X'ten · Hareket aktarımı
Camera movement prompts for AI video
Prompts that spell out the camera: dolly push-ins, orbits, crane rises, whip pans, FPV dives and handheld follow shots. Each one is shown next to the clip it made, so you can see how the wording turns into motion before you reuse it in Seedance, Kling or Veo.
Cozy cinematic café coffee moment
Created this cozy cinematic coffee video, featuring the beautiful atmosphere of a modern café. The video begins with a warm cup of coffee placed on a marble table, creating a calm and inviting mood. Close-up shots capture the espresso flowing smoothly from the coffee machine into the cup. The rich coffee details, creamy texture, and warm lighting make every scene feel realistic and visually appealing. The background shows a lively café environment with natural movement and a relaxing atmosphere. Smooth camera movements and cinematic angles bring the coffee-making process to life. The combination of soft lighting, detailed visuals, and ambient café vibes creates a premium cinematic experience. This video captures the simple beauty of coffee moments in a creative and engaging way.
MiniDV Japanese travel memory montage
Create exactly 30 seconds of photorealistic live-action Japanese travel memories, edited with quick hard cuts, seen through an affectionate adult boyfriend's first-person handheld consumer camcorder. This is a MULTI-SCENE MONTAGE, NOT one continuous take. Begin with a frightening cinematic near-miss at a suburban railway platform, then move immediately into warm, playful, intimate everyday travel. The emotional arc is alarm → relief → affectionate curiosity → golden-hour quiet → joyful night → tender laughter. No disaster imagery. No dramatic movie grading. Output ultrawide 21:9, requested 1080p, high-bitrate encode, with deliberately authentic early-2000s MiniDV aesthetics baked into the image. Use the full wide frame, no 4:3 pillarboxes. SAME ADULT HEROINE IN EVERY SHOT: SAME ADULT HEROINE IN EVERY SHOT: A very beautiful young adult Korean woman, age 25–28, with luminous fair warm-toned skin and realistic pores, expressive dark-brown eyes, softly sculpted youthful cheeks, naturally beautiful full rose-pink lips with a subtle satin finish, natural understated Korean makeup, and a warm spontaneous smile. Preserve her facial geometry, body proportions and identity across every cut, without idealized plastic smoothing. Thin dark-metal oval prescription glasses with clear lenses, eyes always readable. Long, naturally dark-brown to soft-black hair, smooth and healthy with subtle natural movement; no blonde hair, no bangs. One consistent modern Korean Gen Z outfit: an elegant loose ivory ankle-length A-line dress with full-length sleeves and a high neckline, subtle delicate embroidery, paired with a lightweight oversized beige cardigan jacket, clean cream sneakers, and a small structured taupe crossbody bag. Fabric remains naturally flowing, fully opaque and modest during movement. Glasses, hairstyle, outfit and bag remain identical throughout. Contemporary Korean styling, youthful, effortlessly elegant and understated. No spaghetti straps, revealing neckline, exposed midriff, or body-hugging styling. There is no supplied reference image: establish this original Korean heroine from these text details and maintain her exact identity. Never morph into another woman, change her ethnicity, facial features, hairstyle, or body proportions. MINIDV STYLE LOCK: Feels like a cherished home-video cassette recorded around 2003, not polished cinema pretending to be vintage. Gentle optical softness, low contrast, softly washed highlights, warm natural skin, fine luma noise, mild color bleed, subtle consumer-camera edge enhancement, restrained DV compression texture, slightly faded blues and greens, soft halation, occasional exposure pumping when moving between shade and sunlight. Real handheld wide-angle proximity, uncentered composition, brief autofocus searching that settles on her face. Noticeable motion smear on fast pans and night lights, but keep expressions legible. No heavy glitch overlays, fake VHS tracking bars, film burns between scenes, timestamps, captions, subtitles, logos, or excessive chromatic aberration. Noon feels airy, sunset feels honey-gold, night feels softly smeared and imperfect. Do not use green-noir grading or glossy blockbuster teal-orange. POV AND EDIT RULES: The camera operator is her adult boyfriend, mostly unseen, occasionally one anatomically correct forearm or hand enters. She repeatedly checks where he is, leans toward him, looks back, shares a discovery, and smiles directly into the lens. Small natural gestures, no fashion posing, no slow-motion beauty commercial. Hold each simple action long enough to read. Hard cuts between locations, no morphing transitions. Geography evokes a remembered Kansai day through Uji and Nara, with ordinary bookshops and shopping streets between famous places. Late spring or early summer green foliage; no contradictory autumn leaves, snow or cherry blossoms. Do not squeeze in extra landmarks or relocate Mount Fuji into Nara. EXACT 30-SECOND EDIT: 00.00–04.00 — SUBURBAN PLATFORM, DANGER THEN RELIEF Open with danger already readable: she has distractedly stepped too near the platform edge while peering down the line, one hand shading her eyes. She remains on the platform, never on the rails. An approaching maroon-and-cream commuter train is already visible behind her and rapidly grows larger. From first-person POV, one boyfriend hand reaches to her upper arm; upon contact, the arm immediately retracts in ONE continuous movement, pulling her backward toward the lens and behind the tactile safety strip. She visibly moves back one or two quick steps; camera operator steps back naturally with her. Only once she is clear does the train rush past behind her. Its wind flutters her overshirt and dress hem, never lifts or removes her hijab or exposes her body. Her eyes widen behind glasses, lips part with a brief startled gasp, then relief begins in her expression. No contact with the train, injuries, lingering on the edge, or repeated pulling motion. Believable distances and train direction; staged fictional action with a clearly successful rescue. Keep one continuous shot for this opening. 04.00–06.00 — SUNNY GRASS Abrupt warm hard cut: first-person handholding as she guides the viewer along a grassy path. She turns toward the camera with a relieved bright laugh. Sunlight catches her ivory sleeves and navy floral hijab pin; green grass drifts softly behind. One hand linked naturally with the operator, no duplicate fingers. 