Japanese Live-Action Reference-to-Video Prompt Workflows VIDEO Breakdown (Prompts Included)
AI Generated Video
Recreating these Japanese live-action reference-to-video prompt workflows is straightforward with Pollo AI. Start by generating each reference sheet to lock the characters, wardrobe, props, and location; then upload it as the global reference for the matching video prompt. The prompts below are structured for CMS readers who want reusable image-to-video workflows, stable character continuity, and clean Japanese dialogue scenes.
Step 1
Open pollo.ai, select Image Generation, paste the image prompt below and click Generate. Set GPT Image 2.5 as the default image model. Adjust the aspect ratio according to your video needs; multiple sizes including 16:9 and 9:16 are supported.
Image Generation Screen
VIDEO1
Image Prompt
CREATE ONE SINGLE ULTRA-DETAILED HORIZONTAL 16: 9 professional cinematic production-design board for a 15-second photorealistic Japanese live-action coming-of-age drama. Matte-black background, clean gray borders, yellow numbered section labels and professional Japanese typography. Exact layout: production header and palette across the top, character references at upper-left, environment and set design at upper-right, exactly six chronological storyboard frames across the lower-middle, and technical/lighting notes along the bottom. All photographic content must be sensitive, natural and completely live action.
All visible writing must be Japanese only. Title:「届かなかった声」. Header:「尺：15秒」「カット数：6」「画面比：16:9」「季節：晩夏」「時間帯：夕方」「色彩設計」. Palette:「川面青」「若葉緑」「制服紺」「生成色」「夕焼桃」.
SECTION 1: 「1 キャラクター設定」.「人物A：女子学生・18歳」has a gentle round-oval face, realistic fair skin, soft brown eyes and a straight chestnut-brown bob. Two small anatomically correct behind-the-ear hearing aids are visible. She wears a cream school cardigan, white blouse, navy ribbon, navy skirt, dark socks and brown loafers, carrying one light school bag and one worn communication notebook. Show a large portrait, front/side/back/three-quarter turnarounds and expressions labeled「通常」「注意」「不安」「驚き」「安心」「微笑み」.「人物B：男子学生・18歳」has a slim build, short slightly messy black hair, natural tired eyes, white school shirt, loose navy tie, dark trousers, sneakers and black backpack. Show matching views and expressions labeled「通常」「緊張」「後悔」「ためらい」「謝罪」「安堵」.
SECTION 2: 「2 環境・セットデザイン」. Large hero photograph of a quiet pedestrian bridge above a broad Japanese river in late-afternoon sunlight. Blue-green water, visible koi, grassy embankments, bicycle path, distant apartments, summer trees, metal railings and soft peach-colored light. Include smaller detail references of the communication notebook, realistic hearing aid, koi beneath rippling water and sunlight on the railing. Add a blocking diagram labeled「少女の位置」「少年の進行方向」「橋の手すり」「川」「安全な距離」「撮影位置」.
SECTION 3: 「3 絵コンテシークエンス」. Exactly six horizontal frames. Frame 1: peaceful macro shot of koi swimming beneath rippling river water, caption「川面の鯉と波紋」. Frame 2: medium-wide rear shot of the girl standing alone at the bridge railing, caption「少女が橋で魚を見つめる」. Frame 3: medium telephoto shot of the nervous boy approaching slowly and stopping several steps behind her, caption「少年がゆっくり近づく」. Frame 4: intimate facial close-up as she turns and recognizes him, with one realistic hearing aid visible, caption「少女が振り返る」. Frame 5: eye-level two-shot as the boy returns the communication notebook with both hands and bows, caption「少年がノートを返して頭を下げる」. Frame 6: wide final shot of both walking side by side across the bridge with respectful space between them, caption「二人が並んで橋を歩く」.
SECTION 4: 「4 撮影・照明設定」. Labels:「カメラ：35ミリ映画用カメラ」「レンズ：35ミリ・50ミリ・85ミリ」「照明：夕方の自然光」「画調：川面の青緑と柔らかな夕焼け色」「雰囲気：後悔・尊重・癒やし・再会」. No existing movie title, creator name, anime, manga, floating symbols, bullying, violence, injury, exaggerated crying, kissing, inaccurate hearing aids, disrespectful disability depiction, malformed fingers, plastic skin, extra main characters, wardrobe changes, English text, watermark or random letters.
Step 2
Once your image is ready, go to Video Generation. Upload your prepared image or pick assets from your Creations library. Select reference-to-video mode and paste your video prompt. You can also upload a video for motion reference.
Note: Add @your-image for image reference and @your-video for video reference inside the prompt. Pick any of these models: Seedance 2.5, Seedance 2.0, MiniMax H3 or Wan 3.0. Click Generate after confirming all settings.
Video Generation Screen
Video Prompt
CREATE A 15-SECOND HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action video. Use [@BridgeReferenceSheet] as the GLOBAL REFERENCE for the exact identities, faces, hairstyles, body proportions, wardrobe, bags, hearing aids and communication notebook of the girl and boy. Whenever either character appears, preserve the corresponding person from [@BridgeReferenceSheet] exactly. Use [@BridgeReferenceSheet] as the reference for the same bridge, river, railings, koi, embankment, surrounding buildings, trees, sunset light, color palette and quiet emotional tone. Use the storyboard images inside [@BridgeReferenceSheet] as guidance for the action and composition, but turn them into a connected live-action sequence instead of directly animating separate panels. Never show the entire reference sheet, black background, character turnarounds, expression sheets, diagrams, captions, grids or palette blocks from [@BridgeReferenceSheet] .
Sensitive premium Japanese live-action youth drama with real actors, natural skin pores, realistic eyes, anatomically correct hands, accurate hearing-aid placement and respectful disability representation. Restrained acting, natural late-afternoon light and subtle camera movement. Six clean cinematic cuts with no fades, morphs, transparency overlays, floating graphic symbols or slideshow presentation.
0.0–1.8s: Begin with the river environment from [@BridgeReferenceSheet] . Macro moving shot above the water as several koi swim beneath realistic ripples. Small pieces of fish food fall individually onto the surface and create expanding circular waves. Sunlight moves across the water with the fish instead of remaining frozen.
1.8–4.0s: Rear medium-wide shot using the girl and bridge composition from [@BridgeReferenceSheet] . The girl from [@BridgeReferenceSheet] stands safely beside the railing, watching the koi below. The camera approaches slowly from behind. Afternoon wind moves her hair and clothing naturally. Preserve the exact hearing aids from [@BridgeReferenceSheet] without enlarging them or making them glow.
4.0–6.2s: Long-lens medium shot of the boy from [@BridgeReferenceSheet] approaching with the same communication notebook shown in [@BridgeReferenceSheet] held carefully in both hands. He slows several respectful steps behind her, gathers his courage and softly calls in Japanese:「西宮。」 She becomes aware of him and turns only after he calls. The boy stops rather than walking too close.
6.2–8.2s: Intimate close-up of the girl from [@BridgeReferenceSheet] as she recognizes him. Her initial surprise changes into cautious attention. One hearing aid is naturally visible from the correct side angle. Cut to the boy from [@BridgeReferenceSheet] lowering his gaze, taking one nervous breath and tightening his hands slightly around the notebook. No exaggerated crying or theatrical facial movement.
8.2–12.0s: Eye-level two-shot using the character positions and bridge environment from [@BridgeReferenceSheet] . The boy bows once and returns the notebook with both hands. He says quietly in Japanese:「これ、返したかった。」 He then performs the correct Japanese Sign Language phrase corresponding to「友達になれるかな」using slow, deliberate and anatomically accurate hand movements. Do not generate fast random gestures or extra fingers. The girl carefully watches his hands, accepts the notebook and replies softly:「ありがとう。」
12.0–15.0s: Wide lateral tracking shot on the same bridge from [@BridgeReferenceSheet] . They begin walking side by side with comfortable space between them. The girl looks toward the boy and gives a small genuine smile. He notices her expression and gradually relaxes. The camera continues tracking while the river glows behind them. End on their quiet forward movement, without an embrace or romantic pose.
