Chipmunk Anime Fight Sequence

Super casual real smartphone home video footage, sunny summer beach day in Japan (Okinawa-style emerald water and white sand), outing with friends, natural mobile phone camera with slight authentic handheld shake, normal frame rate with smooth natural motion, rapidfire montage with constant quick jump cuts every 1-2 seconds like scrolling through phone memories, unpolished authentic phone recording of a mixed group laughing, playing in the sand, snacking on Japanese convenience-store treats and clicking pictures, pure raw home video feel, no cinematic polish or heavy effects. Use the provided reference photo as the strict ONLY visual reference for the main woman. Maintain her exact appearance with zero deviation: [her described features]. Generate a mixed group of Japanese friends of all ages around her on the beach. 0-2.5s: Shaky handheld rapid cuts, main woman laughing with friends near the shoreline of a Japanese beach, wind in her hair, quick flashes of feet in the sand and small waves, a traditional beach house (umi-no-ie) with noren curtains blurred behind. 2.5-5s: Abrupt jump cuts, close-up of her smiling with sunglasses pushed up, then the group splashing water and clicking selfies, a Ramune soda bottle catching the sun. 5-7.5s: Fast shaky, she sits on a beach towel chatting animatedly, onigiri, Japanese snacks and cans of beer scattered around, natural sunlight flare. 7.5-10s: Quick cut close-up, big warm smile toward camera, playful peace-sign, then jump to the group building a small sandcastle together. 10-12.5s: Abrupt edit, the adult friends gathered under a beach umbrella, passing cans of beer around and clinking the beer cans together in a casual toast. 12.5-15s: Final rapid transition, main woman relaxed, sitting among friends as the sun lowers, soft smile, calm Japanese-summer beach memory ending with a gentle phone sway. Natural smartphone video quality, slight real handheld shake, smooth normal frame rate motion, authentic casual interactions and physics, stable main character consistency, unpolished home phone recording, no pro stabilization or effects. Negative: no cinematic polish, no on-screen text, subtitles, logos, or watermarks; hands with exactly five fingers.

같은 젊은 한국 남성이 영상 내내 일관되게 등장하는 초사실적 시네마틱 라이프스타일 영상. 얼굴·날카롭고 또렷한 눈매·차분한 표정·어두운 레이어드 미디엄 헤어·실버 이어커프·체형을 모든 씬에서 동일하게 유지. 첨부 사진을 얼굴과 헤어의 엄격한 기준으로 사용. 리얼리즘: 모공과 결이 보이는 실제 피부, 왁스·플라스틱 같지 않게. 고운 35mm 필름 그레인, 자연스러운 모션 블러. 실제 그림자와 대비가 있는 강한 방향성 조명 — 평평한 균일 조명 금지. 완급 있는 속도, 모든 게 슬로모 아님. 실제 시네마 카메라 촬영본처럼, AI 생성·부자연스러운 느낌 금지. 의상 원칙: 얼굴과 헤어는 내내 동일. 의상은 각 장소에 맞게 씬별로 자연스럽게 바뀐다 — 모든 컷에서 같은 옷이 아니다. - 시그니처 룩(드레스룸 이후, 와인바): 네이비 코튼 집업 자켓(카라, 오픈) 안에 다크 브라운 니트 폴로, 워시드 차콜블랙 스트레이트 진 + 가죽 벨트, 블랙 가죽 스니커즈, 미니멀 시계 — 첨부 사진 그대로. - 아침 기상 씬: 무지 크림색 반팔티 + 크림색 반바지(편한 잠옷), 자켓 아님, 긴바지 아님. - 헬스 씬: 운동복만 — 다크 탱크탑 + 트레이닝 쇼츠. 헬스 씬 첫 프레임부터 이미 탱크탑 차림, 어느 순간에도 자켓 없음. 각 씬 의상은 아래 타임코드에도 다시 명시 — 그대로 따를 것. 스타일: 럭셔리 한국 라이프스타일 광고, 포토리얼 시네마틱 다큐, 부드러운 핸드헬드, 웜 파스텔 그레이딩, 얕은 심도, 프리미엄 남성복 광고 미감, 4K HDR, 24fps, 16:9. 0-3초 — 아침, 한강 펜트하우스: 한강이 내려다보이는 통창의 럭셔리 미니멀 펜트하우스, 해 뜰 무렵. 무지 크림색 티셔츠와 크림색 반바지 차림. 따뜻한 빛 속에 자연스럽게 깨어 일어나 앉아 강 뷰를 바라본다. 느리고 차분하게. 3-6초 — 드레스룸 코디: 웜 오크 선반의 우아한 워크인 드레스룸. 카메라 앞에서 시그니처 룩을 완성해간다 — 브라운 니트 폴로를 입고, 그 위에 네이비 자켓을 걸치고, 카라를 정리하고, 차콜 진에 벨트를 매고, 시계를 확인하고, 전신 거울 앞에서 머리를 매만진다. 마크로 클로즈업: 원단 질감, 카라 잡는 손, 손목 시계. 이제 시그니처 룩 완성. 6-9초 — 한강변 스포츠카: 시그니처 룩 그대로, 한강 강변 도로로 나오면 매끈한 럭셔리 스포츠카가 기다린다. 다가서서 문을 여는 그를 바로 위에서 내려다보는 항공(버즈아이) 앵글로만 촬영 — 드론이나 비행 기기는 화면에 절대 보이지 않음. 강변 고속도로를 달리고, 아침 햇살이 윈드실드에 렌즈플레어, 뒤로 도시 스카이라인. 9-12초 — 헬스장 운동: 큰 창의 프리미엄 헬스장으로 컷. 첫 프레임부터 이미 다크 운동복 탱크탑 + 트레이닝 쇼츠(자켓 없음). 절제되고 파워풀한 동작 — 덤벨, 케이블 머신, 집중된 호흡. 근육 컨트롤·자세·차분하지만 강렬한 표정·사실적 피부와 옅은 땀·자연광 클로즈업. 12-15초 — 루프탑 와인바, 친구들과: 다시 시그니처 네이비 자켓 룩으로, 골든아워의 우아한 오픈에어 루프탑 와인바, 서울 스카이라인과 강 뷰. 균형 잡힌 친구 모임과 둥근 테이블에 앉음 — 다른 남자 최소 2명과 여자 2명, 모두 세련된 차림, 동등하게 함께. 평범한 친구 모임이지 한 남자를 여자들이 둘러싼 구도 절대 아님. 다 같이 웃으며 와인잔을 들어 건배, 켜지는 도시 불빛. 스카이라인 항공 리빌로 마무리. 오디오는 자연 앰비언트만: 발소리, 새, 멀리 차 소리, 스포츠카 엔진, 헬스 기구와 호흡, 잔 부딪는 소리, 잔잔한 대화, 강가 앰비언스. 음악·자막·로고·워터마크 없음. 네거티브: 화면 글자·자막·브랜드 로고·워터마크 금지. 얼굴·헤어 변경 금지. 인물 복제 금지. 드론·비행 기기 노출 금지. 헬스 씬 자켓 금지. 한 남자를 여자들만 둘러싼 구도 금지. 왁스·플라스틱 AI 피부 금지. 손가락 다섯 개. 신체 왜곡 금지.

