Dolly & Track
Dolly Shot
A dolly shot moves the entire camera toward or away from the subject on wheels or rails, which is the key difference from a zoom: the perspective and depth actually shift as the camera travels, instead of the lens just magnifying a flat image. A dolly in (push in) tightens the frame from a medium shot to a close-up and is the go-to move for building tension or emotional weight; a dolly out (pull out) starts close and pulls back to reveal the wider scene, often used as a reveal or an ending beat. Keep the speed constant and the framing locked on the subject for a clean, professional result.
Tracking Shot
A tracking shot moves the camera parallel to a moving subject, matching its speed so the subject stays roughly the same size and position in frame while the background slides past behind it. It's the standard move for a walking conversation, a car following another car, or a character being led through a space, because it keeps the audience moving with the subject instead of watching from a fixed point. A side tracking shot (sometimes called a truck shot) shifts left or right; a front or rear tracking shot leads or follows directly. The camera path should stay smooth and level for the effect to read as intentional rather than shaky handheld footage.
Pan & Tilt
Pan Shot
A pan shot pivots the camera horizontally from a fixed position, the way your head turns to look from one side of a room to the other, without the camera itself changing location. It's the simplest way to reveal a wide environment, follow a subject moving across the frame, or connect two separate points of interest in a single unbroken shot. A slow pan feels observational and calm; a fast pan raises energy and is often used to whip attention from one subject to another. Keep the rotation speed even from start to end so the motion reads as controlled rather than accidental.
Tilt Shot
A tilt shot pivots the camera vertically from a single fixed position, up or down, without the camera moving through space. Tilting up a subject's body toward their face is a classic reveal move that builds anticipation before showing who they are; tilting down from the top of a building or landscape to ground level establishes height and scale in one continuous motion. Because the camera doesn't travel, a tilt is easy to keep steady, and pairing it with a slow, even speed reads as deliberate and cinematic rather than an accidental camera wobble.
Zoom
Zoom Shot
A zoom shot changes the lens's focal length to make the subject appear closer or farther, while the physical camera position never moves, which is what separates it from a dolly shot. Because the perspective and depth relationships in the frame don't change, a zoom reads as flatter and more compressed than a dolly, which is exactly why some directors prefer it for a deliberately artificial, voyeuristic, or surveillance-camera feel. Zooming in draws attention to a detail fast; zooming out reveals context. For AI video prompts, describing it as a lens-only change (not camera translation) is the detail that keeps the model from generating a dolly instead.
Dolly Zoom
The dolly zoom (also called the vertigo effect, after Alfred Hitchcock's Vertigo) combines two opposite moves at once: the camera physically dollies backward while the lens zooms in, or dollies in while the lens zooms out, timed so the subject stays the same size throughout. The trick is what happens around the subject: the background stretches or compresses dramatically, creating an unmistakable feeling of vertigo, dread, or a sudden realization. It's a harder move to describe to an AI video model than a plain dolly or zoom, so naming both movements explicitly and specifying that the subject's size must stay constant is what makes the effect land.
Crash Zoom
A crash zoom pushes the zoom shot to its most extreme: instead of a smooth, gradual change in focal length, the lens snaps from wide to tight (or tight to wide) almost instantly, landing hard on the subject with a jolt. It's the go-to move for a shock beat, a comedic punchline, or a sudden reveal, because the abruptness itself carries the emotional hit. Unlike a slow zoom, which builds tension gradually, a crash zoom skips the buildup entirely, so pairing it with a hard stop and a beat of stillness right after the snap is what sells the effect.
Drone & Crane
Drone Shot
A drone shot (or aerial shot) is captured from a flying camera, giving it freedom of movement no ground-based rig has: it can rise straight up, sweep forward over a landscape, or orbit a subject from above. It's the standard opener for establishing a location, whether that's a coastline, a city skyline, or a single building, because the height instantly communicates scale in a way eye-level footage can't. For AI video prompts, specifying the altitude, the direction of travel, and how much of the scene should be visible below gives the clearest results, since 'drone shot' alone leaves the model guessing at height and speed.
Crane Shot
A crane shot mounts the camera on a mechanical boom arm, letting it rise or descend smoothly through vertical space, often while also moving forward or backward at the same time, in a way that feels more controlled and weighty than a drone. Craning up and away from a scene is a classic ending move, giving the audience a final god's-eye perspective as the shot pulls back; craning down into a scene, from above a crowd or building down to a single character, works as a dramatic introduction. The combination of vertical and horizontal movement in one continuous, unbroken shot is what makes a crane shot instantly recognizable.