Use 35mm bridge and river shots, 50mm two-shots and 85mm facial close-ups with natural depth of field and realistic motion blur. Sound design: river water, distant traffic, birds, wind, footsteps, clothing movement and notebook paper. Add a subtle, emotionally restrained Japanese piano score. All dialogue must be Japanese only, with no subtitles or narration. No anime, manga, cel shading, floating X symbols, bullying flashback, violence, injury, kissing, disrespectful hearing-aid depiction, random sign language, duplicated notebook, extra main characters, crowds, changing clothes, face drift, malformed fingers, plastic skin, logos or watermark.
VIDEO2
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese LIVE-ACTION character-and-environment reference sheet for a supernatural coming-of-age road adventure. This image will only be used as a global identity, wardrobe, prop and location reference for later video generation. It must contain detailed references for exactly ONE female protagonist and exactly ONE male protagonist, plus one empty primary location and several isolated key-prop details. ABSOLUTELY NO STORYBOARD, no chronological frames, no action sequence, no characters acting inside the location, no timeline, no shot numbers, no camera directions, no dialogue, no battle images and no narrative progression.
LAYOUT: A clean matte-black professional reference-sheet background divided into exactly three large vertical sections. Left section: female protagonist. Center section: male protagonist. Right section: primary location and key props. Use thin gray separators and only three simple Japanese section labels:「女子主人公」「男子主人公」「主要ロケーション」. At the top center, place the simple Japanese title「すずめ」. No English, Chinese, descriptive paragraphs, maps, arrows, technical notes, illegible small text or random symbols.
MALE PROTAGONIST: An original, natural-looking 17-year-old Japanese high-school girl with a slim but healthy teenage build, realistic height and believable ordinary attractiveness. Soft oval youthful face, natural light Japanese skin with visible pores and minor imperfections, dark-brown expressive eyes, a small natural nose and straight dark-brown shoulder-length hair tied into a practical low ponytail. Light uneven bangs and a few loose strands around her cheeks. No celebrity resemblance, no adult glamour, no beauty filter, no heavy makeup and no exaggerated anime eyes. She wears a clean white short-sleeved Japanese school blouse with a dark-red ribbon at the collar, a dark forest-green pleated knee-length skirt, dark ankle socks and brown leather loafers. She carries exactly one small brown school shoulder bag. Show one large chest-up identity portrait, exactly four consistent full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, curious, surprised, determined and relieved smile. Every view must show the exact same face, hairstyle, ribbon, clothes, bag and proportions. Full-body views must include complete shoes, natural arms and correct hands with five fingers.
MALE PROTAGONIST: An original, natural-looking 21-year-old Japanese university student with a tall, lean build and believable adult proportions. Quiet angular-oval face, natural light Japanese skin with visible pores, dark eyes and medium-length slightly wavy black hair that falls loosely around his forehead and ears. Calm, observant and serious, but not model-like or celebrity-like. He wears a muted teal-green lightweight overshirt left naturally open over a plain off-white T-shirt, straight dark charcoal trousers and worn practical dark sneakers. A small antique brass key hangs from a thin dark cord around his neck. He carries exactly one dark canvas crossbody bag. Show one large chest-up identity portrait, exactly four consistent full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, focused, alarmed, determined and gentle smile. Preserve the exact same person, hairstyle, clothing, necklace, bag and body structure across every view. Correct shoulders, hands, fingers and complete shoes.
PRIMARY LOCATION: A large empty photorealistic hero image of an abandoned Japanese hillside resort ruin near the sea on a clear late-summer afternoon. Cracked concrete foundations, remains of low demolished buildings, rusted railings, overgrown grass, small wildflowers, shallow rainwater puddles, scattered stones and distant green mountains descending toward a deep-blue ocean. In the center of the ruins stands exactly ONE freestanding weathered white wooden door in a complete rectangular frame, upright and physically touching the ground despite having no surrounding wall. The closed door has aged white paint, a dark metal handle, subtle cracks and believable weight. Behind it is only the same real ruin environment; do not show a portal in this reference image. Brilliant blue sky with large detailed white clouds, natural sunlight and realistic shadows. No people, silhouettes or human reflections anywhere in the location section.
KEY PROP DETAILS: Beneath the main location image, show exactly three isolated reference images. First: the same antique brass key and dark neck cord used by the male protagonist. Second: one small handmade yellow wooden children’s chair with exactly three legs because one original leg is missing; worn paint, rounded backrest and believable wood grain. The chair is an ordinary motionless prop in this image, with no face, eyes, mouth, arms or magical glow. Third: a close material view of the weathered white door handle, aged paint and lock. These are object references only, not story frames.
VISUAL QUALITY: Premium Japanese live-action casting and real-location photography, realistic skin pores, individual hair strands, correct anatomy, physical fabric textures, natural daylight, clean blue sky, detailed clouds and balanced exposure. The location should feel vast, quiet, abandoned and slightly mysterious without looking like a fantasy illustration. No anime, manga, cel shading, cartoon characters, 3D-rendered people, cosplay, plastic skin, celebrity likeness, excessive makeup, dramatic artificial lighting, fog, haze, glowing portal, red energy, smoke creature, disaster, damaged bodies, extra characters, duplicated faces, mixed identities, changing clothes, malformed hands, extra fingers, floating chair, four-legged chair, multiple doors, storyboard panels, action poses, numbered frames, captions, watermark or logo.
Video Prompt
CREATE AN EXACTLY 28-SECOND HORIZONTAL 16: 9, 30fps, premium ultra-photorealistic Japanese live-action short. Reproduce the narrative order, shot progression, emotional rhythm and coastal atmosphere of the provided reference video: a Japanese schoolgirl rides a red bicycle down a steep coastal road, passes a mysterious young man walking uphill, is unexpectedly captivated by him, brakes and looks back, then the sequence ends with a visual reveal of a solitary door standing inside a flooded abandoned ruin.
GLOBAL
REFERENCE LOCK: Use [@CoastalDoorReferenceSheet] as the exact identity and appearance reference for the FEMALE PROTAGONIST and MALE PROTAGONIST, including their faces, hairstyles, ages, body proportions, skin tone, clothing, shoes, bags and accessories. Use [@CoastalDoorReferenceSheet] as the visual reference for the coastal Japanese landscape, descending hillside road, blue sea, town, green mountains, bright summer sky, abandoned flooded ruin and freestanding door. Whenever either protagonist or either location appears, maintain the corresponding design from [@CoastalDoorReferenceSheet] with no identity, wardrobe or environmental drift.
REFERENCE-SHEET EXCLUSION: [@CoastalDoorReferenceSheet] is only an identity, wardrobe, prop and environment database. It is NOT the opening frame and NOT a storyboard. Never show its black background, Japanese headings, character turnarounds, portrait grid, repeated facial expressions, divided panels or multiple copies of either character. The finished video contains exactly ONE girl and exactly ONE young man. Do not animate the character-sheet poses.
VISUAL
STYLE: Completely photorealistic Japanese live action with real actors, real skin pores, natural facial asymmetry, individual hair strands, believable eyes, physical school clothing, realistic bicycle movement and natural sunlight. Bright coastal summer colors with clear blue sky, detailed white clouds, rich green hills and sparkling sea, but no overexposure, artificial fantasy glow, excessive lens flare, haze, plastic skin or beauty-filter appearance. Use direct cinematic cuts only. No fades, dissolves, morphing, transparency overlays, still-image animation or slideshow presentation.
