Photo to Life

0.0–3.8 Its free hand lifts off the railing within half a second and sweeps loose strands behind its ear, leaving flyaways in the breeze. @avatar1 is already mid-sentence, relaxed American English, creator tone: "No one is telling you how to make AI characters look this real." Brow lift and head shake on "no one"; chin dip straight into the lens on "this real". Open-palm turn, a soft point on "you", hand resting near its collarbone. Eyes mostly on the lens with small drifts to the preview. Breeze in its hair, the banana leaves and the bougainvillea. Sun glitter on the sea, two or three gulls as distant specks. 3.8–4.5 "The skin" @avatar1 says it as the phone arm pushes 15–20cm closer to its cheek and eyes — a physical arm move, never a zoom — while its fingertips glide across its cheekbone and down its cheek in one pass. Focus snaps onto pores, sheen, peach fuzz and asymmetry. No smoothing. Its voice warms from mic proximity, one small audible breath. Do not reset the framing. 4.5–5.1 "hair" @avatar1 angles the phone screen right, putting its hair against the hazy water, head following slightly. Its free hand lifts a section near its face and lets it fall. A stronger gust separates individual strands against the sea. Focus locks on the strands; exposure shifts for the bright water then settles. 5.1–6.4 "the way they talk and move" The phone eases back to the opening framing. On "talk", relaxed rhythmic gestures near its chin including a small open-hand roll. On "move", a casual shoulder roll and side sway while its hand sweeps down and outward. Believable, never choreographed. The frame sways with its body; slight blur on the fast hand pass. Necklaces swing and catch light. 6.4–6.9 silence @avatar1 finishes, holds a small closed-mouth smile, blinks. A bird flies in from upper right, lands on the white pillar by the railing and turns its head toward the left edge of frame. Focus stays on @avatar1, so it sits slightly soft. 6.9–7.9 @avatar2 On the bird's head turn @avatar2 walks in from screen left — striped shoulder and forearm first, then its face — and settles just behind its right shoulder, leaning toward the phone like joining a selfie. Into the lens, playful half smile, laid-back slight Californian: "But I will." Chin dip and brow lift on "I". Right hand holds the Canon at its chest, strap over the shoulder, left arm relaxed. @avatar1 turns slightly toward @avatar2 with a knowing smile, no dialogue. Focus stays on @avatar1 so @avatar2 sits in depth; its voice reads farther from the mic. 7.9–9.0 flash @avatar2 raises the Canon to its right eye, left hand cradling the lens, aimed straight down the phone lens; the phone briefly refocuses toward it. At 8.4s the shutter clacks and the flash fires into the camera — highlights clip, the image blows to pure white within a few frames and holds pure white to the end. No ghost image, no vignette, no fade back. Breeze and gulls continue faintly under the white. AUDIO Phone mic only. Its voice close and clear, louder on "The skin"; @avatar2 farther back. Faint wind and leaves, distant gulls, a wing flutter as the bird lands, the shutter at 8.4s. No music.

3 steps1.2K144.6K

ASSET MAP: @image1 — PERSON_A, present-day, exactly as shown. Photocopy face, high braided bun, hair pins, skin, age, teal velvet embroidered jacket and earrings from the still. @image2 — the old photograph containing PERSON_B. PERSON_B’s identity, age and appearance come only from that photo. CHARACTER LOCK — HIGHEST PRIORITY: PERSON_A’s face, hairstyle, skin tone, build and apparent age must stay exactly as shown in @image1 throughout. PERSON_B, once animated, must stay exactly the person shown in @image2 — their exact facial features, apparent age (elderly, do not change it), skin tone, hair, beard and expression range, preserved from the photo. Do not beautify, de-age, age up, or restyle either person. Preserve each person’s gender presentation exactly as shown. No other people enter the scene. Whenever this text and either reference image disagree, follow the image. RESTRAINED PHOTO-WAKING RULE: PERSON_B begins completely still, exactly as printed in @image2 — aged photo texture, tone and grain visible. Their first movements are small and natural only: a soft blink, a faint warm smile, gentle breathing, the smallest head tilt. Do not give PERSON_B fast or large head turns or exaggerated motion before they have fully stepped out of the photo plane — abrupt motion on a photographic likeness distorts the face. Only after they have visibly left the frame do they move naturally and fully. 16:9. TOTAL DURATION: 12 SECONDS. OPENING: already outdoors with PERSON_A holding the framed print. Never a studio headshot as frame 0. 0-3s: Static establishing shot outdoors — a garden, porch or sunlit yard. PERSON_A stands calmly holding a framed print of the old photo showing PERSON_B — visibly aged, faded, almost entirely black-and-white or sepia-toned. Soft golden-hour light, light breeze moving PERSON_A’s hair and clothing gently. Camera holds still. 3-6s: Inside the frame, PERSON_B gently comes to life — a soft blink, a faint warm smile forming, gentle breathing, the smallest natural head tilt. Unlike a slow color-warming, here the old photo visibly brightens from black-and-white/sepia into full natural color in one gentle wave, without changing PERSON_B’s identity, age or appearance — only the color state changes. PERSON_A’s eyes widen with delighted surprise. 6-10s: PERSON_B steps out of the photo’s plane into PERSON_A’s space, takes PERSON_A’s hand, and gives them one gentle, playful spin. Both are laughing, hair and clothing swept by the motion. The camera orbits slowly around them at mid-distance as they turn, keeping both faces visible at least once during the spin. 10-12s: They come to a stop still holding hands, both breathless and smiling, facing each other. Camera settles into a steady wide shot showing them fully against the outdoor setting, rather than pushing into a close-up. CINEMATIC STYLE: Warm golden-hour outdoor light, natural handheld energy during the spin, settling to a steady wide frame at the end, shallow-to-moderate depth of field, live-action realism — preserve real skin texture and imperfections on PERSON_A; PERSON_B’s texture may echo the photo’s grain slightly even once alive, as a subtle visual signature that they came from the photograph. No subtitles. No on-screen text or watermark. Hands have five fingers each, anatomically correct.