THEME PARK DATE VLOG — 15sec Subjects: A Korean couple in their early 20s, filming their own date on a phone. THE WOMAN = FIRST attached photo. Match her exact eye shape, nose, jawline and hair. Do not substitute a different woman, do not age her up. THE MAN = SECOND attached photo. Match his face and hair exactly. Both faces stay identical in every cut. They wear the exact outfits from the attached photos in every scene. Look: real footage shot on a phone, not AI-generated. 26mm lens, deep focus, true skin texture with visible pores, natural flush, faint sensor noise at night, mild HDR, neutral iPhone color, unedited candid handheld, natural exposure clipping the sky. No cinematic grading, no studio lighting, no bloom. Every scene is inside a large theme park — never a road or rural area. (0:00–0:01) — Park Entrance Selfie from his outstretched arm, low angle, both faces fill frame. Ferris wheel and castle spires behind, sky blown white. She points up, he laughs. Dialogue (woman): "놀이공원 놀러왔어요!" Cut: hard on drop (0:01–0:02) — Ticket Gate Handheld from behind, bobbing with her steps. They push through the turnstile with ticket stubs, she skips a step ahead. Cut: fast on beat (0:02–0:03) — Carousel Both on painted horses under spinning bulbs, rising out of sync. She reaches for his hand and misses as the horses drift. Cut: on movement (0:03–0:05) — Roller Coaster Drop Tight two-shot from just in front of them, harnesses locked over their chests. They are dropping fast — the world behind them rushes upward past their shoulders in a vertical blur, their hair lifting off their scalps. She screams with her eyes squeezed shut, he howls with laughter at her into the wind. Hard vibration, bright sky clipping. Dialogue (woman): "야 이거 너무 빨라!" Cut: on her scream (0:05–0:07) — Claw Machine Fail Over-shoulder through the glass at a claw machine full of plush bears. The claw grips one, lifts it over the pile, drops it just before the chute. His shoulders slump, forehead on the glass, she covers her mouth laughing. Cool arcade LED. Cut: single beat (0:07–0:08) — Cotton Candy She holds a huge pink cotton candy by her face and tears a piece into his mouth. He nods big, eyes closed. Golden light rims her hair. Cut: snap (0:08–0:10) — Night Parade, from inside the crowd Chest-height handheld from behind their shoulders. They stand deep in a dense crowd, heads of strangers filling the foreground around them. Far beyond the crowd a huge illuminated float rolls slowly past with sequined dancers waving from its deck. A bear mascot near the float waves both paws their way — she rises on her toes, waves back grinning, and elbows him, who cracks up. Confetti drifts through the colored light. Cut: snap (0:10–0:12) — Ferris Wheel Cabin at Night ★ SIGNATURE Interior two-shot, camera never visible. They sit facing each other in the glass cabin, the park glowing below. She looks out, then turns to him, and they smile without speaking. Lights diffused, faces sharp. Speed: ~60% slow motion Cut: soft on her turn (0:12–0:13) — Fireworks Look-Up Both tilt their heads back, faces lit in bursts of gold and pink shifting on their skin. Small awed open-mouth smiles. Natural night exposure. Cut: sharp after burst (0:13–0:15) — Ending, Under the Ferris Wheel Close selfie held long from his outstretched arm, the lit Ferris wheel turning behind. She leans her head onto his shoulder and stays. He looks down at her with a soft smile. The frame holds as the wheel turns. Dialogue (man): "우리 내년에도 또 오자." Cut: fade to black Ambient audio only — crowd chatter, ride mechanics, wind, laughter. No on-screen text, subtitles, logos or watermarks. No face drifting between cuts. No visible phone, selfie stick or tripod in frame. No barriers, fences or railings in frame. Nothing moves in reverse. No waxy plastic AI skin, no beauty smoothing, no uncanny CG face in close-ups, no doll-like symmetry. Five fingers per hand, no malformed anatomy, no duplicate people.