0.0–1.7s — SHOT 1, BICYCLE DETAIL: Extreme low-angle close tracking shot beside the front wheel of the red city bicycle ridden by the girl from [@CoastalDoorReferenceSheet] . The wheel rotates forward on sunlit asphalt, spokes blur naturally, the tire follows the downhill road and the pedals turn at a believable speed. Her loafer remains correctly placed on the pedal. Camera tracks parallel to the wheel with subtle road vibration. The bicycle must not float, wobble unnaturally or change design.
1.7–4.8s — SHOT 2, COASTAL DESCENT: Cut to an enormous frontal wide shot looking uphill along the steep coastal road from [@CoastalDoorReferenceSheet] . The girl rides downhill toward the camera on the correct side of the empty road. The camera retreats smoothly in front of her. Behind her, the road descends toward a brilliant blue sea and a densely built coastal town. Utility poles, guardrails and green hills create strong perspective. She pedals normally, then relaxes into a gentle coast while maintaining both hands on the handlebars. No cars, pedestrians or other cyclists.
4.8–6.8s — SHOT 3, GIRL CLOSE-UP: Frontal 85mm moving close-up of the girl from [@CoastalDoorReferenceSheet] while she continues cycling. Her complete face is clear. Wind moves her ponytail, bangs and red uniform ribbon naturally. She initially smiles at the beautiful weather, then notices someone farther down the road. Her smile fades into quiet curiosity as her eyes shift slightly past the camera. Natural skin texture and restrained acting; no exaggerated wide anime eyes.
6.8–9.2s — SHOT 4, YOUNG MAN REVEAL: Long-lens full-body shot down the same road. The young man from [@CoastalDoorReferenceSheet] walks uphill toward her at an unhurried pace, carrying the same bag shown in [@CoastalDoorReferenceSheet] . He remains on his side of the road and looks focused on his destination. The sea, town and road geography remain identical to the previous shot. His hair and open outer shirt move in the sea breeze.
9.2–11.2s — SHOT 5, MALE PORTRAIT: Natural eye-level close-up of the young man from [@CoastalDoorReferenceSheet] as he raises his gaze and notices the approaching bicycle. Sunlight softly illuminates one side of his face. His expression is calm, slightly tired and mysterious rather than romantic or model-like. He slows his steps but does not stop in the centre of the road.
11.2–14.1s — SHOT 6, CROSSING PATHS: Clean side-profile wide shot parallel to the road, matching the reference video’s lateral composition. The girl from [@CoastalDoorReferenceSheet] cycles downhill from left to right while the young man from [@CoastalDoorReferenceSheet] walks uphill from right to left. They pass each other with approximately 1.2 metres of safe lateral clearance. Neither swerves into the other. As they draw level, he turns his head slightly toward her and asks in natural Japanese:「ねえ、この辺りに廃墟はない？」 His delivery is calm and clearly audible without speaking too quickly.
14.1–16.4s — SHOT 7, MOMENT OF EYE CONTACT: Low, close side angle with bright sky behind them. The bicycle continues moving forward while the girl turns her eyes toward the man for one brief moment. Their gazes meet as they pass, but their bodies continue in opposite directions. Her hair sweeps across her cheek in the wind. Use natural backlight with visible facial detail; do not turn either person into a silhouette.
16.4–18.5s — SHOT 8, DELAYED REACTION: Extreme facial close-up of the girl from [@CoastalDoorReferenceSheet] after she has passed him. Her curious expression changes into sudden, quiet fascination. Her eyes open slightly wider, lips part and she momentarily forgets to pedal, but the reaction remains realistic and restrained. She softly repeats to herself in Japanese:「廃墟……？」 No comedic shock face, distorted eyes or frozen expression.
18.5–21.2s — SHOT 9, BRAKING AND TURNING: Rear three-quarter tracking shot following the girl down the hill. She squeezes both brake levers, the bicycle slows progressively and the rear wheel remains aligned with the road. She places her left foot on the asphalt only after the bicycle nearly stops. Keeping her right hand on the handlebar, she turns her upper body and looks back uphill. Her bag and ponytail settle naturally after braking.
21.2–24.3s — SHOT 10, LOOKING BACK: Over-the-shoulder composition from behind the girl. The young man from [@CoastalDoorReferenceSheet] now stands approximately ten metres uphill, framed by green mountains and blue sky. Spatial continuity must clearly show that they have just passed each other. The girl calls to him in Japanese:「山の上に、古い温泉街があります！」 He turns toward her, gives a small appreciative nod and replies:「ありがとう。扉を探してるんだ。」 Use natural Japanese lip synchronization. Do not place them suddenly close together.
24.3–28.0s — SHOT 11, MYSTERIOUS DOOR REVEAL: Direct hard cut to the abandoned location from [@CoastalDoorReferenceSheet] . Begin with a wide forward-moving shot across perfectly still shallow water inside a large roofless ruin. Rusted steel roof frames, broken concrete walls, climbing vegetation and blue sky reflect naturally in the water. In the exact centre stands one solitary weathered white door in its complete frame, surrounded by water but no wall. The camera glides slowly toward it while small ripples spread across the water. The door remains completely closed and motionless. No characters enter this shot. End on a closer symmetrical view of the door as a faint metal creak is heard from behind it, but it must not open yet.
CAMERA AND CONTINUITY: Use 24mm coastal wides, 35mm bicycle tracking shots, 50mm side-profile coverage and 85mm facial close-ups. Natural 180-degree-shutter motion blur, realistic rolling wheels, stable road geography and smooth professional camera movement. The girl always travels downhill toward the sea; the man always travels uphill toward the mountains. After passing, they must remain on opposite sides of each other. No character teleportation, reversed travel direction or repeated encounter.
SOUND DESIGN: Natural bicycle chain, rotating tires, light brake sound, sea wind, insects, distant birds, fabric movement, footsteps and quiet coastal ambience. Dialogue is spoken naturally in Japanese only. Do not generate subtitles, narration or on-screen captions. Keep generated music minimal or absent because the official BGM will be added separately during editing. At the final location cut, gradually reduce the road ambience and introduce distant water droplets, empty-building resonance and one subtle door creak.
STRICT
NEGATIVES: No door-closing action, no portal opening, no supernatural smoke, no worm, no earthquake, no disaster, no cat, no animated chair, no three-legged chair running, no additional students, no cars, no collision, no falling from the bicycle, no man blocking the bicycle, no character standing dangerously in traffic, no romance embrace, no touching, no duplicated actors, no multiple doors, no wardrobe change, no face drift, no malformed hands, no missing bicycle parts, no reversed wheel rotation, no floating bicycle, no anime, manga, cel shading, 3D game graphics, cosplay, reference-sheet layout, black panel background, random text, title card, logo or watermark.
VIDEO3
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action CHARACTER AND ENVIRONMENT REFERENCE SHEET for a youthful Japanese fantasy romance. This image is exclusively an identity, costume and location reference for future video generation. Show exactly ONE female protagonist, exactly ONE male protagonist and an empty primary environment. ABSOLUTELY NO STORYBOARD, no body-swapping scene, no running scene, no hand-writing scene, no comet-action sequence, no timeline, no numbered frames, no maps, no diagrams and no narrative progression.
LAYOUT: Elegant matte-black production-reference background divided into exactly three large sections. Left: female protagonist identity. Center: male protagonist identity. Right: empty mountain environment. Use thin gray separators and only the Japanese labels「女子主人公」「男子主人公」「主要ロケーション」. Add the small Japanese title「黄昏の名前」at the top. No English, Chinese, dense description, technical notes, arrows or random lettering.