06.00–07.50 — OLD BOOKSHOP Shoot through a gap between shelves of well-used Japanese books. She playfully bends into the gap from the far side, notices the lens, gives a small conspiratorial smile. Dusty window light, crowded wooden shelves, autofocus briefly catches book spines before her eyes. Avoid legible invented titles. 07.50–09.00 — QUIET STREET CROSSING Low, casual waist-to-shoe framing of her walking across an ordinary Japanese zebra crossing on a pedestrian signal. Ivory dress hem sways over trainers. Tilt just enough to glimpse her looking back. No fetishized body framing; modest everyday clothing detail and footsteps are the focus. 09.00–10.50 — NARROW LANE She walks ahead between traditional wooden storefronts, potted plants and hanging fabric curtains. Her eyes glance sideways first, then she turns her face back to find the camera, smiling softly. Overhead cables and bicycles at the side establish lived-in Japan; no staged crowd. 10.50–12.50 — UJI RIVERSIDE, ADDED RESEARCHED SCENE At the scenic Uji River in Kyoto Prefecture, film her by a pedestrian riverbank with the historic bridge in the background. Begin with sunlight moving on water, tilt casually to her leaning lightly on the safe railing. A breeze moves her cuff; she points out the view, then looks at the operator rather than posing for the landscape. Keep her face central to the memory, river sound underneath. 12.50–14.00 — UJI MATCHA, ADDED RESEARCHED SCENE Close handheld view of her holding a simple matcha soft-serve outside a small tea shop. She takes a tiny taste, raises her eyebrows appreciatively, then offers the cone toward the lens with a playful smile. Rich natural tea-green against ivory fabric; mild sunlight overexposure, soft storefront background. No visible brand text. 14.00–15.50 — SMALL RAMEN SHOP Close-up of steaming vegetarian mushroom-and-vegetable ramen on a wooden counter or small table, then a quick natural focus pull to her sitting opposite, smiling through wisps of steam. Cozy lamps, spoon and bowl sounds, genuine restaurant ambience. No pork slices or alcohol props. Camera remains handheld, not a glossy food advertisement. 15.50–17.00 — MANGA SHOP She draws one manga volume from a shelf, opens it and lowers her eyes with a curious smile. Brief close shot of hands and page texture; no readable dialogue or famous copyrighted character emphasis. Maintain same glasses and floral pin. 17.00–18.50 — NARA PARK DEER Boyfriend POV offers one designated deer cracker, held flat and released gently to a deer in Nara Park. The woman is softly out of focus a few steps behind on green grass, laughing at the deer's polite approach. No grabbing antlers, hugging or chasing wildlife. Mild camera wobble as the deer approaches. 18.50–20.00 — QUIET READING Sitting on a park bench, she flips a manga page. Close details of her hands, gold botanical embroidery and lace cuff, then tilt to her peaceful face behind her glasses. Rustle of paper, leaves and distant birds. Natural, modest, unposed. 20.00–22.50 — UKIMIDO PAVILION, ADDED RESEARCHED SCENE Nara Park's hexagonal wooden Ukimido pavilion with its cypress-bark roof on Sagiike Pond, reached by a wooden bridge. From the public bridge, she pauses safely by the railing and turns back toward the camera; warm light and green trees reflect in still water. Start wide enough to read the actual pavilion silhouette, then walk a little closer as she extends her hand toward the viewer. Quiet water, wooden footfalls, no temple ritual invented. Intimate location memory rather than a drone tourism shot. 22.50–24.50 — GOLDEN-HOUR MEADOW In Nara's grassy park with deer grazing at a respectful distance, honey-gold backlight edges her hijab and sleeves. She turns gently toward the camera, dress fabric swaying with the evening breeze. A small genuine smile, peaceful eye contact, slight exposure bloom. No spinning near deer, no flying fabric revealing skin. 24.50–27.00 — LIVELY NIGHT SHOPPING STREET Blue-hour-to-night hard cut. Warm shop lanterns, illuminated signs and pedestrians in a Kansai shopping street. She suddenly jogs lightly several steps ahead, then looks back laughing, waiting for the operator. He follows; handheld camera movement leaves convincing soft DV light streaks. No vehicles in her running path; stay in a pedestrian area. Preserve facial readability between movement smears. 27.00–28.00 — TACTILE MEMORY INSERT One brief close handheld detail of her hand adjusting her embroidered cuff beside her taupe bag, ivory skirt hanging naturally; warm shop light blooms softly off the stitching. A casual recorded detail, not an erotic crop. 28.00–30.00 — CAMCORDER MEMORY ENDING A second small silver early-2000s consumer camcorder is held into the foreground by the operator's free hand, its side LCD flipped outward. Current POV remains on the original recording camera; make this a clearly separate physical device, not an impossible view of itself. The second camcorder points toward the woman; its LCD shows the same woman in the same pose at the same moment, synchronized without a duplicate person standing nearby. She notices that he is still filming, briefly gives a playful tiny pout with her naturally rose-toned lips, then breaks into a real warm laugh looking directly at him. Frame drifts toward a bright shop lamp and highlights bloom to soft white over the final fraction of a second. End on warmth, not alarm. AUDIO: Real synchronized location sound ONLY, no background music, no soundtrack, no narration, no scripted dialogue. Opening: rolling rail clatter, approaching train rush, wind blast, a single surprised gasp, shoes stepping back. Then light footsteps, grass and breeze, natural soft laughter, book pages, distant shop ambience, river water, ice-cream wrapper rustle if present, ramen bowls and chopsticks, park birds, deer footsteps, wooden bridge creaks, evening breeze, night crowd and distant vehicles. Audio hard-cuts naturally with each location; subtle room-tone tails are acceptable. Finish with her authentic laugh. No artificial whooshes between edits or dramatic suspense score. FINAL QUALITY RULES: Exactly one stable adult heroine, full modest coverage, consistent hijab and glasses, anatomically natural hands, no face drift or changing ethnicity, no glossy skin filter. Requested high-bitrate file must preserve intentional MiniDV texture rather than remove it. The viewer should feel these are precious imperfect memories from a real day together, not a compilation of polished commercial travel footage.
This AI video is an absolute masterpiece.
