, then hold the completed meaning for at least two seconds. Keep the target continuously visible, anatomically and physically plausible, safely framed and naturally lit. Use at most one simple camera cut; no fast zoom, montage or dramatic acting. Language-neutral: no necessary speech and no embedded writing. Preserve culturally respectful, disability-inclusive representation.", "negative_prompt": "no captions, labels, readable text, letters, numbers, logos, brands or watermark; no jump cuts, time lapse, fast camera movement, flicker, strobe, looping discontinuity or motion smear; no extra fingers, fused limbs, changing identity, changing clothing, morphing objects or disappearing props; no unsafe, frightening, coercive, humiliating or medically distressing behavior; no target hidden by hands, crop, foreground objects or shallow focus at the teaching moment; no music as a meaning cue and no speech required to understand the concept", "generation_settings": { "aspect_ratio": "16:9", "duration_seconds": 8, "camera_motion": "locked", "seed": "record exact provider seed when available", "reference_strength": "preserve semantics and composition; do not clone a real identity", "generate_audio": false }, "continuity_review": [ "same participant identity and anatomy", "same clothing and assistive devices", "same target object and quantity", "hands remain anatomically plausible", "background objects do not appear or disappear", "action begins and ends logically" ], "delivery": { "video_filename": "eye_context_03.mp4", "caption_filename": "eye_context_03.captions..vtt", "audio_description_filename": "eye_context_03.audio_description..wav", "poster_filename": "eye_context_03.poster.png" }, "status": "script_ready_review_required" } ], "approval": { "clinical": false, "cultural": false, "accessibility": false, "rights": false, "technical": false } }