FEMALE PROTAGONIST: An original, believable 17-year-old Japanese high-school girl from a rural mountain town. Natural oval face, realistic light Japanese skin with subtle pores, expressive dark-brown eyes and long straight black hair tied into a practical high ponytail with one distinctive braided dark-red hair cord. Light natural bangs, no celebrity resemblance, no heavy makeup and no exaggerated anime features. She wears a white short-sleeved school blouse, dark-red neck ribbon, navy pleated knee-length skirt, dark socks, black loafers and one simple brown school bag. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, confused, surprised, warm smile and emotional. Preserve the exact same facial structure, hair cord, hairstyle, uniform and proportions in all views.
MALE PROTAGONIST: An original, believable 17-year-old Japanese high-school boy from Tokyo with a slim natural build, realistic youthful skin, quiet dark eyes and short tousled black hair. He has an ordinary attractive face with no celebrity resemblance and no model styling. He wears a white school shirt, slightly loosened dark tie, charcoal school trousers, understated sneakers and one dark backpack. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, confused, determined, surprised and restrained smile. Identical face, hair, wardrobe, backpack and proportions across all images.
ENVIRONMENT: A large empty photorealistic hero view from a grassy mountain ridge overlooking a vast crater lake and a small Japanese town during luminous late-summer twilight. Deep-blue mountain silhouettes, calm lake reflecting the sky, tiny warm town lights, long grass moving lightly in the wind, detailed violet-blue clouds and a coral glow along the horizon. A distant blue-white comet may be visible high in the sky as a stable environmental element, but it must not split, explode, fall or dominate the image. Beneath the main environment image, include exactly three small location-detail crops: wind moving through ridge grass, distant town lights reflected on the lake and twilight clouds above the mountains. No people, silhouettes or human reflections.
Premium live-action realism with natural skin, individual hair strands, correct hands, physical clothing, real outdoor lighting and restrained color grading. No anime, manga, cel shading, illustration, 3D characters, cosplay, plastic skin, celebrity face, glowing bodies, magical particles, meteor impact, town destruction, characters writing on hands, running characters, duplicated people, mixed identities, malformed fingers, wardrobe drift, storyboard images, numbered shots, map, timeline, captions, watermark or logo.
Video Prompt
Create an exactly 28.0-second horizontal 16:9, 30fps premium ultra-photorealistic Japanese live-action fantasy-romance sequence.
REFERENCE LOCK: Use [@TwilightNameReferenceSheet] only as the exact identity, face, hairstyle, body proportion, school clothing, red braided hair cord, shoes, bags and appearance reference for the girl and boy, and as the visual reference for the crater lake, grassy ridge, stone steps, distant town, mountains, twilight clouds and natural color palette. Preserve [@TwilightNameReferenceSheet] whenever either protagonist or the location appears.
The reference is not a storyboard or first frame. Never show its black layout, labels, expression panels, turnaround figures, maps or repeated bodies. Exactly one girl and one boy.
STYLE: Real Japanese actors in a real mountain landscape, natural skin pores, individual hair strands, correct hands, physical clothing and believable running. The comet may be fantastical, but its light must interact naturally with the clouds, lake, grass and actors. No body disappearance, invisible time barrier, teleportation, transparent people or confusing duplicated timelines. Both protagonists remain physically present together after they meet. Direct cinematic cuts only.
0.0–2.8s — Enormous aerial establishing shot of the crater lake and mountain town from [@TwilightNameReferenceSheet] . Coral sunset remains along the horizon while the upper sky turns violet-blue. A distant comet trail is visible but secondary. Camera descends toward the ridge as wind moves the grass consistently.
2.8–5.5s — Side tracking shot of the girl from [@TwilightNameReferenceSheet] running up the stone steps. Her steps, breathing, ponytail, red hair cord, skirt and bag respond naturally. She hears a distant voice and looks toward the ridge ahead.
5.5–8.2s — Reverse tracking shot of the boy from [@TwilightNameReferenceSheet] running along the opposite approach path. He calls:「誰かいるの？」The girl answers from offscreen:「ここにいる！」Their voices come from logically opposing directions.
8.2–10.8s — High wide shot establishes both paths converging on the same ridge. The girl enters from lower screen-left and the boy from upper screen-right. Both remain simultaneously visible and run toward the central crest without crossing or switching paths.
10.8–13.2s — Eye-level two-shot as they arrive and stop approximately two metres apart. They bend slightly to catch their breath, then slowly look up at each other. Wind moves their hair and clothing; their surprise remains restrained and human.
13.2–15.8s — Alternating 85mm close-ups. The girl asks:「私たち、会ったことある？」The boy answers:「ずっと、探してた気がする。」Maintain matching eyelines and accurate Japanese lip synchronization.
15.8–18.0s — Medium two-shot. They take one slow step closer. The girl asks:「君の名前は？」The boy removes one black marker and replies:「忘れないように、書こう。」All dialogue finishes completely by 18.0 seconds.
18.0–20.8s — Tight insert of their hands. The boy gently supports her wrist and writes slowly on her open palm. Ink appears only beneath the moving marker tip. Hands remain stable with five fingers each. The only legible Japanese text is「好きだ」; no random letters or additional writing.
20.8–23.2s — Close-up of the girl looking at her palm, then back at him. Her eyes become moist and she smiles. He closes the marker and lowers his hand. Both remain visibly present in the same physical space; neither fades away.
23.2–24.6s — Low wide shot behind them as the distant comet brightens and begins separating into two trails. Blue-white light moves naturally across the clouds and lake. They turn toward the sky together.
24.6–28.0s — Final continuous wide shot lasting 3.4 seconds. The two protagonists stand side by side on the ridge with natural space between them, silently watching the comet divide above the lake. The girl closes her written hand against her chest; the boy glances briefly toward her and then back to the sky. No dialogue after 18.0 seconds. End on the vast landscape without impact, explosion or disappearance.
SOUND: Diegetic ambience only—mountain wind, moving grass, footsteps, breathing, insects, distant town sounds, marker cap click and subtle low atmospheric vibration from the comet.
ABSOLUTELY NO BGM, song, piano, orchestra or generated melody. No narration, subtitles or titles.
NEGATIVES: no disappearing characters, no duplicated timelines, no body swapping, no town destruction, meteor impact, explosion, excessive particles, glowing bodies, extra characters, face drift, costume changes, malformed hands, instant writing, anime, manga, cel shading, reference-sheet layout, logo or watermark.
Create Japanese Live-Action Reference-to-Video Prompts on Pollo AI
Upload your reference sheet and prompt to generate a polished character-locked AI video workflow.
Start Creating Free
VIDEO4
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action CHARACTER AND ENVIRONMENT REFERENCE SHEET for a quiet rain-themed youth drama. This image is only a clean global visual reference for later video generation. It must contain character identity references for exactly ONE male protagonist and exactly ONE female protagonist, plus an empty environment reference. ABSOLUTELY NO STORYBOARD, no chronological frames, no action sequence, no timeline, no shot list, no arrows, no blocking diagram and no narrative images.
LAYOUT: A refined matte-black production-sheet background divided into exactly three large sections. Left section: male protagonist identity sheet. Center section: female protagonist identity sheet. Right section: environment reference. Use clean thin gray separators and only three simple Japanese labels:「男子主人公」「女子主人公」「主要ロケーション」. At the top, include the small Japanese title「雨音の庭」. No English, Chinese, descriptive paragraphs, technical notes, tiny text or random symbols.