3 steps

ASSET MAP: @image1 — PERSON_A, the child, present-day, exactly as shown. Photocopy face, silver-blonde bob, skin, apparent age, build, and the pink hoodie tracksuit from the still. @image1 is identity and clothing only — not the outdoor courtyard. @image2 — the old photograph containing PERSON_B. PERSON_B’s identity, age and appearance come only from that photo. CHARACTER LOCK — HIGHEST PRIORITY: PERSON_A’s face, hairstyle, skin tone, build and apparent age must stay exactly as shown in @image1 throughout — do not age them up or down, do not alter their proportions. PERSON_B, once animated, must stay exactly the person shown in @image2 — their exact facial features, apparent age exactly as the photo shows, skin tone, hair and expression range, preserved from the photo. Do not beautify, de-age, age up, or restyle either person. Preserve each person’s gender presentation exactly as shown. No other people enter the scene. Whenever this text and either reference image disagree, follow the image. CONTENT SAFETY: PERSON_A remains fully clothed throughout in ordinary, everyday clothing exactly as shown in @image1. All physical contact between PERSON_B and PERSON_A is limited to a gentle hug and a hand resting on the shoulder or head — warm, familial, G-rated, nothing else. The scene is entirely wholesome, supervised-feeling and safe in tone throughout. RESTRAINED PHOTO-WAKING RULE: PERSON_B begins completely still, exactly as printed in @image2 — aged photo texture, tone and grain visible. Their first movements are small and natural only: a soft blink, a faint warm smile, gentle breathing, the smallest head tilt. Do not give PERSON_B fast or large head turns or exaggerated motion before they have fully stepped out of the photo plane — abrupt motion on a photographic likeness distorts the face. Only after they have visibly left the frame do they move naturally and fully. 16:9. TOTAL DURATION: 12 SECONDS. SET: a cozy living room. Soft warm indoor light. Not the outdoor courtyard from @image1. OPENING: already in the living room with PERSON_A and the framed print. Never a studio headshot as frame 0. Never the attached still as frame 0. 0-3s: Static establishing shot in a cozy living room. PERSON_A (the child) sits on the floor or a rug, near a small side table holding a framed print of the old photo showing PERSON_B — visibly aged, faded tone, worn paper texture. Soft warm indoor light. Camera holds completely still. 3-6s: Inside the frame, PERSON_B gently comes to life — a soft blink, a faint warm smile forming, gentle breathing, the smallest natural head tilt, the printed texture gradually gaining living depth without changing PERSON_B’s identity, age or appearance. PERSON_A looks up from what they were doing, eyes widening with curious wonder. 6-10s: PERSON_B steps out of the photo’s plane, kneels down to PERSON_A’s eye level so they are face to face, and says warmly (no audio needed, can be silent or a soft audible “Hello, little one”). PERSON_B opens their arms; PERSON_A leans in for a gentle hug, and PERSON_B rests a hand softly on PERSON_A’s head or shoulder, beaming with quiet joy. PERSON_A giggles, a little shy but happy. 10-12s: They stay close, PERSON_B smiling down at PERSON_A with warmth. Camera holds the same steady eye-level framing — no push, no orbit — letting the moment sit quietly. CINEMATIC STYLE: Soft warm indoor light, minimal camera movement, shallow depth of field, live-action realism — preserve real skin texture and imperfections on PERSON_A; PERSON_B’s texture may echo the photo’s grain slightly even once alive, as a subtle visual signature that they came from the photograph. No subtitles. No on-screen text or watermark. Hands have five fingers each, anatomically correct. MUSIC: original warm instrumental score throughout all 12 seconds — gentle piano and soft strings, emotional but restrained, no lyrics, no vocals, no copyrighted songs. Music is clearly audible from frame 1, under the giggle. Not silent. Not ambient room tone only.

