Two Men at a Dark Bar

Inside an enormous ancient temple with towering stone pillars, cracked marble floors, floating dust particles, and powerful beams of sunlight streaming through the ceiling, a beautiful young female martial artist wearing an elegant flowing pink hanfu-inspired combat dress performs an extraordinary combat dance. The sequence opens with an ultra-wide cinematic shot as she glides gracefully across the polished stone floor. Her silk sleeves and layered skirt flow naturally with every movement while the camera circles her in a smooth 360-degree motion. She suddenly launches into an acrobatic spinning kick, flipping through the air with flawless martial arts precision. The camera switches to an extreme low-angle shot emphasizing her height and power as her dress ripples realistically. The action transitions into dramatic slow motion as she lands softly before instantly accelerating into rapid spinning footwork. Every movement creates realistic cloth simulation, subtle dust bursts, and perfectly synchronized body mechanics. She leaps high beneath a brilliant shaft of heavenly light. The camera follows from below while volumetric lighting surrounds her, creating an angelic silhouette. Her expression remains calm, focused, and fearless. As she begins rotating in midair, glowing golden energy ribbons emerge around her body, spiraling outward like magical dragon-shaped currents. Sparks, floating embers, and swirling particles react naturally to her motion. The camera pulls back into a breathtaking wide shot as the golden energy tornado expands across the temple floor, illuminating the ancient architecture with warm orange reflections. Dust rises naturally while light scatters realistically through the atmosphere. The final moment freezes in epic slow motion as she floats gracefully at the center of the glowing spiral beneath the heavenly beam, surrounded by rotating rings of golden energy. The camera slowly cranes upward, revealing the magnificent temple from above before fading to black.

Female Anime Character: ZERO-RED with two large futuristic energy rifles on each hand with integrated magenta glows. Framing the face, light red piercing eyes, confident and soft airbrush rendering facial expressions and long crimson high red ponytail. Full body view silver cyber combat suit layered in segmented armor plating, with pink-magenta illuminated energy cores. Use @ character ref, the only character and visual style reference. Start with action and end with action. A female cyber gunfighter skates through a dark frozen arena with a slight pink and reflective surface, at quick speed, performing strong yet elegant ice-ballerina movements while shooting large numbers of enemies with her two rifles. She attacks during pirouettes, spiral glides, sweeping turns, spins and rapid skating transitions without ever slowing down. Many enemies surround and rush her from all directions. Every skate movement leaves electric-magenta glow trails, static energy and light red spray. Every gunshot creates vivid crimson muzzle flashes and sharp graphic impact bursts. Keep the environment dark, moody and high-contrast, with bold red, magenta and black posterized anime visuals. No sword, no slow motion, no static posing, no static enemies, full action throughout.

High-quality 3D animated cartoon, stylized family-animation look, oversized heads, big expressive eyes, exaggerated proportions, smooth rounded clay-like 3D surfaces, subsurface scattering skin, vibrant lighting. NOT realistic, NOT live action. A young man in a cluttered bedroom, oversized flannel shirt, messy hair, pulls an old jacket from the back of a closet. 0–4s: medium shot, he digs through the jacket's pockets absentmindedly, half-interested, checking each one and finding nothing. 4–7s: his hand pauses in the last pocket. Fingers close slowly around something folded and papery deep inside. Building anticipation on his face. 7–9s: whip-cut to a tight close-up as he pulls out a crumpled twenty-dollar bill, holding it between two fingers, eyes going huge in disbelief. 9–12s: warm golden light blooms around the bill like it's glowing, dust motes drifting through the beam, his jaw dropping open in pure joy. 12-15s: slow push-in as his face melts into a huge triumphant grin, both fists pumping the air, bill held high like a trophy. Camera: medium searching shot, whip-cut to his face on discovery, slow dramatic push-in as he raises the bill into the light. AUDIO: quiet rustling fabric while searching, a single soft gasp the instant he finds it, a bright angelic choir "aaah" swelling as the golden light blooms, a triumphant horn flourish as he pumps his fists. Vibrant saturated colors, warm golden treasure-light against a muted bedroom palette, smooth stylized 3D character design, family animation aesthetic, no text, no watermark, 16:9. FORBIDDEN: no named studio or franchise references, no readable text or logos, no realistic rendering, no design drift, no visible brand logos.