MALE PROTAGONIST: An original, natural-looking 18-year-old Japanese male student and aspiring shoemaker. Slim but healthy build, believable height, youthful oval-angular face, fair natural Japanese skin with visible pores and slight imperfections, dark-brown eyes, straight black hair with soft natural layers, subtle uneven fringe, clean-shaven, no celebrity resemblance and no model-like glamour. He wears a clean white long-sleeved school shirt with the sleeves casually rolled to the forearms, dark charcoal straight trousers, black belt, dark practical sneakers and a simple black backpack. He carries one transparent umbrella and one plain sketchbook. Show exactly one large chest-up identity portrait, exactly four consistent full-body turnaround views—front, left profile, back and three-quarter—and exactly five small facial close-ups: neutral, focused, slightly nervous, gentle smile and quietly concerned. Every view must show the exact same person, hair, face, clothing and proportions. Full-body views must include complete shoes and natural hands with five fingers.
FEMALE PROTAGONIST: An original, natural-looking 27-year-old Japanese woman with believable adult proportions, a softly oval face, realistic fair skin with pores, dark-brown eyes, straight shoulder-length dark-brown hair with light natural layering, restrained everyday makeup and no celebrity resemblance. She is attractive in a quiet, ordinary and mature way, not glamorous or fashion-model-like. She wears a cream blouse, muted teal lightweight cardigan, dark navy knee-length skirt and simple low-heeled shoes. No jewelry except tiny understated earrings. Show exactly one large chest-up identity portrait, exactly four consistent full-body turnaround views—front, left profile, back and three-quarter—and exactly five small facial close-ups: neutral, thoughtful, faint smile, sadness and relief. Preserve the exact same identity and wardrobe in every view. Correct hands, fingers, shoulders and body proportions.
ENVIRONMENT: A large empty photorealistic location image of a secluded wooden garden shelter during steady early-summer rain. Dense vivid green foliage, moss-covered stones, wet stone path, pond with realistic rain ripples, dripping leaves, dark wet timber, an empty wooden bench and distant city buildings naturally softened by rainfall. Beneath the main location image, include exactly three small environment-detail crops showing rain striking green leaves, concentric ripples on the pond and wet wood beside a transparent umbrella. These are environmental details only, not story frames. No people, silhouettes, faces or human reflections anywhere in the environment section.
Premium real-location photography, natural overcast daylight, realistic rain direction, accurate wet-surface reflections, natural colors, clear material texture and cinematic depth of field. No anime, manga, illustration, cel shading, 3D characters, cosplay, plastic skin, beauty filter, excessive makeup, fantasy glow, sunshine, malformed hands, missing limbs, duplicated people, mixed identities, changing wardrobe, extra accessories, storyboard panels, numbered shots, camera diagrams, captions, watermark or logo.
Video Prompt
Create an exactly
28.0-second horizontal 16:9, 30fps premium ultra-photorealistic Japanese live-action rain drama.
REFERENCE LOCK: Use [@RainGardenReferenceSheet] only as the exact identity, face, hairstyle, body proportions, clothing, shoes, umbrella, backpack and sketchbook reference for the male and female protagonists, and as the visual reference for the rain-soaked Japanese garden, wooden shelter, bench, stone path, pond, foliage and muted green palette. Match [@RainGardenReferenceSheet] in every shot.
The reference is not a storyboard or first frame. Never show its black layout, headings, portrait grids, repeated expressions, turnaround figures or multiple versions of either protagonist. Exactly one male student, one adult woman, one transparent umbrella and one sketchbook.
STYLE: Completely photorealistic Japanese live action with real skin texture, individual damp hair strands, realistic eyes, correct hands, physical clothing and genuine rain. Quiet, mature and emotionally restrained. Direct cinematic cuts only; no fade, morph, transparent overlay, slideshow or frozen reference poses.
0.0–2.8s — Macro moving shot through the garden from [@RainGardenReferenceSheet] . Raindrops collect on leaves, fall onto moss-covered stones and create overlapping rings across the pond. Camera passes beneath a wet branch and reveals the wooden shelter in the distance.
2.8–5.7s — Low side tracking shot of the male protagonist from [@RainGardenReferenceSheet] walking along the stone path beneath the single transparent umbrella. His shoes create small splashes. The umbrella responds to wind and remains firmly held in one hand.
5.7–8.2s — Shoulder-level following shot as he approaches the shelter and notices the woman from [@RainGardenReferenceSheet] sitting alone. He stops outside, folds the umbrella with clear hand movements and places it upright beside the entrance. It never duplicates or disappears.
8.2–10.5s — Medium shot inside the shelter. The woman turns from the rain, recognizes him and gives a faint smile. She says:「また会ったね。」He nods and sits at the opposite end of the bench, preserving respectful space.
10.5–12.8s — Intimate close-up of the student. He looks toward the rain and replies:「雨の日は、ここに来たくなるんです。」Cut to her listening quietly, with natural rain reflections in her eyes.
12.8–15.7s — Insert shot of his hands opening the single sketchbook and beginning a shoe drawing. The pencil creates each line only where it touches the page. The woman leans slightly to see and asks:「靴を作ってるの？」
15.7–18.3s — Eye-level two-shot. He continues drawing and answers:「はい。いつか、自分の手で作りたいんです。」She looks from the page to his face, impressed by his seriousness. Their shoulders and hands remain separate.
18.3–20.8s — Close-up of the woman. She looks briefly at her shoes, then toward the garden and says softly:「歩き出すための靴なら、素敵ね。」Her emotion remains restrained without crying.
20.8–23.5s — Wide exterior view as rain becomes heavier and water pours from the roof edge. The student protects the sketchbook; the woman moves the single umbrella farther under the roof. As the shower softens, she says:「また、雨の日に。」He answers:「はい。」All dialogue finishes completely by 23.5 seconds.
23.5–28.0s — Continuous silent walking shot lasting the full final 4.5 seconds. They leave the shelter and walk side by side beneath the single transparent umbrella. No one speaks. The student holds it; the woman walks naturally within its cover without clinging to him. Camera tracks backward in front of them for 2.5 seconds, then lets them pass and watches from behind for another 2 seconds as they continue deeper into the garden. Their shared walk lasts longer than three seconds and never begins only at the final moment.
SOUND: Rain on leaves, pond ripples, roof drips, wet footsteps, umbrella fabric, pencil on paper, clothing and distant birds only.
ABSOLUTELY NO BGM, song, piano, orchestral score, humming or generated music. No narration, subtitles or titles.
NEGATIVES: no kiss, embrace, sexual framing, extra visitors, alcohol, duplicated umbrella or sketchbook, instantly completed drawing, barefoot close-up, sunshine, artificial fog, face drift, wardrobe change, malformed hands, anime, manga, cel shading, reference-sheet layout, logo or watermark.
VIDEO5
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action CHARACTER AND ENVIRONMENT REFERENCE SHEET for a nostalgic spring youth romance. This image will be used only to lock the protagonists’ identities, clothing and primary location during later global-reference video generation. It must contain exactly ONE male protagonist, exactly ONE female protagonist and an empty environment reference. ABSOLUTELY NO STORYBOARD, no train-action sequence, no chronological images, no timeline, no shot numbers, no blocking diagram and no narrative progression.
LAYOUT: Matte-black professional reference-sheet background divided into exactly three large sections. Left: female protagonist identity sheet. Center: male protagonist identity sheet. Right: empty railway-crossing environment. Use thin gray borders and only three Japanese labels:「女子主人公」「男子主人公」「主要ロケーション」. At the top, include the small Japanese title「桜の距離」. No English, Chinese, paragraphs, technical data or tiny unreadable writing.