3 steps

ASSET MAP: @image1 — PERSON_A, present-day, exactly as shown. Photocopy face, hair, skin, age, teal athletic top and gold necklaces from the still. @image2 — the old photograph containing PERSON_B. PERSON_B’s identity, age and appearance come only from that photo. CHARACTER LOCK — HIGHEST PRIORITY: PERSON_A’s face, hairstyle, skin tone, build and apparent age must stay exactly as shown in @image1 throughout. PERSON_B, once animated, must stay exactly the person shown in @image2 — their exact facial features, apparent age (whatever age the photo shows — elderly, do not change it), skin tone, hair and expression range, preserved from the photo. Do not beautify, de-age, age up, or restyle either person. Preserve each person’s gender presentation exactly as shown. No other people enter the scene. Whenever this text and either reference image disagree, follow the image. RESTRAINED PHOTO-WAKING RULE: PERSON_B begins completely still, exactly as printed in @image2 — aged photo texture, tone and grain visible. Their first movements are small and natural only: a soft blink, a faint warm smile, gentle breathing, the smallest head tilt. Do not give PERSON_B fast or large head turns or exaggerated motion before they have fully stepped out of the photo plane — abrupt motion on a photographic likeness distorts the face. Only after they have visibly left the frame do they move naturally and fully. 16:9. TOTAL DURATION: 12 SECONDS. OPENING: already in the room with PERSON_A and the framed print. Never a studio headshot as frame 0. 0-3s: Static establishing shot. PERSON_A stands or sits calmly, holding or seated beside a framed print of the old photo showing PERSON_B — visibly aged, grain, faded tone, slightly worn paper texture. Quiet, still, a little anticipatory. Camera holds completely still. Natural indoor or soft outdoor lighting that makes both PERSON_A and the framed photo look like they belong in the same room. 3-6s: Inside the frame, PERSON_B gently comes to life — a soft blink, a faint warm smile forming, gentle breathing, the smallest natural head tilt. The old photo’s color and depth slowly warm and deepen, the flat printed texture gradually giving way to living dimension, without changing PERSON_B’s identity, age or appearance. PERSON_A notices, their expression shifting to quiet wonder. 6-10s: PERSON_B steps out of the photo’s plane into PERSON_A’s space. They reach out and gently cup PERSON_A’s cheek in one hand, smiling warmly and studying their face as if seeing them for the first time in years. PERSON_A closes their eyes for a moment, leaning slightly into the touch, a small emotional breath escaping. No embrace, no words — just the gesture held for a beat. 10-12s: PERSON_A opens their eyes and meets PERSON_B’s gaze, both smiling softly. Camera slowly pushes in a few inches, or holds steady. CINEMATIC STYLE: Soft natural or warm indoor light, gentle handheld micro-movement, shallow depth of field, live-action realism — preserve real skin texture and imperfections on PERSON_A; PERSON_B’s texture may echo the photo’s grain slightly even once alive, as a subtle visual signature that they came from the photograph. No subtitles. No on-screen text or watermark. Hands have five fingers each, anatomically correct.