Cinematic 15-second commercial, realistic Minecraft Creeper cake creation, every ingredient is real food while the final design is Minecraft-inspired. Ultra-fast pacing, hard cuts, macro cinematography, first-person POV, dynamic camera, motion blur, shallow DOF, HDR, 4K, realistic textures, no text, no subtitles, no numbers, no logos, no watermarks. POV sifts flour into a bowl, eggs crack, sugar and butter whip into a silky batter, flour folds in, batter pours into round cake pans. Rapid transition into oven POV as sponge layers rise with steam. Fresh layers are sliced perfectly level, thick cream spreads smoothly between each layer, cake stacks higher with fast rotating camera. Thin crumb coat is applied, then vibrant green fondant drops over the cake and stretches tightly across every edge. Fondant smoother glides across the surface creating perfectly sharp sides. Black and brown chocolate squares snap into place one by one forming Creeper faces and pixel dirt blocks. Green buttercream grass grows rapidly around the edges through fast piping macro shots. Small edible brownie cubes and chocolate blocks land onto the top like Minecraft terrain. A realistic edible Creeper figure is assembled piece-by-piece and placed on top of the cake. Camera performs rapid orbit shots, top-down spins, extreme macro close-ups of fondant texture, chocolate pixels, buttercream grass and realistic cake details. Finish with an energetic hero reveal: first-person hands present the completed two-tier Minecraft cake while the camera circles quickly, ending on an ultra-macro close-up of the Creeper face, realistic frosting texture and pixel decorations under cinematic lighting.

ã€é£Žæ ¼ã€‘竞技体æ“平衡木大片（Artistic Gymnastics Balance Beam），电视转播+电影感混åˆè´¨æ„Ÿï¼ˆ16:9 Cinematic, Photorealistic），高速摄影å‡æ ¼ï¼ˆSuper Slow-mo），体æ“馆暖顶光+深色背景影调，国家队级动作规格 ã€æ—¶é•¿ã€‘15秒 ã€åœºæ™¯ã€‘体æ“比赛馆：一æ¡æ ‡å‡†å¹³è¡¡æœ¨ï¼ˆ10厘米宽木é¢ï¼‰å±…于画é¢ä¸­å¤®ï¼Œæœ¨ä¸‹é“ºè“色缓冲垫，背景观众席和è£åˆ¤å¸­è™šåŒ–æˆæš—色å—，顶棚射ç¯æŠŠæœ¨é¢ç…§å¾—å‘亮，空气中有é•ç²‰å¾®å°˜ ã€è§’色】主角 [00:00-00:02] 镜头1：上木å‰ï¼ˆThe Chalk） 近景：主角站在平衡木一端下方，åŒæ‰‹äº’æ“é•ç²‰ï¼Œç™½è‰²ç²‰å°˜åœ¨å°„ç¯å…‰æŸ±é‡Œç¼“缓飘散；她抬眼望å‘木é¢ï¼Œä¸‹é¢Œæ”¶ç´§ï¼Œå‘è£åˆ¤æ–¹å‘举手示æ„。 ã€éŸ³æ•ˆã€‘全场安é™ï¼Œåªæœ‰é•ç²‰æ‘©æ“¦çš„沙沙声和一次深呼å¸ã€‚ [00:02-00:04] 镜头2：上木（The Mount） 侧é¢ä¸­æ™¯ï¼šå¥¹åŠ©è·‘两步è¸è·³æ¿è…¾èµ·ï¼ŒåŒæ‰‹æ’‘木顺势æˆç›´è§’支撑å†ç«‹èµ·ï¼Œç¨³ç¨³ç«™ä¸Š10厘米宽的木é¢ï¼ŒåŒè‡‚展开æˆä¸€æ¡ç›´çº¿ï¼Œè„šå°–绷直。 ã€éŸ³æ•ˆã€‘è¸æ¿ä¸€å£°é—·å“，è½æœ¨åŽæœ¨èº«è½»å¾®çš„震颤声。 [00:04-00:08] 镜头3：木上翻腾串（The Acro Series, Super Slow-mo） å‡æ ¼æ…¢åŠ¨ä½œï¼Œæœºä½æ²¿æœ¨ä¾§å¹³è¡Œè·Ÿç§»ï¼šå¥¹åœ¨æœ¨ä¸Šå®ŒæˆåŽæ‰‹ç¿»æŽ¥ç›´ä½“åŽç©ºç¿»çš„连接串，åŒè„šæ¯æ¬¡éƒ½ç²¾å‡†è½å›žå·´æŽŒå®½çš„木é¢æ­£ä¸­ï¼Œè½æœ¨çž¬é—´è„šè¸å¾®è°ƒã€åŒè‡‚快速展开控制平衡，å‘髓甩出的碎å‘æ悬在空中，é•ç²‰ä»Žè„šä¸‹ç‚¸èµ·ä¸€å°å›¢ç™½é›¾ã€‚ ã€éŸ³æ•ˆã€‘å‡æ ¼æ®µæ¯æ¬¡æ‰‹æŽŒå’Œè„šæŽŒæ‹åœ¨æœ¨é¢ä¸Šçš„"ç °"声被拉长放大，木身震颤的嗡声清晰å¯é—»ã€‚ [00:08-00:10.5] 镜头4：跳步与转体（Leap + Turn） æ¢å¤å¸é€Ÿï¼Œæ­£ä¾§é¢æœºä½ï¼šå¥¹èµ·è·³å®Œæˆä¸€ä¸ªäº¤æ¢è…¿å¤§è·³ï¼ŒåŒè…¿åœ¨ç©ºä¸­åŠˆæˆä¸€æ¡ç›´çº¿ï¼Œè½æœ¨åŽç´§æŽ¥å•è„šç«‹è½¬ä¸¤å‘¨ï¼Œå¦ä¸€æ¡è…¿ç›˜äºŽè†ä¾§ï¼Œè½¬é€Ÿå‡åŒ€ï¼Œè£™æ‘†å¼çš„动作线æ¡å¹²å‡€åˆ©è½ã€‚ ã€éŸ³æ•ˆã€‘跳起è½æœ¨ä¸¤å£°è½»å“，转体时åªæœ‰è„šæŽŒä¸Žæœ¨é¢çš„摩擦声。 [00:10.5-00:13] 镜头5：下法（The Dismount） 侧é¢æœºä½å¾®ä»°æ‹ï¼šå¥¹æœæœ¨ç«¯è¿žæŽ¥ä¸¤ä¸ªåŽæ‰‹ç¿»åŠ é€Ÿï¼Œä»Žæœ¨ç«¯è…¾ç©ºè€Œèµ·ï¼Œå®Œæˆç›´ä½“åŽç©ºç¿»è½¬ä½“720度，身体在空中拉æˆä¸€æ¡ç´§ç»·çš„线，è½åœ°åŒè„šç ¸è¿›åž«å­â€”—纹ä¸ä¸åŠ¨ï¼Œä¸€æ­¥æœªæŒªï¼ŒåŒè‡‚瞬间上举亮相。 ã€éŸ³æ•ˆã€‘腾空åŽé£Žå£°ï¼Œè½åž«ä¸€å£°åŽšå®žçš„é—·å“，éšåŽå…¨åœºçˆ†å‘出掌声与欢呼。 [00:13-00:15] 镜头6：亮相定格（The Salute） æ­£é¢ä¸­æ™¯ç¼“慢推近：她ä¿æŒäº®ç›¸å§¿åŠ¿ï¼Œèƒ¸å£å¿«é€Ÿèµ·ä¼ï¼Œé¢è§’有汗ç å光，éšåŽè½¬å‘è£åˆ¤å¸­å†æ¬¡ä¸¾æ‰‹è¡Œç¤¼ï¼Œå˜´è§’上扬，画é¢åœ¨å¥¹çš„笑容上定格。 ã€éŸ³æ•ˆã€‘掌声æŒç»­ï¼Œå®šæ ¼é…一声上扬收æŸéŸ³ã€‚ ã€å¤‡æ³¨ã€‘所有木上动作åŒè„šå¿…须始终è½åœ¨æœ¨é¢ä¸Šï¼Œè½æœ¨çž¬é—´å…许å°å¹…手臂调整但身体ä¸æ‘‡æ™ƒï¼›ä¸‹æ³•è½åœ°å¿…é¡»"é’‰"在垫上纹ä¸ä¸åŠ¨ï¼ˆStuck Landing），这是全片最é‡è¦çš„讆忆点

A 15-second ultra-realistic cinematic action scene of a fearless woman battling a group of wild lions in an ancient desert arena. Shot 1 (0–3 sec): Extreme wide cinematic shot of an ancient ruined arena surrounded by rocky mountains at sunset. A powerful woman wearing a rugged warrior outfit stands alone in the center while three massive lions slowly approach from the shadows. Dust moves through the air, dramatic atmosphere, intense tension. Shot 2 (3–6 sec): Low-angle action shot as the lions charge toward her. The camera moves quickly around the scene as she dodges the first attack with incredible agility, sliding across the ground while dust explodes around her. Realistic lion movement, detailed fur, cinematic motion blur. Shot 3 (6–10 sec): Fast-paced combat sequence. The woman uses her skills and intelligence to fight back, avoiding attacks and overpowering the lions one by one. Dynamic camera angles, close-ups of her determined expression, slow-motion moments showing powerful movements, realistic physics and intense action choreography. Shot 4 (10–13 sec): The final lion makes a dramatic leap toward her. The camera rotates in slow motion as she blocks the attack and defeats the lion with a powerful final move. Dust fills the air as the arena becomes silent. Shot 5 (13–15 sec): Epic hero shot. The woman stands victorious in the center of the arena as the defeated lions retreat into the distance. The camera slowly pulls back, revealing the vast landscape, golden sunset, wind moving her hair and clothes, cinematic victory moment. Style: Ultra-realistic Hollywood action film, epic scale, realistic animal movement, dramatic lighting, cinematic camera work, detailed textures, 4K quality, natural motion, intense atmosphere. Negative Prompt: cartoon, fantasy CGI look, unrealistic animals, blurry, bad anatomy, extra limbs, unnatural fighting, blood, gore, distorted face, flickering, low quality, watermark.