A cinematic, photorealistic 10-second dark-fantasy martial arts fight in 16:9, shot like a high-budget film with aggressive handheld camera, whip pans, low ground-level tracking shots, motion blur, flying dust and sparks, ultra-intense energy. Setting: a vast red-orange desert battlefield with ancient sandstone ruins and a stone doorway in the distance, cracked dry earth, scattered swords and broken weapons lying on the ground, hazy bright sky with harsh sunlight and long shadows, thick dust and smoke in the air. Color grade: warm burnt-orange and red tones, hazy highlights, deep contrast, gritty film grain. Characters: MISTY, a young woman fighter with short-to-medium dark hair, a white cloth blindfold over her eyes, all-black outfit (black jacket, black pants, black boots), two katanas strapped across her back. Her power: long, living black hair-like tendrils that lash out from her like whips, coil into a spherical shield, and rise as a huge black wall with a giant dark humanoid figure looming behind it. DORJE, an elderly bald monk with a weathered, stern face, wearing a deep maroon robe, barefoot or in simple sandals. His power: fire; glowing red-orange energy ignites in his fists and he throws fireballs that explode on impact; he can leap high into the air. Shot 1 (0–1s): Extreme close-up, backlit. Misty swings her arm in a punch while Dorje's glowing red-orange fist and arm lock against hers; dust and tiny black fragments fly. Cut: Misty ducks and spins, Dorje's fireball bursts beside them in a red glow. Shot 2 (1–2s): Low ground-level shot. A huge sweeping curtain of black tendrils whips across the ground in front of the camera, a large orange fireball erupts behind it, and Dorje stands in a crouch in the red explosion. Then Misty sprints toward him as red-black smoke billows behind her and Dorje swings his robe away. Shot 3 (2–3s): Wide shot. Misty and Dorje trade close-range strikes in the open desert, her fist hitting his robe, then a high kick that he blocks, katana hilt visible on her back, swords scattered on the sand. Shot 4 (3–4s): Misty spins with a coiled black tendril shield around her arm while a dark smoke cloud billows behind. Dorje lunges in with a huge fireball in his hand and strikes her shield; the black coil swirls around the flame. Shot 5 (4–6s): Low angle. A towering wall of black tendrils erupts from the ground behind Misty, with a giant dark humanoid figure rising above it. Dorje leaps high into the air with a glowing red fireball burning in his hand while Misty stays rooted below. Dorje comes crashing down into a massive orange fireball explosion that fills the frame. Shot 6 (6–7s): Dark, dusty extreme close-up of a spiraling black tendril sphere in the red haze. Cut to wide: Misty whips a long black tendril in a wide arc at Dorje through the dust, with sparks flying. Shot 7 (7–8s): Fireballs explode around Misty with a huge black smoke plume as she dashes through them with a katana hand, while Dorje is thrown upside down in the air above, a red flame on his fist. Then a wide shot: Misty stands in a combat stance with fire exploding behind her, Dorje faces her with a glowing red fist, the stone doorway in the background. Shot 8 (8–10s): Intense close fight. Black tendrils lash like a giant whip across the frame with orange sparks everywhere, Dorje blocks with glowing red fists. Then the camera pushes into a glowing fire cloud, where a coiled black tendril sphere sits in the explosion, then dark smoke and dust billow across the cracked ground and everything falls quiet. Audio: heavy cinematic percussion and low sub-bass drones, explosive fireball booms, whooshing whip sounds for the tendrils, sharp impacts, crackling fire, rumbling ground, wind and dust.
Rainy industrial escape thriller
Create a 30-second ultra-photorealistic live-action action-thriller at night during heavy rain. A young woman trapped inside an old industrial building discovers a partially open mechanical loading door and times her escape as two pursuers close in. She slides beneath the heavy door, struggles with realistic friction and backpack movement, rolls onto the rain-soaked pavement, and escapes into a dark service lane. Show realistic human biomechanics, inertia, momentum, gravity, friction, center-of-mass shifts, realistic loading-door mechanics, backpack physics, rain and water dynamics, wet-surface traction, vehicle spray, and physically accurate material interactions. Use handheld chase cinematography, low floor-level shots during the escape, close-ups of the mechanical door, exterior tracking shots, and a final static wide shot. Camera movement should have natural operator inertia and subtle handheld shake. Use only diegetic sound: heavy rain, breathing, footsteps, metal mechanisms, water splashes, distant traffic, echoes, and thunder. Natural nighttime lighting with realistic wet reflections, shadows, depth of field, and motion blur. End with the woman hidden behind a concrete wall as the pursuers run past, followed by an empty rain-soaked street and a cut to black.
Astronaut falls through clouds to spacecraft
REFERENCE IMAGE 1 = EXACT MALE CHARACTER REFERENCE IMAGE 2 = EXACT 30-PANEL STORYBOARD Create a 30-second ultra-realistic Hollywood sci-fi cinematic video following REFERENCE IMAGE 2 exactly in sequence, composition, camera movement, action and visual progression. CHARACTER: Use the male character from REFERENCE IMAGE 1 as the ONLY main character. Preserve his exact identity, facial structure, eyes, eyebrows, nose, lips, jawline, beard, moustache, hairstyle, hairline, skin tone, apparent age and body proportions. Never change his face or make him look like another actor. Keep his identity perfectly consistent whenever his face is visible. Transform his clothing into a realistic futuristic white EVA astronaut suit while keeping his exact face, hair characteristics and body proportions. Suit: white armored panels, dark gray/black technical sections, black gloves, transparent helmet visor, chest controls, life-support backpack, hoses and realistic mechanical joints. Keep the same suit throughout. SPACECRAFT: Use ONE identical futuristic spacecraft throughout. Dark gunmetal/graphite body, large angular armored fuselage, broad swept wings, central cockpit, detailed mechanical structures and large twin rear engines with amber/orange lights. Never redesign, reshape or duplicate the spacecraft. STYLE: Photorealistic high-budget Hollywood science-fiction film. ARRI Alexa cinematic look, realistic skin and metal textures, physically accurate reflections, volumetric clouds, atmospheric haze, HDR, cinematic depth of field, subtle film grain, realistic motion blur and natural lighting. 16:9 widescreen. No cartoon/anime/illustration/cheap CGI appearance. 30-SECOND SHOT SEQUENCE: 01 | 00:00–00:01 Extreme wide aerial view above a gigantic ocean of clouds. Deep blue sky, massive white/gray clouds, warm sunrise in upper-right. Camera slowly moves forward. 02 | 00:01–00:02 Camera travels through thick volumetric clouds. Golden sunlight breaks through. Tiny distant figure begins appearing in center. 03 | 00:02–00:03 Reveal the male astronaut falling through the clouds, very small in frame, arms and legs extended. 04 | 00:03–00:04 Astronaut rapidly approaches camera. White EVA suit becomes clearly visible. Dramatic perspective and atmospheric depth. 05 | 00:04–00:05 Astronaut rotates naturally during free fall. Camera follows him. Face becomes visible through visor. 06 | 00:05–00:06 Medium close-up. Astronaut stabilizes himself and looks ahead. Clouds softly blurred behind him. Preserve exact face from Reference Image 1. 07 | 00:06–00:07 He brings both hands toward his chest and looks down at a small futuristic mechanical device. 08 | 00:07–00:08 Closer shot. He carefully inspects the metallic gold/bronze device with both black-gloved hands. 09 | 00:08–00:09 Extreme close-up of hands and device. He activates a mechanism. Small amber light begins glowing. 10 | 00:09–00:10 Macro device shot. Intricate mechanical components move while the central core glows brighter golden-orange. 11 | 00:10–00:11 Close-up of astronaut's face. Golden device light reflects naturally across his face and helmet. Focused expression. 12 | 00:11–00:12 A gigantic spacecraft begins deploying behind him. Large mechanical structures unfold and amber lights appear. 13 | 00:12–00:13 Frontal reveal. Wide spacecraft wings extend left and right behind the astronaut. Massive scale. 14 | 00:13–00:14 Spacecraft becomes fully deployed: dark gunmetal body, angular wings, central cockpit and mechanical sections. Camera pulls back. 15 | 00:14–00:15 Low-angle underside shot. Massive spacecraft passes directly above camera through the clouds. 16 | 00:15–00:16 Cut inside cockpit. Same male character is seated as pilot in the same astronaut suit. Futuristic controls and canopy visible.
