The Girl and the Paper Boat

shot: type: R2V reference: <Picture 1> camera: fixed, eye-level, medium-wide shot capturing full body throughout, no cuts duration: 15-second continuous sequence, constant tempo throughout character: appearance: delegated entirely to <Picture 1>, no additional description style_separation: character: cel-shaded anime style, locked to reference image environment: photorealistic indoor dance studio floor, neutral wall, soft even lighting action: sequence_name: thomas_flare_to_windmill_combo segments: - segment: intro_setup duration: approx 0-3s description: standing neutral pose transitions down into crouched base position, both palms grounding shoulder-width apart, core braced, hips settling to a fixed rotation height - segment: flare_combo duration: approx 3-9s support_pattern: alternating two-hand support, hand-to-hand transfer, no static single-hand hold description: two continuous flare rotations - right leg sweeps laterally at constant angular velocity, left leg sweeps diagonally downward-forward as support hand transfers, right leg continues diagonally upward with legs held widely split throughout, hips remain the stationary rotation axis with no vertical bounce; cycle repeats once more at identical speed - segment: transition_to_windmill duration: approx 9-11s description: on the final flare rotation, upper back and shoulder lower to the floor while legs continue their circular momentum, torso rolling from the flare's hip-axis rotation into a back-and-shoulder-axis rotation, hands lift off the floor and tuck near the chest as the transition completes - segment: windmill_execution duration: approx 11-15s description: two continuous windmill rotations - legs swing in a wide circular arc overhead while the upper back and shoulders roll across the floor in constant-velocity rotation, legs stay extended and split for momentum, no hands touching the floor for support during this segment physics_constraints: - hip axis remains at constant height during the flare segment, no vertical bounce - leg sweeps driven by hip rotation and knee flexion, not by jumping or floating - hand-to-floor contact during the flare is continuous except the brief instant of transfer between palms - the flare-to-windmill transition is a smooth momentum carry-over, not a stop-and-restart - during the windmill segment, rotation is driven by shoulder-and-back contact with the floor, not by pushing off with hands - rotation speed stays constant within each segment, no acceleration bursts - legs stay maximally extended and split during all rotation segments to convey centrifugal force lighting: soft studio lighting, consistent throughout, no flicker

Begin in a DARK, GRITTY, PHOTOREALISTIC modern city street at night. The opening MUST look completely live-action and realistic — wet pavement reflecting streetlights and neon signs, dark buildings, passing cars, and several pedestrians walking along the sidewalk. The people look tired, serious, and unhappy. Moody cinematic nighttime lighting, realistic human faces, realistic clothing, realistic city environment. [0.0–1.2s] Fast cinematic establishing shot moving through the gloomy photorealistic street. Pedestrians walk silently with tired expressions. Everything feels cold, dreary, and lifeless. [1.2–2.0s] QUICK CUT to a close-up of an old-school portable BOOMBOX sitting on the sidewalk. A human hand reaches into frame and forcefully presses the PLAY button. CLICK. Immediately an energetic, joyful, extremely catchy upbeat song EXPLODES from the speakers. [2.0–4.8s] The boombox speakers suddenly release an enormous, spectacular TIDAL WAVE OF MUSIC. CRITICAL: This is NOT a water wave. It is a gigantic colorful 2D HAND-DRAWN CARTOON WAVE constructed entirely from hundreds of animated musical notes, treble clefs, music staffs, colorful sound lines, rhythm symbols and swirling music graphics. The musical-note wave moves physically like a huge breaking ocean wave, rapidly surging outward from the boombox and racing down the street. The wave itself is already CARTOON ANIMATION while the untouched city ahead of it remains completely PHOTOREALISTIC. As the musical wave touches anything, that object INSTANTLY TRANSFORMS from realistic live-action into colorful hand-drawn cartoon animation. The transformation boundary travels visibly with the wave. It sweeps across pedestrians: REALISTIC HUMAN → COLORFUL CARTOON CHARACTER. Their tired expressions immediately become huge joyful smiles. They begin dancing, bouncing and moving happily to the music. The wave hits realistic cars: REAL CAR → FUN COLORFUL CARTOON CAR. It hits buildings, sidewalks, streetlights and signs: REALISTIC CITY → BRIGHT PLAYFUL CARTOON CITY. Cartoon streetlights bend and bounce rhythmically. Cartoon signs wiggle happily. Cars bounce slightly to the beat. The camera dynamically follows the rapidly expanding transformation. [4.8–7.0s] The camera quickly CRANES UP above the street into a wide aerial view. The enormous musical-note wave races outward block after block across the city. Behind the wave is an entirely colorful, joyful HAND-DRAWN CARTOON WORLD. Ahead of the wave remains dark photorealistic nighttime reality. As the final portion of the city transforms, the transformation races upward into the sky. DARK REALISTIC NIGHT → BRIGHT BLUE CARTOON DAYTIME. A huge cheerful YELLOW CARTOON SUN rises behind the newly transformed cartoon skyline. The cartoon sun opens its eyes and forms a gigantic HAPPY SMILE. FINAL IMAGE: A completely transformed colorful cartoon city filled with smiling cartoon people dancing in the streets, bouncing buildings and cars, floating musical notes everywhere, bright blue sky, and the enormous smiling cartoon sun above the skyline. The upbeat music continues triumphantly through the final frame. CRITICAL TRANSFORMATION RULE: Everything begins PHOTOREALISTIC LIVE ACTION. ONLY the musical-note wave begins as cartoon animation. Nothing transforms until the musical wave physically touches it. The cartoon transformation must visibly spread outward with the moving wave. By the final frame, EVERYTHING is colorful hand-drawn cartoon animation. NO dialogue. NO text overlays. NO subtitles. NO water in the musical wave.