FORMAT / START] Generate a 15-second 16:9 cinematic action scene. Action begins immediately. No reference images provided — generate both characters from the text description. [CHARACTER LOCK] A is The Runner. Preserve exact: woman early 30s, lean athletic build, dark hair tied back and soaked flat, grey tank top, dark cargo pants, bare forearms, low compact stance, contained energy. She is calm. She wins. B is The Enforcer. Preserve exact: woman mid 30s, taller and heavier, short bleached hair, black tactical jacket, gloved hands, upright forward-leaning posture, volatile energy. She attacks first. She loses. Do not swap roles. Only these two characters. No extras. [SETTING] Rooftop of a residential block, night, heavy rain. Standing water across flat concrete, ventilation units, a low parapet, laundry lines strung between poles with wet sheets snapping in the wind. A single service lamp on the stairwell housing — hard white light, everything else in blue shadow. City glow far below. No crowd. [MOTION RULES] Readable action, clear silhouettes, short pause before contact, visible setup before every strike, hard impact with real physical weight, short hit-stop, controlled smear only on fast-moving limbs, clear geography. Every cut continues the same exchange. No slow motion, no speed ramping — full real-time speed throughout. Movements grounded and physically motivated. Each technique flows directly into the next without discontinuity. A moves with economy and precision. B moves with aggression and momentum. Rain and standing water react to every impact. [BEAT 1 — ENTRY] Both women face each other across the wet roof, three meters apart. B advances fast, right hand cocked. A stands relaxed, weight centered, hands low, eyes on B's sternum. A wet sheet swings between them and passes. The gap closes. [BEAT 2 — FIRST EXCHANGE] B throws a heavy right cross. A slips outside, lets it pass her ear, and fires a short left hook into B's ribs mid-extension. B grunts, arm drops. Water sprays from both their feet. [BEAT 3 — COUNTER] A is already inside. She catches B's collar with her left hand and drives a short right elbow across the cheekbone — dense contact, proximity not distance. B's head rotates. Her back foot slides out on the wet concrete. [BEAT 4 — RECOVERY EXCHANGE] B tears free and charges, driving full bodyweight into a tackle. A absorbs one step back, drops her weight, hooks an arm under B's shoulder and pivots — using B's own momentum to throw her over the hip into standing water. Heavy flat landing. Spray erupts. [BEAT 5 — FINISH] B pushes up onto one knee, no guard left. A steps in and drives a knee into her sternum. B folds and goes flat on her back in the water. A straightens, exhales once, rain running off her, and walks past toward the stairwell without looking back. Final image: A silhouetted in the service lamp, B down in the water behind her, sheets still snapping in the wind. [AUDIO] No music. Only: rain on concrete, wet footwork, hook into ribs, elbow on bone, boot sliding on wet floor, body hitting standing water, one long exhale at the finish, sheets flapping in wind. [NEGATIVE CONTROL] No random cuts, no role swap, no outfit change, no face morphing, no extra characters, no teleporting, no unclear contact, no comedy, no subtitles, no logos, no UI, no slow motion, no speed ramping, no wire work, no supernatural movement, no weapons. Every position change physically motivated and continuous