Picnic girl and tiger-hat cat
A cinematic 30-second short film, photorealistic, Korean riverside picnic aesthetic, late afternoon golden hour fading into night.A young East-Asian woman with a short black bob sits cross-legged on a beige gingham picnic blanket on green grass beside a wide river. Behind her: a long blue bridge, distant city skyline, and a tall tower. She wears an oversized white-and-blue pinstripe shirt, loose black necktie, sheer tulle overskirt, dark mini skirt, white socks, and black loafers. In front of her: a plastic cup of iced cola, a cardboard fried-chicken box, a small fluffy pom-pom, and extra chicken pieces.A chubby gray-and-white tabby cat wearing a knitted beige tiger hat with ears, stripes, and cartoon teeth sits and walks on the blanket.Shot 1 (0–3s): Wide shot. Girl calmly eats a fried chicken drumstick. The cat rolls on its back, then stands and walks toward her, staring at the chicken.Shot 2 (3–6s): Girl looks at the cat and says clearly in English, soft but firm: “No. This is mine.” The cat rears up on hind legs and bats at her arm like a tiny tiger.Shot 3 (6–10s): Wind blows her hair. She flinches, eyes squeezed shut, mouth open in a comic yell as the cat slaps her sleeve, then she sighs and says clearly: “Okay, okay. Eat. Just eat.” Shot 4 (10–16s): Cat grabs a drumstick from the box with both paws, sits upright like a person, and nibbles it with eyes half-closed, blissful. Girl watches, then takes another bite of her own chicken.Shot 5 (16–19s): Girl stands up, gathers her tulle skirt, and says clearly: “I’ll come back. Leave it.” She walks out of frame to the left. Cat stays sitting, still holding the drumstick.Shot 6 (19–23s): Time-lapse. Sky turns orange sunset, then deep blue dusk. City lights and bridge lights bloom. Cat finishes eating, flops onto its side, rolls on the blanket, belly up, hat still on.Shot 7 (23–30s): Night. Close-up of the cat sitting neatly on the blanket beside leftover chicken and the cola cup. Soft city bokeh behind. Cat looks slightly guilty, then looks at the empty space where the girl sat, ears drooping. No more dialogue. Soft ambient river wind and distant city hum. End on a still close-up of the tiger-hatted cat
Korean girl runs through historical marketplace
Created a beautiful Korean girl walking confidently through a traditional Korean marketplace, wearing an elegant historical outfit, with long black hair flowing naturally and a detailed ancient Korean street surrounding her. At one second, the camera gently moves closer as she walks through the busy crowd, keeping her face clear, realistic, and consistent. At two seconds, show her passing wooden market stalls, traditional Korean houses, hanging lanterns, and merchants while people move naturally around her. At three seconds, she suddenly begins running through the crowded marketplace, with her hair, clothes, and body movement responding naturally to her speed. At four seconds, use a smooth cinematic tracking shot from the front as she runs between the stalls while the background moves naturally. At five seconds, switch to a dramatic side angle showing her determined expression, realistic footsteps, flowing hair, and detailed traditional clothing. At six seconds, follow her from behind as she runs deeper into the ancient street, with sunlight, dust, lanterns, and moving people creating a cinematic atmosphere. At seven seconds, smoothly move the camera around her while maintaining perfect character consistency, realistic facial features, natural skin texture, and believable motion. At eight seconds, finish with a wide cinematic shot of the Korean girl continuing forward through the beautiful historical marketplace, with photorealistic details, natural lighting, realistic physics, smooth camera movement, and no distortion.
Luxury Skincare Commercial Sequence
Cinematic luxury skincare ad. 0-3s: Woman in white silk robe holds a gold serum bottle in soft light. 3-6s: Macro of a golden oil drop falling from a glass pipette. 6-10s: Close-up of woman applying serum to her cheek with a gold tool. 10-13s: Macro of liquid ripples. 13-15s: Hero shot of the bottle on a pedestal with dappled shadows. Warm, elegant, high-end aesthetic.
THE WRONG WOMAN TO MESS WITH Action Comedy.
Ultra-realistic American live-action, grounded stunt choreography, dynamic handheld cinematography. Scene: Inside a late-night NYC subway, Character A sits calmly as a huge, arrogant man blocks two seats with his backpack and bullies a commuter. She stands and says, “Move your bag.” He towers over her: “Make me.” A fast, continuous-feeling sequence of realistic close-quarters action follows. The man reaches for her, swings clumsy punches, charges, and attempts repeated shoves. Character A skillfully dodges, pivots around subway poles, redirects his momentum, and uses controlled counters. He repeatedly stumbles into empty seats and handrails, finally landing safely on a padded bench with his backpack bouncing onto his lap. Dynamic tracking shots, tight close-ups, readable full-body choreography, realistic balance, practical stunt weight, and physical comedy. At 21 seconds, he tries to rise. She calmly raises an open hand and says, “We're done.” Embarrassed by watching commuters, he pulls his backpack onto his lap and moves over to let the displaced commuter sit. In the final three seconds, recreate the opening wide shot: she sits calmly opposite him while he behaves respectfully, leaving seats available. As the doors open, she exits without looking back. He quietly apologizes to the commuter. Visual & Audio Rules: Exactly 30 seconds, 24 fps, 16:9 landscape, consistent character identity and wardrobe, authentic NYC subway, cold fluorescent lighting, cinematic 35mm texture, subtle film grain, natural train movement, realistic reactions, and grounded American action-comedy tone. Audio only: train rumble, wheel squeals, footsteps, clothing movement, seat creaks, exertion, dialogue, and door chimes. No music, subtitles, narration, blood, graphic injuries, weapons, superhuman moves, slow motion, speed ramps, or exaggerated impacts. Keep passengers safe and subway layout consistent.