KARAOKE ROOM VLOG — 15 SEC Subjects: Two Korean women in their early 20s at a Korean noraebang, close friends. WOMAN A — pale yellow collared knit, cream balloon mini skirt, low ponytail, bright and loud. WOMAN B — sheer plum cardigan over a dark slip dress with chiffon hem, low bun, quieter until she sings. Both faces stay identical in every cut, each in her own outfit throughout. Shoes kicked off at the door — barefoot the whole video, including on the sofa. Look: real footage, not AI-generated. 26mm lens, deep focus, real skin texture with pores and sweat shine, natural flush, sensor noise in dark corners, neutral iPhone color, unedited handheld. Neon strips and a disco ball throw shifting pink, blue and orange across the walls, but their faces stay clearly lit and never look CG. No cinematic grading, no studio lighting, no bloom, no slow motion. Room: one cramped windowless karaoke room barely bigger than the sofa — walls a couple of steps away on every side, always visible behind them. Bench sofa, low table in front, mirrored wall, two mics, a tambourine. On the table: two smartphones face-up, a water bottle, a bowl of snacks. Nothing else — no alcohol, no soju or beer. Every screen shows ENGLISH only — song title, lyric lines, music video. No Hangul or CJK anywhere in the room. (0:00-0:01) — Bursting In Handheld from inside as the door swings open and the two pile in, A first, laughing. Dialogue (A): "We're singing till our voices die!" Cut: hard on the door (0:01-0:02) — Picking a Song Over-shoulder on the remote in A's hands, thumb stabbing buttons while B leans in pointing at the English titles. Cut: on beat (0:02-0:04) — A Belts It A stands barefoot on the sofa, mic in both hands, head thrown back mid-note, hair swinging, low ceiling above her. B below shakes a tambourine hard. Camera bounces. Cut: on the tambourine (0:04-0:05.5) — B Takes the Mic Tight on B on the sofa, mic held close in both hands, eyes shut, brows drawn, a slow ballad. A's hand enters frame swaying a lit smartphone like a lighter. Cut: push in slightly (0:05.5-0:07) — B Goes Big The chorus hits — B rises off the sofa mid-note, one arm flung out, head tipped back, voice cracking on the high note. A doubles over laughing and claps. Camera jolts back to hold both. Cut: on the last note (0:07-0:09) — The Score Both freeze staring at the wall screen, which flashes the score in huge digits: 98. B throws her arms up screaming, A collapses face-down. Dialogue (B): "Ninety-eight?! No way!" Cut: on her collapse (0:09-0:11) — Duet Chaos Wide two-shot that already fills the little room, both barefoot up on the sofa, sharing one mic, bouncing completely off the beat. The mirrored wall doubles the chaos. Cut: on a bounce (0:11-0:13) — Catching Their Breath They drop back down side by side, flushed and panting, passing the water bottle, phones still on the table. A nudges B, their eyes meet and they start giggling. Cut: soft on the laugh (0:13-0:15) — Handing Over the Mic A pushes her mic straight at the lens with a challenging grin, B leaning in laughing and pointing at the camera. Both sweaty and wrecked, waiting for an answer. Dialogue (A): "So? You're not gonna sing?!" The shot ends on a live lit frame. Ambient audio only — backing track, real singing, screaming, tambourine, laughter. No subtitles, logos or watermarks. No camera rig, selfie stick or tripod in frame — the table phones are props, never a filming device. No face drifting between cuts. No fade to black. No alcohol. No Korean or Chinese characters anywhere — text is English only, and the only number on the score screen is 98. No shoes on the sofa. No wide empty hall or stage. No waxy plastic AI skin, no beauty smoothing, no uncanny CG face, no doll-like symmetry. Five fingers per hand, no malformed anatomy, no duplicates.

