Camera & Shots

What Is a Dolly Shot?

Also called dolly in, dolly out, push in, pull out, dolly

A dolly shot moves the whole camera through space on a wheeled platform, most often straight toward or away from the subject. Because the camera changes position rather than focal length, the background shifts in scale relative to the subject, which is what makes a dolly in feel different from a zoom.

Why moving the camera is not the same as zooming

Both a dolly in and a zoom in make the subject fill more of the frame, and that is the last thing they have in common.

A zoom is an optical crop. The camera stays where it is and the lens magnifies, so every spatial relationship in the frame is preserved and simply enlarged. Nothing moves past anything else.

A dolly shot changes the camera's actual position. Objects at different distances shift at different rates, near things sweep past faster than far things, and the relationship between subject and background is continuously rewritten. That parallax is the entire perceptual difference, and it is why a dolly feels like travelling and a zoom feels like inspecting.

The practical consequence is that a dolly in compresses the background less than an equivalent zoom does. Move closer at the same focal length and the background stays wide and grows relatively bigger. That is the classic push-in look: the subject arriving while the world behind them opens up.

The moves and what they carry

  • Dolly in, or push in. Concentration. Slow enough and the audience never notices, which is why it is the standard move under a scene that has to intensify without a cut.
  • Dolly out, or pull out. Revelation and release. Almost always revises the meaning of the frame you started in, and conventionally ends things.
  • Lateral dolly. Reveals depth by sliding foreground past background, useful for making a flat space feel three-dimensional.
  • Dolly with the subject. Holds a walking character at constant size while the world streams behind them.

Speed carries as much meaning as direction. A crawl reads as psychology. A fast push reads as shock and usually needs sound to survive.

Prompting a dolly shot

Video models default to a zoom because it is the easier synthesis, so the prompt has to specify what a zoom cannot do.

  • Name the movement as camera travel, not magnification: camera moving steadily toward her, not getting closer.
  • Describe the parallax. background compressing behind her, or for a lateral move foreground posts sweeping past the distant treeline. This is the single most effective clause, because it describes something a zoom physically cannot produce.
  • State the start and end framing: from a medium shot to a close-up. Without a destination the model tends to drift and reverse.
  • Lock the subject if the subject should not move: static subject. Otherwise models often walk the person toward the lens instead, which looks similar in a thumbnail and wrong in motion.
  • Give the speed a word: slow, steady, creeping. Unqualified movement is usually rendered too fast.

For a push in that has to be exact, a first and last frame model is often more reliable than a text prompt. Supply the wide framing and the close framing as the two keyframes and let the model produce the travel between them, which removes the question of whether it understood the word dolly at all.

The prompt for this

A starting point that reliably produces the effect. Adjust the subject and setting; keep the technical clauses.

Slow dolly in on a woman sitting alone at a kitchen table, camera moving steadily toward her from a medium shot to a close-up, background compressing behind her, static subject, 35mm, dim overhead light

Try Dolly Shot yourself

Open the generator with a starting point already filled in.

AI Video GeneratorAI Motion ControlAI First & Last Frame Video

Frequently asked questions

What is the difference between a dolly in and a zoom in?
A zoom changes focal length from a fixed position, so it magnifies the existing image and the spatial relationships stay identical. A dolly in physically moves the camera, so perspective changes: the background grows relatively larger and the sense of entering the space is real. A zoom looks at something; a dolly goes to it.
How is a dolly shot different from a tracking shot?
The terms overlap and usage varies by crew. In common practice a dolly shot means movement along the lens axis, in or out, while a tracking shot means travelling alongside a subject. Both may use the same dolly and track, so the distinction is about direction rather than equipment.
What does a slow push in do dramatically?
It concentrates attention without the audience registering an edit, so it tends to read as growing understanding or growing dread. Its strength is that it is nearly invisible when slow. A push in that takes eight seconds to travel a metre will hold a scene together in a way a cut cannot.
What is a pull out for?
Revealing context, and usually revising the meaning of what you have been looking at. A face becomes a face in a crowd, a room becomes a room in a ruin. Pulling out at the end of a scene is also a conventional way to release tension, which is why it so often closes a sequence.
Why do AI models turn a dolly into a zoom?
Because both produce a subject getting larger, and a zoom is easier to synthesise from a single viewpoint. The tell is the background: in a real dolly in it changes shape as you pass through space, while in a zoom it simply scales. Naming the parallax explicitly is what pushes a model toward the correct one.

Related terms

Back to Camera & Shots