FEMALE PROTAGONIST: An original, natural-looking 17-year-old Japanese high-school girl with a believable ordinary appearance, softly oval youthful face, natural light Japanese skin with visible texture, dark-brown almond-shaped eyes, small natural nose and straight dark-brown hair tied into a loose low ponytail. Light wispy bangs and a few loose strands near her cheeks. No celebrity resemblance, no heavy makeup, no airbrushed skin and no adult glamour. She wears a crisp white short-sleeved Japanese school blouse with a subtle navy collar, navy pleated knee-length skirt, dark navy socks and polished black loafers, carrying exactly one rectangular navy school bag. Show one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, gentle smile, curious, surprised and quietly sad. Preserve the same face, hairstyle, uniform, bag and proportions in every view.
MALE PROTAGONIST: An original, natural-looking 17-year-old Japanese high-school boy, slightly taller than the girl, with a slim build, believable youthful face, realistic light skin, dark-brown eyes and short slightly wavy black hair naturally disturbed at the fringe. Reserved and thoughtful, not glamorous, muscular or model-like. He wears a white short-sleeved school shirt, dark navy trousers, black school shoes and exactly one simple black shoulder bag worn across his body. Show one large chest-up portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, restrained smile, attentive, startled and lonely. Every view must preserve one consistent identity, wardrobe and body structure. Complete shoes, correct hands and five natural fingers.
ENVIRONMENT: A large empty photorealistic image of one quiet Japanese suburban railway crossing on a clear spring afternoon. Two steel railway tracks, textured gravel, black-and-yellow striped barriers, red warning lights, narrow pedestrian paths, white road markings, utility poles, overhead electrical cables, modest houses, green embankment and dense cherry trees in full bloom. Intensely clear blue sky with large white clouds, vivid but natural spring colors and physically correct sunlight. No people and no train in the main environment image. Below it, include exactly three small location-detail crops: a cherry branch with naturally spaced blossoms, the railway barrier and warning light, and polished steel rails surrounded by gravel and small spring weeds. These are material/location references, not storyboard shots.
Photorealistic Japanese live-action casting, real skin pores, individual hair strands, physical fabrics, natural daylight and accurate anatomy. No anime, manga, illustration, 3D rendering, cosplay, plastic skin, celebrity likeness, beauty filter, extra students, train sequence, falling-petal action frame, characters facing each other across tracks, narrative scenes, duplicated people, different uniforms, malformed hands, missing bags, text blocks, storyboard, maps, arrows, shot numbers, watermark or logo.
Video Prompt
Create an exactly 28.0-second horizontal 16:9, 30fps premium ultra-photorealistic Japanese live-action romantic short.
REFERENCE LOCK: Use [@RailwayCrossingReferenceSheet] only as the exact identity, face, hairstyle, body proportion, school uniform, shoes, bag and appearance reference for the FEMALE PROTAGONIST and MALE PROTAGONIST, and as the visual reference for the suburban railway crossing, rails, barriers, warning lights, cherry trees, houses, utility poles, spring sky and natural color palette. Match [@RailwayCrossingReferenceSheet] whenever either person or the setting appears.
[@RailwayCrossingReferenceSheet] is only a character-and-location reference sheet, never the first frame and never a storyboard. Never show its black background, Japanese headings, divided layout, turnaround figures, facial-expression grid or repeated copies of either protagonist. Exactly one girl and one boy throughout.
CRITICAL GEOGRAPHY LOCK: The two protagonists begin on physically opposite sides of the tracks and remain on opposite sides until the barriers have fully risen after the train.
The tracks run horizontally from screen-left to screen-right. The BOY is always on the NEAR/SOUTH side and occupies screen-left in every shared wide shot. The GIRL is always on the FAR/NORTH side and occupies screen-right in every shared wide shot. The camera never crosses the 180-degree railway axis.
In close-ups, the boy looks screen-right toward the girl; the girl looks screen-left toward the boy. The train always moves screen-left to screen-right. Never place them beside each other before the train, never reverse their sides, never move either person through the train, and never make the girl disappear.
After the final carriage clears, both characters must be simultaneously visible in the same wide frame, still standing in their original positions.
STYLE: Real Japanese actors, natural pores, individual hair strands, accurate hands, physical uniforms, believable train weight and correct railway engineering. Clear spring daylight, blue sky and pale cherry blossoms with natural exposure. Direct cinematic cuts only; no fade, morph, dissolve, transparent overlay, slideshow or artificial beauty filter.
0.0–2.2s — Low macro tracking shot on the FAR/NORTH side: the girl’s loafer steps onto the pedestrian approach and stops safely behind her barrier line. One cherry petal falls and brushes the shoe. Her bag swings naturally. No train is visible or audible.
2.2–4.8s — Frontal medium tracking shot of the girl from [@RailwayCrossingReferenceSheet] walking beneath the cherry trees on the FAR side. She notices the boy across the tracks, slows and gives a small restrained smile. Her eyeline points clearly screen-left across the railway, never toward someone beside her.
4.8–7.2s — Matching medium shot of the boy from [@RailwayCrossingReferenceSheet] arriving on the NEAR side. He stops behind the near safety line, adjusts his shoulder bag once and looks screen-right toward the girl. The far-side girl remains visible as a small background figure across both tracks, proving their separation.
7.2–10.0s — Static symmetrical wide shot that clearly establishes geography: boy screen-left in the near foreground; two complete steel tracks across the middle; girl screen-right on the far side; both barriers raised; no train. Hold this composition long enough to make their opposite positions unmistakable. Camera makes only a very slow forward push.
10.0–12.8s — 85mm close-up of the girl from [@RailwayCrossingReferenceSheet] , looking screen-left. She watches a petal and says naturally in Japanese:「ねえ、桜の花びらって、 [@RailwayCrossingReferenceSheet] 五センチなんだって。」Accurate lip synchronization, calm pace and no subtitles.
12.8–14.8s — Close-up of the boy from [@RailwayCrossingReferenceSheet] , still looking screen-right. He answers softly:「そんなに、ゆっくりなんだ。」The girl replies from the correct distant direction:「また会えるよ。」All dialogue finishes completely by 14.8 seconds.
14.8–17.3s — Return to the same wide geography. Warning lights activate and both barriers descend at the same time. The boy and girl each make one instinctive half-step and stop safely behind their own barriers. Neither changes side. The crossing bell and barrier motors are clearly audible.
17.3–21.8s — Low side wide shot from the NEAR side. A realistic commuter train enters screen-left and travels to screen-right, temporarily blocking their view. The boy remains visible near screen-left behind the foreground barrier. The girl is physically hidden only because the train occupies the middle depth layer; she remains stationary behind the far barrier. Do not make her walk away, teleport or dissolve. Train wind lifts hair, clothing and petals with correct direction.
21.8–24.3s — The final carriage exits screen-right. In the first frame after it clears, immediately reveal BOTH protagonists together in the same wide composition: boy still near-left, girl still far-right, looking directly at each other. The girl has not moved or disappeared. They both relax and exchange a small smile while the barriers rise fully. The boy does not search for her or look around in confusion.
24.3–28.0s — Only after both barriers are completely raised, the girl and boy begin walking toward each other along the marked pedestrian crossing from their original opposite sides. Keep both visible in one continuous wide-to-medium push for the full final 3.7 seconds. Their routes remain straight and logical; they take several natural steps and approach without suddenly jumping position. End as they stop near the center pedestrian zone facing one another, still separated by comfortable space.
No speech during the final 13.2 seconds. The ending uses only expressions, footsteps, spring ambience and train noise fading away.
SOUND: Diegetic sound only—spring wind, birds, distant neighbourhood ambience, footsteps, clothing, petals, crossing bell, barrier motors, rail vibration, train wheels and fading metal resonance.
ABSOLUTELY NO BGM, no song, no piano, no orchestral score and no generated music; music will be added later. No narration, subtitles, title or text.
