AI Video Prompts for Dance

Ballet grand jetés, hip-hop floor work, contemporary partnering, and breakdance power moves — these Seedance 2.0 dance prompts give the model a choreographer's direction and a cinematographer's brief so the movement and the camera make sense of each other at the same time.

Dance video has one rule that every other video genre either borrows or ignores: the camera must understand what the movement means before it can decide where to be. A ballet arabesque means something at full body — the extended leg, the lifted arm, the turned-out hip, the weight balanced on a single point — and it means something different in close-up on the expression in the peak of the extension. A hip-hop isolation means nothing in wide and everything in medium close-up where the precise muscle control is visible. The camera distance is not a neutral choice; it is a reading of the movement, and that reading has to match the genre's grammar. The most common error in AI dance video prompting is describing the movement without describing the camera relationship that makes the movement legible. "A ballet dancer performing" gives the model a generic wide shot of someone in a tutu. "A ballerina in a single-point arabesque, camera holding a three-quarter rear angle at hip height to show the full line from standing foot through fingertips, stage backlight separating the figure from the dark flat behind her, the position held for three counts before the controlled descent" gives the model a specific camera position, a specific light source, and a specific duration — and those three specifications together produce a dance frame that communicates technique. There are four movement registers in dance video, and each requires a different camera and light strategy. Classical ballet uses architecture and line as its primary visual language: the geometry of the body against empty space. The camera distances that serve classical ballet are wide enough to show the full body geometry, angled to reveal the most extreme extension of the held position (typically three-quarter front or rear), and positioned at hip-to-waist height to show both the foot placement and the upper body geometry simultaneously. Light in classical ballet is theatrical and directional — stage backlight that separates the dancer from the flat behind them, front-fill that keeps the face legible, and a hard side key that casts a shadow of the position on the floor, doubling the line. The one exception is the extreme close-up used in documentary and behind-the-scenes work: the face during a pirouette series, the hands during a port de bras, the feet during petit allégro — these exist to reveal the technical precision that the wide stage frame cannot show. Hip-hop and street dance operate from the floor up. The visual grammar is low-angle — the camera below the dancer's center of gravity looking up through the movement, making the trick or combination read against the sky or the ceiling rather than against a background at the same height. This low angle works for a different reason than it does in action video: in hip-hop, the low-angle is not about power, it is about reading the footwork. Footwork is the primary text of hip-hop; the beat drops, the feet respond. A mid-height camera misses the footwork entirely because the feet are obscured by the torso. A floor-level or near-floor camera positioned in front of and below the dancer shows the feet, the low-level moves (windmill, six-step, flare), and the vertical dimension of uprocking and freezes all at once. Light in hip-hop video is either hard and practical (street lights, warehouse fluorescents, the hard shadow of a single overhead source) or dramatic and theatrical (RGB stage washes that color the floor and the dancer separately). Either way, the shadow becomes a second dancer in the frame. Contemporary and modern dance requires the camera to understand the movement from inside its logic rather than observing it from a fixed documentary distance. Contemporary choreography is about weight — the way weight shifts through a transition, falls into the floor and rebounds, accumulates in a drop and releases into an extension. The camera that best captures contemporary movement is one that moves with or in counterpoint to the dancer: a slow push-in as the dancer settles into the floor, a pull-back as the movement expands into the space, a lateral track that stays with the dancer's center through a traveling phrase. Light in contemporary is often working against the camera: side light that creates a strong shadow structure on the body and reveals the weight transfers as plane changes, or natural light from a large window that treats the studio as an architectural space the dancer is inhabiting rather than a stage they are performing on. Breakdance and power moves exist in their own visual grammar where the floor is the stage and the camera on the floor is the ticket into the vocabulary. Spinning on the head, balancing in a freeze on one hand, the windmill's continuous floor momentum — these are incomprehensible from a standing-height camera because the spatial reference is wrong. Place the camera at floor level facing the dancer, and the windmill becomes a continuous circular sweep that fills the frame; the headpin becomes a vertical spin with the floor as the anchoring plane; the freeze becomes a held position at the camera's own eye level where the balance reads as a deliberate architectural hold rather than a falling catch. Slow motion is the other essential tool for power moves: a normal-speed windmill or aerial reads as a blur, but 240fps reveals the precise mechanics of each phase — the entry, the rotation, the exit — that the dancer has spent thousands of hours perfecting. Across all four registers, the pre-beat moment is the overlooked structural element that separates dance video that builds tension from dance video that merely records movement. Every major trick or combination has a preparation: the demi-plié before the grand jeté, the crouch before the aerial, the slight drop of the center before the flare. The preparation is where the audience's breath catches — and the camera that is already framed for the trick before the trick happens gives the viewer the preparation and the execution in a single continuous frame. This is why specifying "beginning from the preparation before the movement" in a dance prompt produces more satisfying video than starting from the peak: the preparation is the grammar; the peak is the punctuation; the resolution is the sentence.