Photorealistic behind-the-scenes smartphone vlog footage of a young Korean girl content creator. Faux vertical smartphone footage naturally cropped into 16:9. Authentic handheld phone aesthetic. No cinematic color grading, no HDR, no beauty filters. Feels like genuine casual vlog footage. CHARACTER — JIYU: Korean girl, age 20, extremely cute rabbit-like face — big round eyes with under-eye aegyo-sal, slightly chubby cheeks, soft coral lips, natural Korean features and real skin texture (no doll-like AI smoothness). Bangs with shoulder-length wavy hair (soft light brown / ash brown). Petite with curvy build. Cozy grey zip-up hoodie loungewear set (zipper open, white inner peeking). Maintain the exact same face, hair, body and outfit identity in every scene — never morph into a different person. SETTING: A simple, slightly messy small studio apartment. Clothes on the chair, laptop and ring light on the desk, empty coffee cups, snack wrappers, tangled cables, cute posters on the wall. Natural, lived-in feeling. LIGHTING: Starts with soft morning light through the window, gradually transitions into warm afternoon and evening indoor lighting. Natural smartphone exposure changes. CAMERA: Simulated smartphone footage with gentle handheld movement, natural walking, subtle autofocus breathing, occasional imperfect framing, realistic phone compression. Feels like a friend casually filming. STYLE & MOOD: Casual, relatable, a bit funny. Chill but passionate young creator — sometimes excited, sometimes chaotic, always sincere. All dialogue in natural Korean, casual and unscripted. SCENE SEQUENCE (15s): CUT 1 (0~2.5s) — Waking Up. Morning. JIYU slowly wakes up in bed, messy hair, squints at the light, dramatically flops back onto the pillow, then forces herself up with a lazy stretch. 대사: "아... 5분만 더..." CUT 2 (2.5~5.5s) — TikTok Live. She props her phone on a ring light, sits cross-legged, starts a live stream — waving brightly, reading comments, laughing, playful hand gestures. 대사: "안녕 얘들아~ 오늘도 왔어!" CUT 3 (5.5~7.5s) — Content Chaos. At the desk editing, too many tabs open, recording a voiceover, slightly overwhelmed. She knocks over a cup while celebrating a good take. 대사: "아 이거 왜 안 되지... 어? 오케이 됐다!" CUT 4 (7.5~10s) — Tteokbokki Break. She eats tteokbokki while scrolling her phone, laughs at a meme, then suddenly gets an idea and rushes back to the desk. 대사: "음~ 진짜 맛있다. 어 이거다!" CUT 5 (10~12.5s) — Shower & Blow-dry. Fresh out of the shower, towel around her shoulders, she blow-dries her wavy hair in front of a mirror, humming, glancing casually at the camera. 대사: "아 오늘 하루 진짜 길었다~" CUT 6 (12.5~13.5s) — Wind Down. Night. Tired but satisfied, she half-heartedly tidies up, sits on the floor with a drink, scrolling her phone. 대사: "오늘도 수고했어, 나." CUT 7 (13.5~15s) — Phone Call Ending. She's on a phone call, smiling and giggling softly. Suddenly noticing the camera, she blurts out and quickly covers the lens with her hand — screen goes dark, end. 대사: "어 남자친구 아니에요! ㅋㅋ" AUDIO: Natural room ambience only. Keyboard typing, phone notifications, hairdryer, tteokbokki being eaten, occasional sighs and small laughs, fan humming. Her casual Korean muttering to herself. No background music. OVERALL VIBE: Funny, relatable, chill — a genuine day-in-the-life vlog of a young Korean creator who's passionate but a little messy and human. CONSISTENCY: Keep JIYU's exact face, hair, and outfit identical across all cuts — never change into a different person. Natural realistic hands and fingers. No on-screen text, subtitles, logos, or watermarks.