NEGATIVES: no same-side placement, no flipped geography, no camera crossing the railway axis, no missing girl, no searching boy, no girl walking away behind the train, no wrong-direction train, no people on active rails, no extra students, duplicates, face drift, wardrobe change, malformed hands, anime, cel shading, 3D graphics, reference-sheet layout, logo or watermark.
VIDEO6
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action CHARACTER AND ENVIRONMENT REFERENCE SHEET for a rain-heavy Japanese youth fantasy. This image is only for locking the protagonists’ identities, wardrobe and rooftop environment during later global-reference video generation. Show exactly ONE male protagonist, exactly ONE female protagonist and an empty primary location. ABSOLUTELY NO STORYBOARD, no praying scene, no rooftop meeting, no cloud-opening action, no sunlight transformation, no chronological frames, no timeline and no narrative sequence.
LAYOUT: Refined matte-black reference-sheet background divided into exactly three large sections. Left: male protagonist identity sheet. Center: female protagonist identity sheet. Right: empty rooftop environment. Thin gray separators and only the Japanese labels「男子主人公」「女子主人公」「主要ロケーション」. Add the small Japanese title「空が晴れる瞬間」at the top. No English, Chinese, descriptive paragraphs, technical notes, maps, arrows or random text.
MALE PROTAGONIST: An original, natural-looking 18-year-old Japanese boy with a slim youthful build, believable ordinary face, realistic light Japanese skin with visible pores, dark-brown eyes and short naturally messy black hair. He looks slightly tired but determined, without celebrity resemblance, model styling or heavy makeup. He wears a dark charcoal T-shirt, light-gray hooded jacket, black trousers, worn dark sneakers and one simple canvas shoulder bag. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, tired, curious, determined and genuine smile. Preserve the exact same face, hair, clothing, bag and proportions in every view.
FEMALE PROTAGONIST: An original, natural-looking 18-year-old Japanese girl with a gentle oval face, realistic skin texture, dark expressive eyes and shoulder-length black hair tied partly back with one small blue ribbon. Ordinary youthful beauty, no celebrity resemblance, no oversized anime eyes, no heavy makeup and no beauty filter. She wears a white blouse, pale-blue hooded jacket, navy knee-length skirt and clean white sneakers. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, gentle smile, focused, hopeful and emotional. Her hands remain in a relaxed neutral position; do not show praying poses. Preserve the exact same ribbon, hair, clothing and proportions.
ENVIRONMENT: A large empty photorealistic image of a rain-soaked Tokyo rooftop overlooking dense realistic urban buildings. Wet concrete, shallow puddles, metal stairs, rooftop door, water tanks, utility structures, power lines, drainage channels, safety fencing and one small weathered red gateway positioned securely far from the roof edge. Enormous dark storm clouds fill the sky, with only a faint naturally brighter area behind them; the clouds must remain closed, with no dramatic sunbeam and no active weather transformation. Beneath the main image, include exactly three small environment-detail crops: rain striking metal railing, realistic puddle reflections and dense storm-cloud texture. No characters, silhouettes, praying figures or human reflections.
Premium photorealistic live action with real Japanese casting, natural skin pores, individual hair strands, correct hands, physical wet fabrics, accurate rain direction and realistic rooftop materials. No anime, manga, illustration, cel shading, 3D characters, cosplay, plastic skin, celebrity faces, glowing eyes, praying pose, magical sunlight, cloud-opening sequence, flying people, dangerous roof-edge positioning, lightning strike, city destruction, duplicated actors, wardrobe drift, malformed hands, storyboard panels, numbered shots, diagrams, captions, watermark or logo.
Video Prompt
Create an exactly
28.0-second horizontal 16:9, 30fps premium ultra-photorealistic Japanese live-action weather-fantasy sequence.
REFERENCE LOCK: Use [@WeatherRooftopReferenceSheet] only as the exact identity, face, hairstyle, body proportion, clothing, blue ribbon, shoes, shoulder bag and appearance reference for the boy and girl, and as the visual reference for the rain-soaked Tokyo rooftop, entrance door, metal stairs, wet concrete, fencing, red gateway, skyline and storm-cloud palette. Preserve [@WeatherRooftopReferenceSheet] precisely.
The reference is not a storyboard or opening frame. Never show its black layout, labels, portraits, expression grid, turnaround figures or multiple character copies. Exactly one boy and one girl.
STYLE: Real Japanese actors, natural pores, individual wet hair strands, soaked physical fabrics, accurate rain direction and realistic rooftop materials. Weather follows strict cause and effect: uninterrupted heavy rain first, prayer second, clouds open third, sunlight travels downward fourth, rain stops only inside the light fifth. Direct cinematic cuts only; no fade, morph, overlay or game-like effect.
0.0–2.8s — Enormous high-angle moving view of Tokyo under relentless rain. Roads and roofs reflect cold gray daylight. Dense clouds fully cover the sky; no blue opening or sunlight exists yet. Camera pushes toward the rooftop from [@WeatherRooftopReferenceSheet] .
2.8–5.5s — Handheld medium tracking shot of the boy from [@WeatherRooftopReferenceSheet] pushing open the rooftop door and stepping into wind-driven rain. He braces against the gust, wipes water from his eyes and notices the girl near the red gateway. The door remains physically behind him and swings partly closed.
5.5–8.0s — Frontal medium shot of the girl from [@WeatherRooftopReferenceSheet] , safely far from the roof edge. Rain runs from her hair, ribbon and sleeves. She turns toward the boy and gives a calm smile. He approaches through the background but stops several metres away.
8.0–10.5s — Reverse close-up of the boy looking from her to the completely closed storm sky. He asks:「本当に、晴れさせられるの？」Heavy rain continues at the same strength.
10.5–12.8s — Close-up of the girl. She looks upward and says:「空に、お願いしてみる。」She takes a steady breath while rain continues striking her face and clothing.
12.8–15.2s — Low-angle medium shot. She brings both hands together slowly with correct fingers, closes her eyes and says clearly:「ねえ、今から晴れるよ。」Her line finishes completely by 15.2 seconds. No weather change begins before the sentence ends.
15.2–17.8s — Over-the-shoulder shot toward the sky. Only now do the clouds begin rotating and separating gradually. A small circular blue opening appears directly above the rooftop. Rain remains heavy around both characters at first.
17.8–20.8s — Majestic low wide shot. Warm sunlight travels visibly from the cloud opening down toward the roof instead of appearing everywhere at once. Wet-surface reflections brighten only where the light reaches. Rain weakens inside that circle while continuing over the surrounding city.
20.8–23.2s — Alternating close-ups of the girl opening her eyes and the boy looking around in disbelief. Backlit raindrops weaken and stop around them. He steps into the light and says:「すごい……本当に晴れた。」The final spoken line finishes completely by 23.2 seconds.
23.2–28.0s — Continuous silent final shot lasting 4.8 seconds. Both protagonists stand side by side inside the sunlit area with natural space between them. They take three slow steps forward across the safe central roof area, stop and look upward. Rain remains visible across the surrounding blue-gray city like a distant curtain. Sunlight expands across only a few neighbouring rooftops. No dialogue during the final shot; finish with water dripping from the railing and birds emerging in the distance.
SOUND: Heavy layered rain, wind, rooftop drainage, metal door movement, wet footsteps, distant traffic, water dripping and birds after the sunlight arrives.
ABSOLUTELY NO BGM, song, piano, orchestra, humming or generated melody. No narration, subtitles or titles.
NEGATIVES: no early sunlight, instant clear sky, wrong rain direction, lightning strike, city destruction, roof-edge danger, flying people, glowing eyes, magical particles, instant dry clothing, duplicated characters, face drift, malformed hands, anime, cel shading, reference-board layout, logo or watermark.