3 steps

ASSET MAP: @image1 — PERSON_A, present-day, exactly as shown. Photocopy face, black bob, red lipstick, gold hoop earrings, pearl and gold chain necklace, black sleeveless cutout dress, skin, age and build from the still. @image2 — the old photograph containing PERSON_B. PERSON_B’s identity, age and appearance come only from that photo. CHARACTER LOCK — HIGHEST PRIORITY: PERSON_A’s face, hairstyle, skin tone, build and apparent age must stay exactly as shown in @image1 throughout. PERSON_B, once animated, must stay exactly the person shown in @image2 — their exact facial features, apparent age (elderly, do not change it), skin tone, silver-grey wavy hair and expression range, preserved from the photo. Do not beautify, de-age, age up, or restyle either person. Preserve each person’s gender presentation exactly as shown. No other people enter the scene. Whenever this text and either reference image disagree, follow the image. RESTRAINED PHOTO-WAKING RULE: PERSON_B begins completely still, exactly as printed in @image2 — aged photo texture, tone and grain visible. Their first movements are small and natural only: a soft blink, a faint warm smile, gentle breathing, the smallest head tilt. Do not give PERSON_B fast or large head turns or exaggerated motion before they have fully stepped out of the photo plane — abrupt motion on a photographic likeness distorts the face. Only after they have visibly left the frame do they move naturally and fully. 16:9. TOTAL DURATION: 12 SECONDS. SET: a quiet residential interior near a doorway or hallway mirror. Soft morning daylight streaming in. Not a restaurant. Not a night city skyline. Not dusk windows. OPENING: already indoors with PERSON_A holding the framed print. Never a studio headshot as frame 0. 0-3s: Static establishing shot indoors, near a doorway or hallway mirror, soft morning light streaming in. PERSON_A stands still, a little pensive, holding a framed print of the old photo showing PERSON_B — visibly aged, faded tone, worn paper texture. Camera holds completely still in a steady medium shot. 3-6s: Inside the frame, PERSON_B gently comes to life — a soft blink, a faint warm smile forming, gentle breathing, the smallest natural head tilt, the printed texture gradually gaining living depth without changing PERSON_B’s identity, age or appearance. PERSON_A looks up, surprised and moved. 6-10s: PERSON_B steps out of the photo’s plane into PERSON_A’s space, reaches over and gently straightens PERSON_A’s collar or smooths a stray piece of their hair — the way a parent or mentor would before someone heads out the door. Then they step back half a pace with a proud, approving smile, giving a small nod. PERSON_A smiles back, a little bashful, eyes glistening. 10-12s: They remain standing together in the doorway light, PERSON_B’s hand resting briefly on PERSON_A’s shoulder. Camera stays in the same steady medium shot — no push, no orbit — holding the two-shot as the moment settles. CINEMATIC STYLE: Soft morning indoor light, minimal camera movement throughout (deliberately steadier than a typical reunion clip — this is a quiet, composed moment, not a rush), shallow depth of field, live-action realism — preserve real skin texture and imperfections on PERSON_A; PERSON_B’s texture may echo the photo’s grain slightly even once alive, as a subtle visual signature that they came from the photograph. No subtitles. No on-screen text or watermark. Hands have five fingers each, anatomically correct.

3 steps

CHARACTER LOCK — HIGHEST PRIORITY: PERSON_A is exactly the person in @image1. Preserve their facial identity, facial structure, eye shape and colour, eyebrows, nose, lips, skin tone, hairline, hairstyle, hair length, texture and colour, freckles and distinctive marks, apparent age, build and body proportions. Preserve their gender presentation exactly as shown. PERSON_B, once animated, is exactly the person shown in @image2 — their exact facial features, apparent age as the photo shows, skin tone, hair and expression range. Do not beautify, de-age, age up, or restyle either person. No other people enter the scene. Whenever this text and either reference image disagree, follow the image. @image1 is PERSON_A, present-day — photocopy face, hair and the navy square-neck look from the still. @image2 is the old photograph containing PERSON_B. PERSON_B’s identity comes only from that photo. RESTRAINED PHOTO-WAKING RULE: PERSON_B begins completely still, exactly as printed in @image2 — aged photo texture, tone and grain visible. Their first movements are small and natural only: a soft blink, a faint warm smile, gentle breathing, the smallest head tilt. Do not give PERSON_B fast or large head turns or exaggerated motion before they have fully stepped out of the photo plane — abrupt motion on a photographic likeness distorts the face. Only after they have visibly left the frame do they move naturally and fully. 16:9. TOTAL DURATION: 12 SECONDS. OPENING: already in the room with PERSON_A and the framed print. Never a studio headshot as frame 0. 0-3s: Static establishing shot. PERSON_A stands or sits calmly, holding or seated beside a framed print of the old photo showing PERSON_B — visibly aged, grain, faded tone, slightly worn paper texture. Quiet, still, a little anticipatory. Camera holds completely still. Natural indoor or soft outdoor lighting that makes both PERSON_A and the framed photo look like they belong in the same room. 3-6s: Inside the frame, PERSON_B gently comes to life — a soft blink, a faint warm smile forming, gentle breathing, the smallest natural head tilt. The old photo’s color and depth slowly warm and deepen, the flat printed texture gradually giving way to living dimension, without changing PERSON_B’s identity, age or appearance. PERSON_A notices, their expression shifting to quiet wonder. 6-10s: PERSON_B steps out of the photo’s plane into PERSON_A’s space. They close the distance and share a warm, gentle hug, arms wrapping around each other’s backs. Both have joyful, emotional expressions, eyes slightly teary. Soft happy laughter, no dialogue needed. 10-12s: Both settle into the embrace. Warm, natural expressions on both faces. Camera slowly pushes in a few inches, or holds steady. CINEMATIC STYLE: Soft natural or warm indoor light, gentle handheld micro-movement, shallow depth of field, live-action realism — preserve real skin texture and imperfections on PERSON_A; PERSON_B’s texture may echo the photo’s grain slightly even once alive, as a subtle visual signature that they came from the photograph. No subtitles. No on-screen text or watermark. Hands have five fingers each, anatomically correct.

3 steps