Seedance Video Prompt - Epic Wuxia, Deadpan Comedy 1. Visual Style & Core Direction Cinematic realistic ancient Chinese wuxia fantasy. High-end historical Chinese fantasy production design, natural skin texture, realistic fabric and hair movement, detailed ancient architecture, misty mountain atmosphere, dramatic cinematic lighting, subtle volumetric fog. The comedy is deadpan and understated, not slapstick. Characters must behave as if they are inside a serious wuxia movie, while the dialogue creates the absurdity. Tone: epic, solemn, restrained, awkward, unexpectedly funny. Use realistic cinematic acting, subtle facial expressions, natural body language, controlled camera movement, shallow depth of field, realistic motion blur. Important: Do not make the characters cartoonish or overly expressive. The humor comes from the contrast between the epic situation and the completely mundane conversation. 2. Character Consistency Senior Sister - Character A Calm, elegant and dignified female martial artist. Extremely polite and composed. She has the aura of a powerful senior disciple, but she is completely clueless about what the enemy leader actually said. Her default expression is serious and respectful. Junior Sister - Character B Younger female disciple, intelligent and quick-witted. She immediately understands the absurd situation. She tries very hard not to laugh. Her reactions should be subtle: confused eyes, suppressed smile, slight pause, awkward glance. Enemy Leader Intimidating male martial arts leader in elaborate ancient Chinese robes and armor. He begins with absolute confidence and delivers an extremely serious challenge at high speed. After realizing that nobody understood him, his confidence gradually disappears. Master Older, respected martial arts master. Extremely calm and authoritative. He observes the situation like a judge and delivers the final line with complete seriousness, making the punchline even funnier. 3. Narrative Logic The entire scene should feel like an epic martial arts confrontation. Expectation: A terrifying enemy leader arrives and gives a long, solemn challenge. Reality: The Senior Sister does not understand the final part because he spoke too quickly. Instead of admitting she did not understand, she simply nods because she feels obligated to respond after listening for so long. The Junior Sister realizes what happened. The enemy leader becomes completely confused. The Master gives the final deadpan verdict. The comedy should escalate naturally: Epic confrontation → polite misunderstanding → awkward realization → authority delivers absurd conclusion. 4. Shot-by-Shot Direction 0-3 seconds - Epic Establishing Shot Wide cinematic shot of an ancient Chinese mountain courtyard surrounded by mist and towering cliffs. The Enemy Leader stands several meters away from the two sisters, surrounded by his disciples. Wind moves their robes and hair. The atmosphere feels extremely serious and dangerous. Slow cinematic push-in. Enemy Leader steps forward and begins delivering a powerful martial arts challenge at extremely high speed. His speech is solemn, intense and theatrical. 3-5 seconds - The Senior's Response Medium close-up of Senior Sister. She listens with an absolutely serious expression. She slowly nods several times as if she fully understands everything. The Enemy Leader finishes his long speech with a dramatic final sentence. Senior Sister remains completely calm. She gives another respectful nod. Tiny pause. No music cue for comedy. Let the awkward silence create the joke. 5-7 seconds - The Whisper Cut to a tight two-shot of the two sisters. Senior Sister subtly leans toward Junior Sister and whispers: Senior Sister: "What did he say? It was too fast at the end." Junior Sister freezes. She looks at Senior Sister. Then slowly looks toward the Enemy Leader. Her expression says: "You just nodded like you understood everything." Keep the reaction subtle and realistic. 