Arc / Orbit Shot
An arc shot moves the camera along a curved path around a subject that stays roughly still, always keeping the lens pointed inward at the same point. A short arc (sometimes called a partial orbit) reveals a new angle of the subject over the course of the move, which is a common way to make a static product or a posed subject feel dynamic without them doing anything. Push the arc all the way to 360 degrees and it becomes a full orbit shot, a move often used for hero reveals, since it turns the subject into the visual center of the frame while the whole environment rotates around them.
Handheld & Rig
Handheld Shot
A handheld shot is filmed with the camera carried in the operator's hands rather than mounted on a stabilizer, which introduces small, organic shake and drift that reads as immediate, urgent, or documentary-style rather than polished. Directors reach for handheld specifically to break the smoothness of a scene, whether that's for a chaotic action beat, a tense conversation, or footage meant to feel like it was captured in the moment rather than staged. The amount of shake matters: too little and it just looks like an unstabilized tripod, too much and it reads as an error, so AI video prompts should specify subtle, controlled shake rather than extreme camera shake.
Gimbal / Steadicam Shot
A gimbal shot (or Steadicam shot, named after the original mechanical rig) uses motorized or counterweighted stabilization to cancel out the operator's footsteps and body movement, so the camera can travel through a space, walk up stairs, or turn corners while staying perfectly smooth and level. It's the opposite intent of a handheld shot: both let the camera move freely through a scene, but a gimbal shot removes the shake entirely for a floating, almost weightless feel, which is why it's the standard choice for long, flowing walk-and-talk scenes or fluid action sequences that need to stay legible.
Static Shot
A static shot is a completely locked-off frame, mounted on a tripod with no pan, tilt, dolly, or zoom of any kind, so every bit of visual interest has to come from what's happening inside the frame rather than from camera motion. It's an underrated choice: a static shot gives a scene stillness and formality, draws attention to composition and blocking, and lets subtle performance or environmental detail read clearly without a moving camera competing for attention. For AI video prompts, explicitly stating that the camera must not move (no pan, no zoom, no drift) is often necessary, since many models default to adding subtle camera motion even when a still frame wasn't requested.
Body-Mounted Camera (Snorricam)
A body-mounted camera shot (often called a Snorricam, after the rig that popularized it) attaches the camera directly to the actor's body, usually the chest, instead of being carried by a separate operator. The result is deeply disorienting: the subject's face stays locked in the same position in frame no matter how they move, while the background spins, tilts, and lurches behind them with every step. It's a favorite for depicting anxiety, intoxication, or a breakdown, since it isolates the subject from a world that suddenly feels unstable around them.
POV & Framing
POV Shot
A POV shot (point-of-view shot) frames the camera as if it were a character's own eyes, moving and reacting the way their head and body would rather than observing them from outside. It drops the viewer directly into the subject's position, which is why it's a favorite in action scenes (a punch flying toward the lens), horror (turning a corner into the dark), and first-person storytelling of any kind. The move works best paired with natural head-movement cues, like turning to look at something, glancing down, or a slight bob while walking, since those small motions are what sell the illusion that the camera is a person rather than a floating lens.
Over-the-Shoulder Shot
An over-the-shoulder shot (often shortened to OTS) positions the camera just behind and to the side of one person, so a sliver of their shoulder and head frames the edge of the shot while a second subject faces the camera beyond them. It's the standard building block of a filmed conversation, because cutting between two OTS shots (one over each person's shoulder) keeps the audience oriented about who is talking to whom and where they're standing relative to each other, without ever losing that spatial relationship. Keeping the foreground shoulder soft and out of focus, with the far subject sharp, is what makes the framing read correctly rather than looking like an accidental obstruction.
Specials
Time-Lapse
A time-lapse compresses a long stretch of real time, hours of a sunset, days of construction, weeks of a blooming flower, into just a few seconds of fast, fluid motion, usually from a single fixed camera position. It's technically a time-manipulation technique rather than a camera movement, but directors often combine it with a slow pan, tilt, or slider move to add production value on top of the sped-up passage of time. The giveaway details that sell a time-lapse are fast-moving clouds or light changes and streaking light trails from moving traffic, both worth naming explicitly in the prompt.
Pass-Through Shot
A pass-through shot sends the camera directly through an object that should stop it, a wall, a window, the surface of water, a closed door, in one unbroken, physically impossible move that reveals whatever is on the other side without ever cutting. Because it can't be filmed for real, it's a shot that only became common once VFX (and now AI video) made it easy, and it works best as a transition device: starting a scene outside a building and passing straight through the window into the room, for instance. Naming the specific surface the camera passes through, and confirming the motion never pauses at the moment of contact, is what keeps the illusion seamless.