OpenZhenZhen AI Prompt Gallery · MiniMax H3 · video

High Knees

High Knees — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

High Knees — AI video creation
High Knees

Prompt

English prompt

subject_definitions: <Subject 1> is the woman in <Picture 1>: East Asian, late twenties, very long straight black hair falling past her waist with a heavy centre part, a slim oval face, dark red lipstick and softly defined brows. She wears a navy long-sleeve bodysuit with a heart-shaped cutout high on the chest, gathered at the left waist, and sheer black tights printed with small evenly spaced black polka dots, on black pointed-toe stiletto heels. <Picture 1> shows the end of her movement, not the beginning: it is the finished side split with one leg raised overhead and held at the calf by her hand. She begins the segment standing on both feet with the leg down. Her face stays composed while her breathing does not, her supporting knee never locks, and her raised leg carries a fine tremor at the limit. <Subject 2> is the hotel dining room in <Picture 1>: a large pale marble floor with grey veining and visible tile seams, a round table in white marble behind her, deep green velvet dining chairs with dark tapered legs and brass tips, a tall crystal chandelier of clustered glass shades overhead, a white floral centrepiece, wine glasses, and a wide window on the left in pale sheer curtains. Warm cove lighting runs along the ceiling edge and the far wall is a flat warm grey. summary: [reference generation] A 15-second 9:16 live-action segment built from <Subject 1> and <Subject 2>, in which she stands at first on both feet in the dining room, is then already halfway into the split with the leg at her waist, breathes harder than she wants to while her face stays composed, and finally arrives at the raised position in <Picture 1> with the leg trembling at its limit. No dialogue. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3], [Shot 4]): fully_preserved - her long black hair, navy heart-cutout bodysuit, polka-dot sheer tights, black pointed stilettos and dark red lips are retained, along with the finished raised-leg position of the reference picture, which is reached only in the final shot. <Subject 2> (appears in [Shot 1], [Shot 2], [Shot 3], [Shot 4]): partially_preserved - the pale veined marble floor with its tile seams, the round marble table, the green velvet chairs, the clustered crystal chandelier and the sheer-curtained window are retained, while the camera moves in closer and the framing changes across the four shots. detailed_description: Live-action, cinematic, bright soft daylight from the window on the left mixed with warm light from the ceiling coves, clean high-key exposure, shallow depth of field and the faintest film grain, all four shots on a static camera with no reframing, no slow motion and no speed ramps. The body behaves the way a body behaves when it does not know it is being filmed: weight never fully settles, small corrections happen mid-gesture, and nothing arrives exactly on the beat it was heading for. [Shot 1] A medium shot from slightly below eye level, the veined marble floor filling the lower half of the frame and the green velvet chairs soft behind her. <Subject 1> is standing on both feet with the leg down, her weight still rocking between them, the toe of her right shoe resting a little ahead of the left. Her left hand hangs at her thigh with the fingertips against the dotted fabric, and her right arm lifts a little as if to settle her hair and then simply does not continue, coming back down. Her eyes are on a point on the floor a few steps away from her, not on the lens. [Shot 2] At 00:04.000, the camera cuts to a medium shot from her left front, looking slightly up the line of the raised leg. The leg is already at waist height and still travelling, the knee still softly bent rather than snapped straight, her right hand closed around her calf just above the ankle, the heel of her standing shoe pressed into the same tile seam it started on. The fabric at her waist is pulling taut out of its gathers. Her free arm goes out to the side to find balance, not yet fully extended. She takes one breath in through the nose and does not let it all the way out. [Shot 3] At 00:08.000, the camera cuts to a medium close shot framed from the collarbone up, the chandelier dissolved into soft bright shapes behind her head. Her mouth is closed and her jaw is held, the face doing its work of looking easy while her breath keeps arriving faster than that, the base of her throat moving with it. Her eyes come to the lens a moment before her face squares to it. Her hair has not moved much at all. [Shot 4] At 00:12.000, the camera cuts to a medium shot from low behind her left shoulder, her body filling the frame in profile with the raised leg running up out of the top edge, the room falling away behind her. She is at the finished position of <Picture 1>, the bodysuit now stretched smooth across her hip where it was gathered before, her standing knee still unlocked, her heel still on that seam. The raised leg carries a fine tremor she is not correcting, and her chest lets the held breath go slowly through the last of the shot. overall_soundscape: A large quiet room with a faint air-conditioning hum and the soft tail of city sound behind the window glass. Her shoe leather creaks once on the marble as her weight moves, the dotted fabric drags against her palm, and from the second shot onward her breathing is the loudest thing in the room, quick through the nose, held, then released long at the end. non_diegetic_music: A single sustained low string note with no rhythm and no build, entering under the first shot and thinning to nothing in the last three seconds so that only her breathing is left. Negative prompt: subtitles, captions, text, words, letters, descriptions, labels, annotations, watermark, signature, writing on screen

Topics and styles

View API model

Explore more AI image and video prompts · Try image generation · Try video generation · Create on the AI canvas

  • Close the Door

    Close the Door — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

  • Forest of Green Light

    Forest of Green Light — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

  • Green Light Forest

    Green Light Forest — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

  • Four Views

    Four Views — explore this zhenzhen-image-g-v2.6-lowprice creation on OpenZhenZhen, with its original prompt and a reusable generation example.

  • Exhibition

    Exhibition — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

  • Self-Portrait

    Self-Portrait — explore this MiniMax H3 creation on OpenZhenZhen, with its original prompt and a reusable generation example.