7-10 seconds - The Absurd Explanation Senior Sister quietly explains, still completely serious: Senior Sister: "He talked for so long... I had to give some response." Junior Sister tries desperately not to laugh. She looks away. The Enemy Leader is still standing in the background. He has overheard enough to realize what happened. His confident expression slowly collapses. He looks genuinely speechless. Hold the awkward silence for a beat. 10-12 seconds - Enemy Leader Reaction Close-up of the Enemy Leader. His intimidating expression disappears. He blinks once. Then again. He looks toward Senior Sister in disbelief. His face communicates: "Did she seriously not understand a single word?" Do not exaggerate the reaction. Keep it deadpan and realistic. 12-15 seconds - Master's Final Verdict Cut to the Master standing calmly behind the sisters. He has been silently observing the entire conversation. He looks at the Enemy Leader. Then calmly says: Master: "At least she was very polite." Cut to the Enemy Leader's completely defeated expression. Hold for a brief awkward silence. End on the Enemy Leader standing speechless while the Senior Sister remains perfectly dignified. 5. Dialogue & Performance Rules Dialogue must sound natural in spoken Mandarin Chinese if voice generation is supported. The Enemy Leader speaks rapidly during his opening challenge. Senior Sister speaks slowly, politely and calmly. Junior Sister reacts with restrained disbelief. Master speaks slowly with absolute authority. Comedy timing is essential: Long serious challenge Serious nod Brief silence Whispered misunderstanding Awkward realization Final deadpan verdict Do not rush the punchlines. Use natural pauses between dialogue lines. 6. Camera & Cinematography Cinematic 24fps look. Realistic camera movement. 35mm lens for establishing shots. 50mm lens for dialogue scenes. 85mm lens for reaction close-ups. Shallow depth of field during close-ups. Natural cinematic motion blur. Subtle handheld movement only when appropriate. Use slow push-ins and controlled cuts. No rapid editing. No exaggerated zooms. No sitcom-style camera movement. 7. Sound Design Ancient Chinese fantasy atmosphere. Wind through the mountain courtyard. Subtle movement of robes and weapons. Distant birds and environmental ambience. The opening challenge should feel powerful and dramatic. After Senior Sister asks what he said, reduce the background sound slightly. Allow a short moment of silence before each punchline. The final Master line should be delivered completely seriously. No laugh track. No comedy music. No exaggerated sound effects. The audience should discover the joke through the characters' reactions. 8. Visual Quality & Negative Prompt Ultra-realistic cinematic Chinese wuxia fantasy, premium film production quality, realistic human anatomy, realistic facial expressions, detailed skin pores and natural skin imperfections, realistic hair strands, physically accurate cloth simulation, natural lighting, cinematic depth of field, atmospheric perspective, volumetric mist, detailed ancient Chinese costumes and architecture. Avoid: cartoon style, anime, exaggerated facial expressions, slapstick acting, goofy movements, modern clothing, modern objects, inconsistent faces, changing character appearance, extra fingers, distorted hands, unnatural body proportions, floating objects, excessive camera shake, excessive slow motion, random subtitles, text overlays, watermarks, logos, laugh track, comedy music, overacting. Critical consistency rule: Keep the same facial identity, hairstyle, costume, body proportions and color palette for each character throughout the entire video. The Senior Sister must remain dignified and serious even when delivering the absurd explanation. The Master must remain completely calm. The Enemy Leader's loss of confidence should be subtle and progressive.