Concept: "Han River Analog Footage × Higgsfield" Create an authentic 15-second analogue home-video recording filmed with a handheld Hi8/VHS-C camcorder. The camcorder is never visible. The viewer only sees the recorded footage, as if watching a digitized family videotape. This is not a vlog, commercial, or cinematic film, just a friend casually filming another friend. Use genuine analogue camcorder aesthetics: soft SD resolution, VHS grain, faint interlacing, analogue colour bleed, mild tracking noise, autofocus breathing, slight exposure fluctuations, imperfect handheld shake, natural motion blur, washed-out colours, and subtle tape hiss. No HDR, beauty filters, cinematic grading, stabilisation, or AI-perfect rendering. A young Korean woman in her early twenties wears a pastel fitted crop top, mid-rise vintage blue denim shorts, white ankle socks, classic white sneakers, a lightweight windbreaker tied around her waist, and carries a simple canvas shoulder bag. Her styling should evoke nostalgic late-90s/early-2000s Korean street fashion without mentioning a specific year. The scene takes place beside the Han River on a warm summer afternoon. Families relax on the grass, cyclists ride past, children play, and the distant Seoul skyline fills the background. She walks naturally toward the riverside railing before stopping to admire a massive floating HIGGSFIELD installation in the river. The word HIGGSFIELD is built from giant three-dimensional block letters, mounted on a realistic floating platform. The sign must appear physically installed in the river with accurate reflections, shadows, and perspective, remaining large, bold, and perfectly readable throughout the scene. The camera gently pans between her and the installation, keeping both in frame. She smiles naturally, takes a few steps closer, then notices the unseen camera operator. She laughs softly, gives a warm wave, points excitedly toward the giant HIGGSFIELD sign, gives a subtle thumbs-up, then looks back at the river with a smile before continuing her walk. A light breeze moves her hair and clothing while a cyclist passes behind her. The camera follows for a few more steps before stopping naturally. Generation Requirements: One continuous 15-second recording with no cuts or transitions. Keep movements natural, identity consistent, anatomy accurate, and interactions believable. The HIGGSFIELD installation must stay prominent, crystal clear, and naturally integrated into the environment. Include realistic ambient sounds such as wind, birds, distant conversations, bicycles, and water. Avoid AI artifacts, duplicated pedestrians, warped anatomy, floating objects, modern cinematic effects, or anything that makes the footage look AI-generated. The final result should be indistinguishable from genuine nostalgic analogue camcorder footage.