Cinematic action scene set in a post-apocalyptic destroyed city during sunset. Two young East Asian female heroines—one with long vibrant bright pink hair in colorful streetwear with yellow gloves, and one with a sharp black bob hairstyle wearing a tight black latex superhero suit—run side-by-side down a cracked, ruined highway. They leap high into the air to battle a massive, menacing black alien monster with glowing red eyes, sharp bio-armor, and dark tendrils. Fast-paced dynamic cuts: sweeping tracking shot, close-up of a high-kick land on the monster's chest, high-angle leap into a smoke cloud, landing together in a low three-point superhero pose looking intense and out of breath. Ends with a extreme close-up of the black alien monster grinning evilly with glowing red eyes and lightning crackling across its face. Photorealistic, 8k resolution, dramatic volumetric lighting, cinematic VFX, Unreal Engine 5 render style.

# 動画仕様 [Image1]＝CARD Aのカード・人物・能力の固定参照。 [Image2]＝CARD Bのカード・人物・能力の固定参照。 [Image3]＝各ラウンドの勝敗、最終スコア、最終勝者を示す完成済み結果画像。 15秒、16:9、720p、24fps。高品質な現代日本アニメ、プレミアムなカードゲーム召喚バトルPV。映画的照明、ホログラムカード、固有色の粒子、滑らかなカメラ。爽快で上品な非流血アクション。 Image1とImage2から顔、髪、瞳、年齢感、体格、種族、衣装、装備、武器、翼、アクセサリー、固有色、移動・攻撃・防御方法、固有スキル、弱点、持久力を読み取る。Image3から左右配置、ROUND 1〜4の勝者、最終スコア、最終勝者を読み取る。人物デザインはImage1とImage2だけを使用。Image3内の人物で顔や衣装を上書きしない。 # 最重要方針 0.0〜3.0秒はImage1とImage2を完成済み2Dカードとして表示。12.2〜15.0秒はImage3を完成済み2D結果パネルとして表示する。カードと結果画像を描き直さず、名前、数値、レアリティ、スキル、勝敗を再生成・変更しない。戦闘中は能力値、勝者名、説明文を表示しない。新規文字は「VS」「ROUND 1」「ROUND 2」「ROUND 3」「ROUND 4」のみ。最初のフレームからCARD Aを左、CARD Bを右へ同じ大きさで同時表示。片方を先に出さない。両者は3.0秒ちょうどに同時召喚され、それ以前に戦闘しない。戦闘中もAは左、Bは右を基本とし、左右を反転させない。 Image3の各ラウンドの勝者、スコア、最終勝者を厳守。レアリティや総合値だけで勝敗を変えない。 # 舞台宇宙、神殿、ホログラムを融合した円形召喚アリーナ。光沢のある黒い床、中央に細い金色線。左側はA、右側はBの固有色で照らす。属性表現は参照カードに存在するものだけ。全ラウンドで同じ舞台を維持する。 # タイムライン ## 0.0〜1.5秒：カード同時表示最初からImage1を左、Image2を右へ同時表示。2枚は同じ大きさと高さで正面を向き、カード全体を収める。各カードから設定に合う固有色の光と粒子が流れる。存在しない属性を追加しない。中央に金色の「VS」。2枚を残したまま緩やかにドリーインする。 ## 1.5〜3.0秒：同時召喚両カードの前へ、それぞれの固有色の召喚陣が出現。人物の輪郭が立体的に浮かび、2.3秒で同時に一歩踏み出す。2.8秒からカードが光へ変わり、3.0秒ちょうどに両者が完全実体化して戦闘開始。この区間に能力比較表、スコア、結果画像を出さない。 ## 3.0〜5.0秒：ROUND 1・機動戦上部に小さく「ROUND 1」。24mm低空トラッキング。両者がカード設定に合う走行、踏み込み、跳躍、飛行、転移、推進、幻影などで中央の有利な位置を争う。Image3でROUND 1を取った側が、予測、敏捷性、位置取りによって最後に有利な場所を確保する。敗者側も一度は先手を取り、能力の魅力を見せる。