基础设定：基于这张 {{Mixed 2}} 的色卡，文字等信息和人物角色图 {{Mixed 1}} ，以及下面的画面内容，生成一段充满日式电影美学与叙事感的短片。仅生成下面15秒的画面内容。无背景音乐风格要求： 整体必须是正宗的“日式青春文艺电影”视觉风格。画面排斥任何高饱和、高对比度的商业塑料感。采用“Natural Light & Airiness”自然光与空气感布光，高光处带有温暖的过曝柔化边缘。通过“极致纯净的环境白噪音”与“极致克制的电影慢镜头”形成极具沉浸感的静谧艺术张力画面内容：Scene 06 (30.0-35.0s) [奔向碧海]: 镜头一采用 24mm 超广角，远景。两人推着自行车来到海边的堤坝。镜头二采用手持跟拍 (Handheld Tracking) 贴地拍摄，捕捉他们冲向沙滩时的背影和飞扬的裙摆，背景是湛蓝、清透得不真实的晴空。 Scene 07 (35.0-41.0s) [浪花与秘密]: 镜头一展示两人在海边堤坝，并排坐下。女生拨弄被风吹乱的头发。镜头二采用缓慢推轨 (Dolly-in)，采用低角度逆光仰拍，阳光穿过他们举起的手臂，在脸上形成柔和的镜头光晕 (Lens Flare)，两人害羞地相视一笑。 Scene 08 (41.0-45.0s) [夏天的风]: 镜头一采用 35mm 焦段，特写喝了一半的蓝紫色塑料瓶被放在沙滩上，旁边滚落着一颗脏篮球，一阵风吹过，扬起细微的沙粒。镜头二画面渐隐至纯白，中心浮现极其纤细、带有易碎感的明朝体字体：“夏天的风”，排版文艺诗意。【日语旁白 (VO-Japanese)】: 真夏の余熱、一瞬の風が吹く。

スタイル: 実写映画。3倍速のハイスピードMV。昭和レトロな日本の銭湯、妖怪、からくり人形、和風サイケデリックを融合したダークファンタジー映像。深紅・紺青・金・黒を基調に、古い木造建築、タイル、湯気、和傘、扇子、仮面、巨大な眼球などを反復モチーフとして使用する。舞台セットは精巧な人形劇や立体切り絵のような質感。高速ズーム、急激なパン、360度旋回、画面分割、万華鏡、増殖、残像、トンネル表現を多用する。画像1・画像2・画像3の人物は、それぞれ参照画像の顔、髪型、衣装、体格、アクセサリー、配色などの特徴を維持し、全シーンで同一人物として統一する。人物の衣装や外見をシーンに合わせて勝手に変更しない。複製・増殖する場合も元となる参照人物のデザインを維持する。 BGM: なし。参照音源の歌声・演奏のみ。ギター、マイク、鐘、湯が流れる音、蒸気、木材の軋み、妖怪の声、からくり装置の作動音などの環境音・効果音は入れる。シーン1：妖怪銭湯ライブ古い巨大な和風銭湯の入口前に画像1・画像2・画像3の人物が並ぶ。中央で画像1の人物がヴィンテージマイクへ向かって歌う。左右では画像2の人物と画像3の人物が楽器を激しく演奏する。カメラは銭湯の巨大な建物全体を映すワイドショットから開始。そのまま床すれすれを高速で前進し、3人の中央へ突入する。画像1の人物の顔をクローズアップした直後、画像2、画像3の人物へ高速パンする。背後の煙突や銭湯から大量の白い湯気が噴き出す。シーン2：増殖する人物と鐘画像1の人物がマイクの前で歌いながら、片手に古い金属製の鐘を持つ。鐘を鳴らした瞬間、背後に画像1の人物と同じ顔を持つ人形や仮面が円形に次々と出現する。カメラは鐘の超クローズアップから開始。鐘が鳴る瞬間に高速ズームアウト。画像1の人物を中心として、同じ顔が曼荼羅状に無数に配置されていることを見せる。背景に並ぶ顔は歌に合わせて一斉に口を開閉する。カメラが画像1の人物の周囲を360度高速旋回。顔の輪が何重にも増殖し、巨大な万華鏡へ変化する。シーン3：画像2の人物と妖怪の湯画像2の人物が銭湯内部で激しく楽器を演奏する。周囲の浴槽から大量の蒸気が立ち上る。その中から巨大な蛇、奇妙な魚、一つ目の生物などの妖怪が次々と現れる。