Mountain vibes, random moves, zero chill.
Cinematic Mountain Travel Montage • Funny • Cool • 15 Seconds • 16:9 Identity Lock: Use @image 1 as the master identity reference. Keep the exact same face, hair, skin tone, clothing, accessories, and body proportions throughout. Video Style: Ultra-realistic mountain travel vlog filmed by a friend using a natural handheld camera, dynamic movements, slight camera shake, and spontaneous expressions. Scenes: - 00:00–00:02: Walk through misty mountains, flip hair, and pose confidently. - 00:02–00:04: Walk backward like a fashion model, nearly lose balance, then act cool. - 00:04–00:06: Enjoy the mountain view as the wind messes up the hair; laugh naturally. - 00:06–00:08: Jog down a safe trail and strike an action-hero pose. - 00:08–00:10: Suddenly perform a funny dance and point at the camera. - 00:10–00:12: Sit on a stable rock, pose like an adventurer, then smile mischievously. - 00:12–00:15: Approach the camera, cover the lens, switch to POV, laugh, spin once, and end with an abrupt cut. Audio: Natural wind, footsteps, clothing rustle, and spontaneous laughter. No background music. Negative Prompt: No identity changes, facial distortion, stiff movements, dangerous actions, actual falls, glitches, teleportation, or CGI. Keep all movements realistic and natural. Output: 16:9 horizontal • 15 seconds • 4K UHD • 24 FPS • Ultra-realistic live action • Natural handheld camera.
Coastal Korean romantic encounter at golden hour
A beautiful Korean girl with long, straight black hair walks through a peaceful coastal Korean town during golden hour, wearing a stylish beige trench coat and carrying a handbag. A handsome Korean boy in a dark leather jacket stands near a railway crossing as a vintage tram passes by. The girl slowly approaches him along the seaside street, with warm sunlight glowing on her face. Cinematic camera movements, natural walking motion, realistic facial expressions, soft ocean breeze, and detailed Korean street scenery. Capture their first eye contact with a gentle romantic atmosphere, shallow depth of field, warm color grading, and photorealistic quality. Create a beautiful 15-second cinematic love story with smooth transitions, natural lighting, and emotional storytelling.
Zombie horror kitchen survivor scene
Create a short 16 9 photorealistic Korean cinematic zombie-horror scene in an abandoned farmhouse kitchen. A crying female survivor from Korea with messy dark hair, dried blood on her forehead, and a dirty olive-green hooded anorak aims a black pistol at her infected lover. She is trembling, tears running down her face, but cannot pull the trigger. The infected man has tangled dark hair, pale dirty skin, milky-white eyes, dark veins, torn clothing, and an increasingly feral expression. He snarls loudly and screams directly at her. She slowly lowers the pistol, drops it beside her, and opens her arms toward him. He suddenly screams again, lunges forward and violently grabs her face. Hard smash-cut to warm golden-hour memories of the same couple happily together in a vast wheat field — laughing, holding each other, walking hand-in-hand through the golden grass, and sitting together beneath the sunset. Make the contrast between the cold present and beautiful memories emotionally powerful. Photorealistic live-action cinema, natural skin and tears, realistic infected eyes and facial expressions, handheld camera, practical-effects horror, 35mm film grain, believable movement and lighting. No gunshot, no gore, no dialogue, no slow motion, no extra plot, no CGI look, no animation, no distorted anatomy, no duplicate characters, no subtitles, no text, no watermark.
Woman still amid frozen synchronized crowd
Create an ultra-realistic cinematic live-action video using the uploaded reference image as the exact visual reference for the woman, her face, hairstyle, outfit, proportions, crowd, environment, lighting, and composition. A young woman with a sleek short dark bob, fitted black turtleneck, black trousers, black shoes, and minimal jewelry walks calmly through a dense crowd. She reaches the center and stops completely still while the surrounding people suddenly begin moving in rapid, perfectly synchronized motion—turning their heads, shifting positions, raising their arms, and stepping in unison. The woman remains completely motionless and expressionless as the chaotic movement suddenly stops, leaving the entire crowd frozen mid-action. The woman remains standing perfectly still at the center of the frozen crowd, calmly looking around while every person remains locked in their exact position. The camera slowly pushes closer toward her as the silent crowd stays completely motionless in the background. She then looks directly into the lens and gives a subtle mysterious smile. Photorealistic skin, realistic anatomy, natural lighting, believable physics, shallow depth of field, authentic handheld camera movement, and natural ambient sound. No anime, cartoon, illustration, CGI look, morphing, duplicated people, distorted faces, teleportation, floating objects, exaggerated effects, subtitles, logos, or watermark.