勝敗は直撃ではなく位置取りで表現。 ## 5.0〜7.0秒：ROUND 2・攻撃戦上部に「ROUND 2」。35mmのサイド寄り全身構図。両者は参照カードにある攻撃方法だけを使用。武器は同じ種類・本数を維持し、素手の人物に武器を追加しない。魔法や属性も参照設定に限定する。 Image3の勝者が技術、攻撃力、戦術、射程、受け流しで優勢になる。敗者側も美しい攻撃を一度完成させる。身体への直撃や負傷ではなく、武器同士、光、風圧、結界、回避で表現。 ## 7.0〜9.0秒：ROUND 3・特殊戦上部に「ROUND 3」。両者を収めて約80度オービット。両者が参照カード記載の固有スキルを発動。加速、連続攻撃、予測、カウンター、固有属性、結界、回復、幻影、召喚など、設定された能力だけを映像化する。 Image3の勝者が相手の技を見切る、上回る、回避する、受け流す、無効化する、回復して立て直す形で優勢になる。敗者側のスキルも省略せず美しく発動させる。 ## 9.0〜11.0秒：ROUND 4・持久戦上部に「ROUND 4」。細かく切らず、両者を同時に見せる全身ショット。持久力の低い側は呼吸を整え、光量や動きがわずかに低下。高い側は姿勢、結界、翼、装甲、武器、呼吸を安定して維持する。汗、流血、転倒では表現しない。最後の攻防で、Image3の勝者が防御、持久力、回復、冷静さによって制する。敗者は倒れず、自ら攻撃を止めて静かに構えを解く。 ## 11.0〜12.2秒：決着カメラが引いて両者の全身を映す。勝者は武器を収め、翼や結界を解除して穏やかに構えを解く。敗者は立ったまま呼吸を整え、相手へ敬意を示す。誇張した勝利ポーズや侮辱なし。両者の固有色の粒子が中央へ集まり、白い光の平面になる。 ## 12.2〜15.0秒：結果表示白い光からImage3が出現。Image3を再描画・再組版せず、完成済み2D画像としてそのまま16:9で表示する。カード、人物、名前、数値、レアリティ、ラウンド結果、スコア、勝者、左右配置を変更しない。新しい文字を重ねない。周囲に細かな固有色粒子だけを追加し、文字を隠さない。 13.0秒までに完全表示し、15.0秒までカメラを固定。ズーム、パン、回転、フェードアウトなし。 # カメラ 0.0〜3.0秒：安定した正面ドリーイン。 3.0〜5.0秒：低空横方向トラッキング。 5.0〜7.0秒：サイド寄り全身構図。 7.0〜9.0秒：約80度のオービット。 9.0〜11.0秒：固定に近い全身ツーショット。 11.0〜12.2秒：左右を維持したワイド。 12.2〜15.0秒：Image3固定。 180度ルールを維持し、高速回転、上下反転、激しい手ぶれ、過剰なズームを使わない。 # 一貫性・禁止事項顔、髪、瞳、年齢感、体格、身長差、種族、衣装、装備、武器、翼、尻尾、アクセサリー、固有色、戦闘スタイルを固定する。各キャラクターの実体は一人だけ。残像や幻影は半透明の単色光とし、顔、肌、衣装、武器を描かない。人物を融合させず、顔や衣装を交換しない。手足、指、翼、尻尾を増減させない。身体と武器を貫通させない。存在しない武器、属性、翼、魔法、召喚物を追加しない。勝敗、スコア、ラウンド結果を変更しない。文字化けした情報パネル、指定外の文字、技名、能力値を表示しない。流血、切断、死亡、重傷、衣装破損、敗者への侮辱、過度な点滅、既存ゲームロゴ、字幕、透かし、署名、黒枠を表示しない。 # 音響実在曲ではないオリジナルBGM。壮麗なオーケストラ、電子音、低い打楽器、疾走感あるストリングス。 0秒はカードと召喚音、3秒でビート開始。ROUND 1は移動音、ROUND 2は武器・風圧・受け流し、ROUND 3は固有スキル音、ROUND 4は呼吸と防御音を強調。12.2秒に上品な勝敗決定音。セリフ、歌、ナレーションなし。

Animate the character walking naturally forward with a stable, realistic walk cycle. Preserve the exact same face, body proportions, hairstyle, costume, colors and accessories throughout the entire shot. Her arms swing gently in coordination with her steps, her clothing and hair react naturally to body movement and air resistance, and each foot makes firm contact with the ground without sliding. Maintain consistent anatomy, five fingers on each hand, stable facial features and correct costume details. No character redesign, morphing, duplicated limbs, costume changes or sudden scale changes.