複数の不気味な仮面も空中へ浮かび上がり、画像2の人物の周囲を漂う。カメラは画像2の人物の手元の超クローズアップから開始。楽器に沿って高速移動し、そのまま画像2の人物の顔へ急接近する。直後にカメラが高速で後退。背後の湯気の中から巨大な蛇の妖怪がレンズへ向かって飛び出してくる。シーン4：画像3の人物とからくり地獄画像3の人物が激しく楽器を演奏する。周囲には巨大な妖怪の頭部、炎のような装飾、木製のからくり装置、巨大な牙を持つ奇妙な機械が配置されている。画像3の人物が強く演奏するたびに、床のからくり装置が音楽と連動して動き始める。巨大な牙が勢いよく開閉し、巨大な妖怪の頭部が咆哮する。カメラは画像3の人物の足元から顔まで一気にチルトアップ。そのまま画像3の人物の周囲を高速旋回する。最後は背後に現れた巨大な妖怪の口の内部へ突入する。シーン5：回転するからくりステージ画像1・画像2・画像3の人物が巨大な円形のからくりステージ上で演奏する。中央に画像1の人物。左右に画像2と画像3の人物。周囲には楽器、和傘、扇子、銭湯の桶、仮面、巨大な眼球が巨大な機械部品のように配置されている。演奏が始まると円形ステージ全体が高速回転する。カメラは真上からステージ全体を俯瞰。そこから螺旋状に降下しながら3人へ高速接近する。回転速度が徐々に上昇し、背景が赤・青・金の円形残像へ変化する。シーン6：千手の人物画像1の人物がマイクの前で歌う。突然、画像1の人物の背後から、画像1の衣装デザインをそのまま反映した大量の腕が放射状に出現する。それぞれの手には和風の扇子が握られている。何十本もの腕が歌のリズムに合わせて一斉に動き、扇子を開閉する。左右では画像2と画像3の人物が激しく演奏する。画像1の人物を中心にカメラが高速で360度旋回。旋回するたびに腕の数が増えていき、最終的に巨大な千手観音のようなシルエットになる。シーン7：サブリミナル・妖怪コラージュ画面が突然、三角形や菱形の細かなパネルへ分割される。画像1の人物の瞳。画像1の人物の口元。画像1の人物の衣装。ヴィンテージマイク。画像2の人物の顔。画像2の人物の手元。画像3の人物の目。画像3の人物の楽器。扇子。鐘。巨大な眼球。蛇。妖怪の口。仮面。これらの超クローズアップ映像を0.1〜0.3秒単位で高速に切り替える。パネル自体も回転、拡大、縮小し続ける。一部の顔や眼球だけが逆方向へ動き、画面全体が巨大な万華鏡のように変形する。画像1・画像2・画像3の人物の顔が交互に一瞬だけフラッシュする。最後は中央に出現した巨大な眼球へすべての映像が吸い込まれる。シーン8：妖怪銭湯大宴会巨大な銭湯の中央で画像1・画像2・画像3の人物がライブ演奏する。周囲を一つ目妖怪、鬼、蛇、魚、提灯のような生物、奇妙な人形たちが埋め尽くす。浴槽から大量の青白い湯気が噴き上がる。建物の正面には巨大な眼球が現れ、演奏する3人を見つめている。カメラは妖怪たちの足元を縫うように高速移動して3人へ接近。画像1、画像2、画像3の人物を連続してクローズアップする。その後、一気に上空へ上昇して銭湯全体を俯瞰する。妖怪たちが3人を中心に円を描いて踊り始め、床の模様まで巨大な渦へ変化する。シーン9：無限に続くからくり箱画像1・画像2・画像3の人物の前に、古い和風の巨大なからくり箱が現れる。箱の正面がゆっくり開く。内部には、画像1・画像2・画像3と同じ人物が小さなステージで演奏している。カメラが箱の内部へ高速ズーム。さらに小さなステージにも同じからくり箱が存在する。その箱の内部にも再び画像1・画像2・画像3の人物がいる。同じ3人、銭湯、妖怪、からくり箱が何重にも連続する無限トンネルになる。カメラはその世界を高速で突き抜け続ける。最後は最深部に現れた画像1の人物の瞳へ急激にズームイン。瞳の中に最初の銭湯ライブを行う画像1・画像2・画像3の人物が映り込み、そのまま暗転する。