This AI schoolgirl smashes through a subway train window in an epic acti
Create a 30-second ultra-realistic, cinematic action scene set in a crowded underground subway station. A young blonde woman with long, messy blonde hair, wearing a dark school blazer, white shirt, plaid mini skirt, and black shoes, confronts a young man with messy black hair dressed in a black school uniform. Scene 1 (0–5 seconds): The woman rushes down a tiled subway staircase, her long blonde hair moving naturally as she descends. The young man follows her, looking surprised and confused. Use dynamic handheld camera movements and dramatic cinematic lighting. Scene 2 (5–10 seconds): The woman reaches the subway platform as a silver train stands beside it. She moves quickly toward the young man while commuters watch in the background. Capture the scene with realistic motion blur, tracking shots, and wide-angle cinematic framing. Scene 3 (10–16 seconds): The two face each other near the train doors. The young man raises his hands defensively while the woman stares at him intensely. Use alternating close-up shots to capture their facial expressions, tension, and subtle body movements. Scene 4 (16–23 seconds): The woman suddenly launches into a fast, choreographed fight sequence. She delivers a powerful spinning strike, sending the young man stumbling backward and falling onto the subway platform. Her hair swings dramatically with the movement. Use fast tracking shots, dynamic camera angles, and realistic physical reactions. Scene 5 (23–27 seconds): The young man struggles to get up while the woman stands confidently beside him. Commuters react with shock and keep their distance. The silver train and tiled subway walls establish a gritty, realistic urban atmosphere. Scene 6 (27–30 seconds): The woman approaches the fallen young man and leans toward him, staring down at him with a cold, intimidating expression. He looks up at her in disbelief. End with a dramatic cinematic close-up. Visual Style: Photorealistic live-action cinematography, Korean action-drama aesthetic, realistic human anatomy, natural skin textures, detailed facial expressions, cool muted color grading, fluorescent subway lighting, subtle film grain, shallow depth of field, realistic motion blur, smooth camera transitions, and high-end movie production quality. Camera & Quality: Dynamic handheld cinematography, low-angle shots, wide tracking shots, dramatic close-ups, 4K resolution, 24 FPS, realistic physics, and seamless scene continuity. Important Consistency Rules: Maintain the same faces, hairstyles, clothing, body proportions, and character identities throughout the entire video. Ensure fluid, believable movements, natural interactions, and consistent subway station details. Avoid cartoon effects, exaggerated physics, distorted faces, extra limbs, and unnatural movements.
Village water-gun battle vlog
Create a 30-second ultra-photorealistic live-action Korean countryside lifestyle vlog showing a playful summer water-gun battle between a young Korean woman and three village children. Begin with the woman walking through a quiet Korean village when the children surprise her with a playful water-gun attack. She laughs, joins the game, and the friendly battle turns into a spontaneous chase through the sunlit village streets. Show authentic Korean countryside details including traditional homes, stone walls, gardens, trees, bicycles, narrow paved roads and warm summer sunlight. Keep all characters, clothing, water guns and environment perfectly consistent. Capture realistic water-gun pressure, fluid dynamics, gravity, air resistance, droplets, splashes, wet clothing, wet hair, running biomechanics, friction, momentum and natural environmental reactions. Water should behave like real pressurized liquid rather than artificial beams. Use modern 2026 handheld cinematic-vlog cinematography with dynamic tracking shots, close-ups, over-the-shoulder angles, realistic autofocus, natural camera inertia and occasional wide village views. Natural diegetic audio only: children's laughter, footsteps, water-gun sounds, water impacts, leaves, birds, insects and authentic Korean reactions. No music, narration or subtitles. End with everyone completely soaked and laughing together as sunlight catches the falling water droplets in the peaceful village street.
Never Underestimate Her Kung Fu Skills! Made with Seedance 2.5 on picsa
Create a 10-second ultra-realistic cinematic martial arts action video featuring a skilled young Asian female warrior with black hair tied in a high ponytail, wearing a traditional white martial arts uniform with black trim, a black waist sash, and black training pants. The scene takes place in a traditional ancient Chinese courtyard with tiled rooftops, stone flooring, and bright natural sunlight under a clear blue sky. Scene 1 (0–2 seconds): A wide cinematic shot shows the female martial artist standing confidently in the courtyard as a black-clad opponent approaches. Use a low-angle camera and a slow forward dolly movement. Scene 2 (2–4 seconds): The opponent attacks with a punch. She swiftly blocks the strike, pivots her body, and counters with precise martial arts movements. Use dynamic handheld camera movement and realistic motion blur. Scene 3 (4–6 seconds): Close-up of her focused face as she dodges another attack and delivers a powerful counterstrike. Capture realistic facial expressions, flowing hair, fabric movement, and detailed fight choreography. Scene 4 (6–8 seconds): Two opponents attack from opposite sides. She performs a fast defensive spin, blocks both attacks, and forces them backward with fluid kung fu techniques. Use a dynamic tracking shot with cinematic action timing. Scene 5 (8–10 seconds): She lands in a powerful low martial arts stance, one palm extended toward the camera, staring intensely at her remaining opponents. Dust rises gently from the stone floor as the camera slowly pushes in. Visual style: Photorealistic, high-end martial arts movie cinematography, 4K detail, natural sunlight, realistic skin texture, accurate human anatomy, smooth movements, authentic kung fu choreography, dramatic camera angles, shallow depth of field, cinematic color grading, crisp details, vertical 9:16 aspect ratio. Audio: Natural courtyard ambience, fast whooshing sounds during punches, sharp impact effects, footsteps on stone, and subtle cinematic background music. Negative prompt: Cartoon, anime, CGI appearance, unrealistic movements, distorted faces, extra fingers, extra limbs, duplicated characters, floating bodies, flickering, inconsistent clothing, poor anatomy, low resolution, excessive motion blur, text, subtitles, watermark.
When you wake up late and panic mode activates!
Create a 10-second stylized, ultra-detailed 3D animated comedy sequence featuring a young man with messy dark hair, big expressive blue eyes, and a gray T-shirt in a cozy bedroom and bathroom. Start with him sleeping peacefully in bed under a blanket → close-up as he suddenly wakes up in shock, eyes wide and mouth open, seeing the alarm clock showing 7:30 AM → he leaps out of bed in a hilarious panic and rushes across the sunlit bedroom with exaggerated cartoon movements → cut to the bathroom, where he frantically brushes his teeth with a blue toothbrush, foam spreading around his mouth → close-up of his confused, worried expression as he pauses mid-brush → end with him looking toward the camera and breaking into a big goofy smile, toothpaste foam still around his mouth. Warm morning sunlight through the window, cozy home interior, expressive facial animation, exaggerated comedic timing, realistic cloth movement, smooth camera transitions, detailed 3D character rendering, soft shadows, cinematic composition, family-friendly animated movie aesthetic. Maintain consistent character design, face, hairstyle, and clothing throughout. No text overlays, subtitles, logos, or watermark.