Cinematic anime short film clip, 15 seconds. Outdoor ancient Thai temple, stone walls and pillars, thick jungle beyond, natural daylight through jungle canopy, heavy bag hanging from stone archway, dust particles in the air. CHARACTER: Use uploaded character sheet as strict visual reference. Female Muay Thai fighter, athletic powerful build, long dark hair tied back in tight bun. Deep red Muay Thai shorts with gold trim, red gloves, red hand wraps, red ankle wraps, bare feet, gold Mongkon headband, toned bare midriff. AUDIO: High energy driving beat loud from frame one never stopping, every single strike synced to the beat, glove elbow knee impact sounds punching through on every hit, bag chain rattling constant, never fading never dropping. 0:00–0:01 — Already throwing, bag already swinging. Wide shot full temple. Cut to low angle. Cut to close on gloves. Three angles in one second. 0:01–0:03 — Jab cross hook combination. Close on left jab landing. Cut to side profile full body. Cut to close on her eyes. Cut to front shot gloves flying. Cut to close on bag impact. 0:03–0:05 — Elbow strikes. Close on right elbow connecting. Cut to overhead looking down. Cut to low angle looking up. Cut to close on gold Mongkon. Cut to wide shot through stone pillars. 0:05–0:07 — Knee strikes, grabs the bag. Close on knee driving in. Cut to side profile full body. Cut to close on her fierce face. Cut to overhead. Cut to close on bare feet pivoting on stone floor. 0:07–0:09 — Rapid punch and elbow combo. Front shot gloves blurring. Cut to behind her looking toward temple. Cut to close on glove connecting. Cut to low angle. Cut to close on bag swinging violently. 0:09–0:11 — High kick. Low angle looking up at the kick. Cut to side profile full kick. Cut to close on foot connecting. Cut to close on her face on landing. Cut to wide shot full body full temple. 0:11–0:13 — Full combination everything — punch elbow knee kick all in sequence. Camera never stops cutting — every single strike different angle, different distance, different height. Gold trim flashing every frame. 0:13–0:15 — Final devastating roundhouse kick. Low angle slow for one frame then lands. Bag swings to absolute maximum. She snaps back into stance. Close on eyes locked forward. Gold Mongkon gleaming. Hero moment. Done. Cinematic cel-shaded 3D anime, semi-realistic CGI, hand-painted textures, Unreal Engine quality, natural daylight through jungle canopy, warm stone temple lighting, digital painting feel, hard-edge brushwork, heavy shadows, not cartoon not Disney not Pixar, film grain, premium anime quality, high energy beat and impact sounds dominant and aggressive throughout, camera angle changing on every single strike, never static, never the same angle twice in a row.

Spiderman vs Captain America

Create a dynamic cinematic commercial video strictly based on the uploaded storyboard. Treat every storyboard panel as a chronological keyframe of a single continuous sequence. Recreate each frame in exact order from first to last, preserving the product, composition, objects, ingredients, environment, lighting, colors, textures, and visual storytelling shown in the storyboard. Generate smooth in-between motion connecting all storyboard panels naturally. Every scene must evolve seamlessly into the next scene without cuts that break continuity. The camera movement and object motion should logically guide the viewer through the storyboard sequence. Animate all visible elements with realistic physics and premium advertising-quality motion. When products open, separate, transform, reveal ingredients, fillings, liquids, textures, interiors, toys, or surprises, these actions must happen naturally and cinematically according to the storyboard progression. When splashes, explosions, particles, crumbs, smoke, steam, powders, droplets, liquids, flying ingredients, or floating objects appear, animate them with realistic high-speed commercial physics. Use dynamic camera language throughout the sequence: macro push-ins, cinematic pull-backs, orbit shots, dolly moves, crane movements, tracking shots, speed ramps, smooth transitions, slow motion, and dramatic reveals. Emphasize premium product presentation with extreme macro detail, realistic materials, glossy reflections, volumetric lighting, shallow depth of field, natural motion blur, and photorealistic textures. Maintain perfect visual continuity between storyboard panels. Do not invent additional scenes. Do not skip frames. Do not repeat frames. Do not rearrange the sequence. The final shot must match the last storyboard panel exactly. Video style: ultra-realistic premium commercial advertisement, cinematic food/product photography, high-end brand campaign quality, dramatic lighting, realistic physics, rich detail, smooth animation, professional product showcase. Duration: 10 seconds. Aspect ratio: match storyboard. Single continuous sequence. No text. No captions. No subtitles. No numbers. No logos added beyond the original product. No watermarks. No UI elements. No scene repetition. Transform the uploaded storyboard into one seamless, dynamic, high-end commercial video while preserving the exact visual narrative and frame order shown in the storyboard.