VIDEO7
Image Prompt
CREATE ONE SINGLE HORIZONTAL 16: 9 ultra-photorealistic Japanese live-action CHARACTER AND ENVIRONMENT REFERENCE SHEET for a sensitive coming-of-age drama. This image is only a clean visual database for the protagonists’ identities, wardrobe, hearing aid, communication notebook and primary location. Show exactly ONE female protagonist, exactly ONE male protagonist and an empty river-bridge environment. ABSOLUTELY NO STORYBOARD, no reunion sequence, no bullying scene, no sign-language action sequence, no walking scene, no timeline, no shot numbers and no narrative frames.
LAYOUT: Clean matte-black production-reference background divided into exactly three large sections. Left: female protagonist identity sheet. Center: male protagonist identity sheet. Right: empty primary environment. Use thin gray separators and only the Japanese labels「女子主人公」「男子主人公」「主要ロケーション」. Add the small Japanese title「届かなかった声」at the top. No English, Chinese, paragraphs, technical specifications, diagrams or random symbols.
FEMALE PROTAGONIST: An original, natural-looking 18-year-old Japanese female student with a gentle round-oval face, realistic fair Japanese skin with visible texture, soft brown eyes and a straight chestnut-brown bob ending slightly below the ears. She wears two small anatomically correct behind-the-ear hearing aids positioned naturally behind both ears. The hearing aids must remain subtle, realistic and identical in every relevant view. She wears a cream school cardigan over a white blouse, navy ribbon, navy pleated skirt, dark socks and brown loafers. She carries one light school bag and one worn communication notebook. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, attentive, uncertain, relieved and softly smiling. Preserve the same identity, hair, hearing aids, clothing and proportions throughout.
MALE PROTAGONIST: An original, natural-looking 18-year-old Japanese male student with a slim build, slightly tired but kind dark eyes, realistic skin and short slightly messy black hair. Reserved, nervous and believable, with no celebrity resemblance, fashionable styling or model-like glamour. He wears a white school shirt, loosely worn navy tie, dark trousers, understated sneakers and one black backpack. Show exactly one large chest-up identity portrait, exactly four full-body turnaround views—front, left profile, back and three-quarter—and exactly five facial close-ups: neutral, nervous, regretful, hesitant and gently relieved. Preserve one consistent identity, hairstyle, wardrobe and body structure. Hands must be relaxed and neutral; do not depict sign language in this reference image.
ENVIRONMENT: A large empty photorealistic image of a quiet pedestrian bridge above a broad Japanese river during late-summer afternoon. Blue-green flowing water, visible koi beneath the surface near the riverbank, grassy embankments, bicycle path, distant apartment buildings, summer trees, clean metal bridge railings and warm peach sunlight. The bridge must be safely designed with realistic proportions. Beneath the main environment image, include exactly three small detail references: a realistic behind-the-ear hearing aid shown as an isolated product detail, the worn communication notebook closed on a neutral surface and koi beneath moving river water. These are object and environment references, not scenes. No people, hands, silhouettes or reflections of characters.
Sensitive premium live-action realism, natural skin pores, realistic hair, correct anatomy, respectful disability representation and soft natural daylight. No anime, manga, cel shading, floating symbols, drawn X marks, 3D rendering, cosplay, plastic skin, enlarged hearing aids, incorrect device placement, sign-language pose, bullying, violence, crying scene, romantic interaction, duplicated people, mixed faces, malformed hands, storyboard frames, shot numbers, maps, captions, watermark or logo.
Video Prompt
Create an exactly
28.0-second horizontal 16:9, 30fps ultra-photorealistic Japanese live-action emotional drama.
REFERENCE LOCK: Use [@BridgeReferenceSheet] only as the exact identity, face, hairstyle, body proportions, school clothing, backpack, hearing aids and communication-notebook reference for the girl and boy, and as the visual reference for the river, pedestrian bridge, railing, koi, embankment, trees, distant buildings and late-afternoon palette. Match [@BridgeReferenceSheet] whenever either person, object or setting appears.
The reference is not a storyboard. Never show its black layout, labels, repeated portraits, product panels, turnaround bodies or multiple copies. Exactly one girl, one boy and one notebook.
STYLE: Sensitive premium Japanese live action with natural pores, realistic hair, correct anatomy, accurate hearing-aid placement and respectful disability representation. Restrained acting, natural late-afternoon sunlight and direct cinematic cuts. No floating graphic symbols, morphs, fades, transparent overlays or slideshow presentation.
0.0–2.5s — Macro tracking shot above the river from [@BridgeReferenceSheet] . Koi swim beneath the surface as small pieces of fish food land individually and create expanding ripples. Sunlight moves across the water naturally.
2.5–5.2s — Medium-wide rear shot of the girl from [@BridgeReferenceSheet] standing safely beside the bridge railing. She watches the koi. Wind moves her bobbed hair and cardigan. One hearing aid is subtly visible from the correct angle, never enlarged or glowing.
5.2–7.8s — Long-lens shot of the boy from [@BridgeReferenceSheet] approaching slowly from the far end of the bridge. He carries exactly one worn communication notebook with both hands. He stops several respectful steps behind her and gathers his courage.
7.8–9.8s — Close-up of the boy. He softly calls:「西宮。」He does not touch her or suddenly move closer. His nervous expression and breath are natural.
9.8–12.0s — Intimate close-up of the girl turning after noticing him. Her expression changes from surprise to cautious recognition. Cut to the boy lowering his gaze, then looking back up. Maintain correct matching eyelines.
12.0–14.5s — Eye-level medium two-shot. The boy takes one measured step closer but leaves comfortable space. He bows once and says:「急に呼び止めて、ごめん。」The girl watches attentively and gives a small nod.
14.5–17.3s — Close insert as he extends the single notebook with both hands and says:「これ、ずっと返したかった。」The girl accepts it only after he releases his grip. Their fingers remain separate and anatomically correct. The notebook never duplicates or changes design.
17.3–19.4s — Close-up of the girl looking at the worn cover. She traces one crease with her thumb, raises her eyes and replies softly:「ありがとう。」Her emotion remains restrained; no dramatic crying.
19.4–22.3s — Medium shot focused on the boy’s upper body and hands. He performs a short, correct and simple Japanese Sign Language phrase corresponding to「友達になれるかな」using slow, anatomically accurate movements while speaking it quietly. No random rapid hand motion. The spoken line finishes completely before 22.3 seconds.
22.3–24.0s — Close-up of the girl. Relief replaces uncertainty. She gives a clear affirmative nod and a small correct sign, then answers:「うん。」The word finishes by 23.0 seconds, leaving the ending free of speech.
24.0–28.0s — Wide lateral tracking shot of them walking side by side across the bridge for the full final 4.0 seconds. No one speaks. Maintain a comfortable respectful distance. The girl carries the single notebook against her chest; the boy wears his backpack normally. Camera tracks parallel for three seconds, then slows while they continue forward. She glances at him and smiles; he returns a restrained smile. Their walk never lasts less than three seconds.
SOUND: River water, wind, birds, distant bicycles, footsteps, clothing and notebook paper only.
ABSOLUTELY NO BGM, song, piano, orchestral score or generated music. No narration, subtitles or titles.
NEGATIVES: no incorrect hearing aids, random sign language, duplicated notebook, floating X symbols, bullying, violence, kiss, embrace, extra protagonists, face drift, costume changes, malformed fingers, anime, illustration, reference-sheet layout, logo or watermark.
Final Summary
Create polished Japanese live-action image and video prompt workflows from reusable reference sheets. Keep character identity, wardrobe, props, environment, timing, camera language, and Japanese dialogue consistent from reference image to final video.Want to bring your own reference-to-video scene to life in minutes? Upload your character or environment references, paste the prompt, and create a cinematic AI video workflow with Pollo AI.