Rainy morning forgotten jacket MiniDV
15-sec MiniDV — Forgot Her Jacket Main Subject: Young Korean woman, 24, naturally attractive, realistic skin, minimal makeup, long dark hair loosely down. Wearing a simple oversized faded lavender T-shirt, loose gray lounge pants, white socks and a thin silver necklace. Preserve her exact identity, facial features, hairstyle, body proportions and appearance throughout. Location: Small old Seoul apartment during a dark, rainy early morning. Bedroom leading into a small entryway, wooden shoe rack, umbrella stand, canvas shoulder bag and rain-covered apartment door. Heavy rain visible outside. Lighting & Atmosphere: Deep early-morning blue-hour ambience. Cold blue-gray light from the rainy window and doorway mixed with a weak warm interior light. Muted colors, soft shadows, slightly underexposed surroundings, cozy but chilly atmosphere. Style: Ultra-realistic early-2000s Sony MiniDV home video filmed by another person holding the camcorder. Completely candid and unstaged. Natural handheld movement, subtle human shake, imperfect framing, gentle reframing, occasional autofocus hunting, mild exposure shifts, faded colors, soft contrast, authentic DV compression, subtle low-light digital noise and microphone noise. Smooth continuous real-time motion throughout. No stuttering, judder, frame skipping, duplicated frames, stop-motion appearance, excessive motion blur, speed changes or low-frame-rate look. No stabilization or modern cinematic movement. 00:00–00:03: She walks sleepily from her bedroom toward the apartment door, carrying her canvas shoulder bag. She slips on her sneakers and reaches for the door. 00:03–00:06: She opens the door and immediately feels the cold rainy air. She pauses, instinctively hugs her arms around herself and looks outside at the heavy rain. 00:06–00:09: She looks down at her thin T-shirt, then back outside. Her expression changes to a small realization: she forgot her jacket. 00:09–00:12: She quickly turns around and walks back toward her bedroom. The camera casually follows her with slightly imperfect handheld framing. 00:12–00:15: She grabs an oversized jacket from a chair, throws it on and pulls it around herself. She looks toward the camera with a small relieved smile, then heads back toward the door. Audio: Natural sound only—steady rain, door opening, distant traffic, footsteps, fabric movement, quiet apartment ambience and faint wind. No music, narration or added sound effects. Goal: A tiny, relatable rainy-morning moment that feels accidentally captured on an old family MiniDV camera. The humor comes from her realizing she dressed completely wrong for the weather, followed by the cozy relief of putting on her jacket. Motion quality: Keep every action smooth, continuous and physically realistic. No rushed or exaggerated movements. The handheld camera should move naturally with the person filming. The vintage look must come from authentic MiniDV texture, autofocus, exposure behavior and muted colors—not from a low or inconsistent frame rate.
Seoul day-in-life lifestyle montage
Create a 30-second photorealistic, highly entertaining Korean Seoul lifestyle video featuring the SAME adult Korean woman throughout. Keep her face, hairstyle, outfit, accessories and overall appearance perfectly consistent in every shot. Character: Adult woman with long soft dark hair, natural realistic skin, fresh minimal makeup, expressive eyes and genuinely playful facial expressions. 0–3 sec: She opens her Seoul apartment door confidently, takes one step outside, suddenly realizes she forgot something, freezes with a funny expression, turns around and runs back inside. Fast comedic camera push-in + whip transition. 3–7 sec: She comes running back outside and catches the Seoul bus at the last second. She looks at the camera with a tiny proud smile. Dynamic tracking shot. 7–11 sec: Seoul convenience store she grabs iced coffee, a cute dessert and snacks. The dessert nearly slips from her hand, but she catches it perfectly and smiles. 11–15 sec: Subway ride through Seoul. She enjoys her dessert by the window, watches the city passing by and makes a playful expression at her reflection. 15–20 sec: Lively Korean street-food market. She tries spicy tteokbokki, reacts dramatically to the heat, laughs and grabs water while her friends tease her. 20–25 sec: Golden-hour Han River. She and her friends ride bicycles together, playfully racing and laughing. Smooth side-tracking camera. 25–30 sec: Nighttime rooftop convenience-store seating. They share snacks and dessert while Seoul glows behind them. Camera smoothly pulls backward and upward, revealing the beautiful Seoul skyline. Style: Premium Korean lifestyle commercial × cinematic vlog, cute, playful, energetic, natural expressions, smooth camera movement, realistic Seoul locations, beautiful lighting, cinematic daylight-to-night transition, photorealistic 16:9 landscape
When the Thunder Emperor Awakens
DURATION: 30 seconds | FORMAT: horizontal 16:9 STYLE: Ultra-photorealistic dark fantasy combat, 8K HDR, 24fps. CHARACTER LOCK — ABSOLUTE PRIORITY image1 = Only reference for the adult MALE Jiangshi protagonist. Preserve his exact male face, hairstyle, outfit, physique, and accessories. Never change his gender, face, or identity. image2 = Only reference for consistent enemy monsters. ENVIRONMENT Vast stormy green battlefield, thousands of monsters, gigantic demons, ruined Taoist temples, shattered altars, navy storm clouds, cyan-white lightning, golden talismans, and black demonic energy. Never show an empty battlefield. 0–5s — LIGHTNING DASH The male Jiangshi dodges monster claws, slides across grass, and throws explosive golden talismans. Lightning-powered punches and kicks destroy enemies. He says: "비켜라." 5–10s — THUNDERBLADE TORNADO TECHNIQUE 1: 뇌부천쇄 · 천뢰풍폭 Thousands of lightning blades form a gigantic golden-cyan tornado, tearing through monster armies and destroying stone altars. He charges forward, shouting: "천뢰를 받아라!" 10–15s — THUNDERSTORM TECHNIQUE 2: 구천뇌우 · 만뢰천폭 He slams a golden talisman into the ground. A colossal Bagua formation appears overhead, unleashing thousands of lightning bolts and three interconnected thunder tornadoes. He shouts: "하늘이여, 쓸어버려라!" 15–20s — LIGHTNING PRISON TECHNIQUE 3: 팔괘천뢰진 · 팔방봉멸 Eight enormous golden sealing pillars rise, connecting sky and ground with cyan lightning. A gigantic octagonal prison traps thousands of monsters in crushing electrical chains. He shouts: "뇌문, 개방!" 20–25s — DEMON KING The formations explode in a massive blue-white shockwave. A mountain-sized demon king emerges from black storm energy. The protagonist dodges, attacks with lightning-powered palms, and launches aerial counterattacks before being blasted backward. 25–30s — FINAL ULTIMATE FINAL TECHNIQUE: 구천뇌제 · 만뢰천벌 He raises a golden talisman as thousands of lightning bolts converge around him. He says: "구천의 벼락이…" then shouts: "네놈을 멸하리라!" A colossal golden-cyan Bagua formation fills the sky. Thousands of gigantic lightning bolts merge into a devastating thunder pillar, engulfing the demon king in an enormous explosion. FINAL FRAME: The recognizable male protagonist stands in the foreground as lightning illuminates the monster army and ruined temples. CAMERA: Dynamic tracking, aggressive whip-pans, fast combat cuts, dramatic low angles, realistic impact shake, continuous battlefield action. NEGATIVE: No female characters, gender swapping, face morphing, identity changes, outfit changes, inconsistent monsters, cartoon CGI, teleportation, empty backgrounds, logos, subtitles, or watermarks.
