Camera Movement in AI Video: Pan, Tilt, Dolly, Truck, Crane, Handheld

The pan rotates the camera horizontally on a fixed axis — left or right. It scans an environment, follows a moving subject, or transitions between two points in the same plane. Slow pans feel contemplative and surveying; fast pans (whip pans) feel kinetic and editorial.

Prompt examples:

• "slow pan left across a moonlit mountain ridge, cinematic, 4K"
• "fast whip pan right, tracking a motorcyclist through neon-lit streets, motion blur, night"

The tilt rotates the camera vertically on a fixed axis — up or down. A tilt-up from feet to face reveals a character dramatically. A tilt-down from a skyscraper’s peak to its base establishes scale. In AI video, tilt is highly effective for product reveals and architectural showcases.

Prompt example: "slow tilt up from cobblestone ground to gothic tower, dawn light, misty atmosphere, cinematic".

The dolly physically moves the camera forward (dolly-in) or backward (dolly-out) along a track. Unlike a zoom, a dolly changes the spatial relationship between subject and background — the background shifts in parallax, creating a three-dimensional sense of moving through space.

Dolly-in increases emotional intensity and intimacy: "slow dolly-in on a woman standing at a window in rain, soft backlight, 50mm, melancholic".

Dolly-out (or pull-back) creates distance, isolation, or reveals — the subject stays the same size while the world around them expands: "dramatic dolly-out revealing a lone figure in a vast desert, golden hour, aerial perspective".

The dolly zoom (Vertigo effect) simultaneously dollies in while zooming out — or vice versa — so the subject stays the same size while the background compresses or expands behind them. It is a powerful disorientation tool: "dolly zoom effect, character standing on rooftop, background buildings stretching, suspense".

The truck (also called tracking shot or crab) moves the camera laterally — left or right — while keeping it aimed in the same direction. It follows a moving subject at a constant distance, maintains a character’s eye line while the background flows past, or slides past a series of objects for a reveal.

Prompt example: "truck right tracking a dancer moving through a studio, natural window light, continuous motion". In AI video, trucking shots work beautifully for fashion, architecture walkthroughs, and product showcases where you want the camera to glide alongside the subject.

The pedestal moves the camera straight up or down without tilting — the entire camera rises or sinks on its vertical axis. This is different from a tilt, which pivots. A pedestal-up floats the camera above a ground-level scene; a pedestal-down sinks it. Prompt example: "slow camera pedestal-up rising from street level to rooftop view, city at night, neon reflections".

The crane shot uses a physical crane arm to sweep the camera through arcs that combine horizontal, vertical, and directional movement simultaneously — soaring up and over a landscape, swooping down toward a subject, or rising from ground level into the sky.

Prompt example: "crane shot rising up and over a medieval castle at dawn, wide lens, cinematic scope, fog in the valley below". Crane movements imply production value and are excellent for opening title sequences, climactic reveals, or emotional summits.

The drone shot (aerial footage) achieves similar movement through a different tool. In AI video prompts, the distinction matters: drone shots imply a modern, real-world aesthetic; crane shots imply a controlled, often more cinematic compositional intent.

Prompt example: "aerial drone shot banking left over coastal cliffs, turquoise ocean, sunset, wide angle". For Floniks AI Video, aerial movement prompts typically require explicit directional language: "camera sweeping upward and forward," "banking left while descending," "orbiting clockwise around the subject."

Handheld camera movement introduces organic, irregular motion that signals authenticity, spontaneity, and urgency. Used in documentary, cinema vérité, war films, and modern drama (think Children of Men or Saving Private Ryan), handheld shooting feels "unmediated" — as if the audience is really there.

Prompt examples:

• "handheld medium shot, journalist running through a crowded protest, kinetic energy, natural daylight"
• "handheld close-up, couple in argument, tight quarters, practical lamp lighting, raw and intimate"

In AI video, handheld prompts sometimes need reinforcing language: "shaky cam," "unstabilized camera," "cinema vérité style," or "natural camera shake" alongside "handheld." Without reinforcement, the model may produce a stabilized shot despite the instruction. Contrast handheld with "smooth gimbal movement" when you want mechanical stability with organic motivation.

Single camera movements are powerful; sequences of movements build cinematic language. A two-shot sequence might open with a wide crane shot descending toward a city street (establishing scale and location) then cut to a handheld medium shot following a character through the crowd (intimacy and energy). Each clip in the sequence is generated separately in Floniks AI Video with its own movement prompt, then assembled.

The Floniks Editor (visual workflow tool) is designed for exactly this pattern. Create a workflow with two or more AI Video nodes in sequence — each with its own movement specification, lighting direction, and style tokens. Connect them with a sequence node that defines the transition. This gives you frame-level control over each movement clip rather than hoping a single long prompt generates a complex compound movement.

Advanced tip: when chaining a dolly-in to a pan, generate the dolly-in with the ending frame’s composition in mind (where the pan will begin). Consistent subject position at the clip boundary makes the cut feel motivated rather than arbitrary.