Physics lab experiment cinematic video
Create a cinematic, realistic educational video set inside a modern university physics laboratory. A confident male physics professor walks into the classroom carrying a coffee cup and a notebook, while students sit attentively in a large lecture hall. The scene then transitions to a hands-on physics experiment where the professor and a group of students gather around a transparent inclined track. A small metallic ball is carefully placed on the track, creating a clear visual demonstration of motion, gravity, acceleration, and energy. Show detailed close-up shots of the transparent rail, metal ball, students observing closely, and the professor guiding the experiment with natural hand movements. Use smooth camera movements, realistic depth of field, natural facial expressions, and authentic classroom interactions to make the experiment feel engaging and believable. Include cinematic laboratory lighting with soft indoor illumination and warm sunlight entering through large windows. End with the professor working at a laboratory table during golden-hour lighting, surrounded by scientific equipment, creating a thoughtful and inspiring academic atmosphere.
Detective investigates case board at night
30-second cinematic procedural mystery sequence. Character A, matching the reference face/outfit throughout, investigates a dim precinct office at night. 0–3s: She pins a photo onto a cluttered case board covered in evidence and tangled red string. 3–6s: She studies an old grainy photograph through a magnifying glass; a blurred background figure looks strangely familiar. 6–9s: Her phone buzzes. She ignores it, comparing handwriting samples and noticing a suspicious match. 9–12s: She cross-checks two old documents and discovers a timestamp discrepancy. 12–15s: Her phone rings; she answers, listening intensely as new information changes everything. 15–18s: She turns sharply toward the board, spotting two previously unrelated photographs. 18–21s: She connects them with red string, revealing a hidden pattern and a familiar face among the evidence. 21–24s: Her eyes widen as she realizes the culprit has been in plain sight all along. 24–27s: She urgently grabs her phone and jacket and rushes out, leaving the illuminated board behind. 27–30s: Slow push-in on the newly connected evidence and overlooked photograph, ending on a quiet, ominous procedural-mystery reveal.
Anime warriors clash in holographic arena
Cinematic high-end 3D anime action, 30 seconds, 24fps, ultra sharp, no watermark, no logo, no text overlay.Two fighters in a vast teal holographic arena: glowing grid floor, fog, tall white marble pillars, a single vertical cyan light beam in the center.Girl: silver-lavender hair with a pink bow, fierce cute face, white tech sports outfit with a small cat graphic and teal trim, white cyber boots with pink/cyan lights. She dual-wields a massive white greatsword with glowing cyan circuit lines.Opponent: fox-eared white-and-gold cracked helmet with one glowing red eye, long black-and-white hair, green varsity jacket over a white crop top, brown plaid mini skirt, white mechanical legs, jet thrusters on the boots. She fights with a flaming gold katana. 0–4s: low-angle start. Girl charges across the wet teal grid, sword huge in foreground, cyan foot trails. They clash — cyan blade vs gold fire katana, sparks and orange embers. 4–8s: close-up lock. Blades grind, helmet visor cracks, cyan targeting grid flashes across both faces. Girl snarls. Opponent jet-jumps overhead, spinning with a flaming slash. 8–14s: they leap onto two facing marble pillars, then sprint and kick off the same giant white pillar mid-air. Speed-line camera. They land on one cube platform and trade brutal punches and sword hits, motion blur, sparks. 14–20s: both drop and sprint low toward camera, blades dragging the floor, a bright spark line between them. Smash cut to mid-air chase around the cyan beam — girl running with the long sword, opponent flipping above with the katana. 20–25s: brief monochrome flash, then they race side by side through rain-streak motion. High kick from the girl hits the helmet in a gold-white burst. Wide shot: both stand on opposite glowing rings, weapons raised, gold sparkles falling. 25–30s ENDING (same as original): overhead crane shot as they dash into the center. Floor cracks like glass. Extreme close clash — cyan vs gold, shards flying, visor splitting. White-out flash. Arena explodes outward: marble cubes, teal glass, smoke and a huge gold-white blast filling the frame. No logo. Hold the explosion.
Tokyo train supernatural horror sequence
Create a 30-second ultra-realistic live-action Japanese supernatural thriller set inside a nearly empty Tokyo underground train station late at night 0–5 seconds: A young Japanese woman runs down the empty platform while repeatedly looking behind her. The last train is about to leave. She reaches the train just before the doors close 5–10 seconds: Inside the almost empty train, only three passengers are sitting silently. The woman catches her breath and looks through the window The train starts moving Suddenly, every passenger turns their head toward her at exactly the same time She freezes 10–16 seconds: The lights inside the train begin flickering. The digital station display changes rapidly between different station names before showing a station that does not exist The train slows down The woman looks outside The train is passing through a completely dark underground tunnel filled with hundreds of motionless people standing beside the tracks 16–22 seconds: The train suddenly stops All three passengers slowly stand up and walk toward her without saying anything She backs away and desperately presses the emergency button Nothing happens Then one passenger quietly points behind her 22–27 seconds: She turns around The train doors are now open But instead of the platform, there is an identical train directly outside Inside that train, she sees herself sitting alone Her reflection slowly looks up and smiles 27–30 seconds: The woman steps backward in terror The doors suddenly close The other passengers disappear The train begins moving again She looks at the window Her reflection is no longer there Cut to black Photorealistic live-action Japanese supernatural horror, authentic Tokyo underground station, realistic modern train interior, natural Japanese characters, cinematic handheld camera, smooth tracking shots, realistic train movement, practical lighting, atmospheric fluorescent flickering, subtle supernatural effects, realistic facial expressions, escalating tension, strong visual storytelling, detailed environments, realistic reflections, diegetic train sounds only, no music, no narration, no excessive gore, no graphic injuries, no animation, no anime, no cartoon, no artificial CGI look, no distorted anatomy, no subtitles, no random text, no logos, no watermark