How to use this atlas
This page is AI for Everyone’s camera-movement prompt atlas. It uses the same set of red-chair, white-column and courtyard diagrams to compare 25 common moves and combinations. The 12 beginner entries practise one change at a time: position, orientation, focal length or viewpoint. The 13 advanced entries then practise timing, following a subject, and combined paths. The levels are only this atlas’s learning order, not an industry-wide difficulty scale. The collection covers common wording, not every possible filming method. Not every entry is a physical move: a static shot is a state, a zoom is a change of focal length, and a straight-down bird’s-eye view is a viewpoint.
The diagrams are original teaching drawings, not AI-generated video or filmed results. The left image shows the path or viewpoint of the camera or subject; the right image compares framing from A to B; dashed lines and faded shapes mark the starting position. Beginner examples keep the chair, columns and other scene objects still so you can see how the camera itself moves. Advanced prompts use replaceable fields: fill in your subject, setting and target framing, and test a complex path in steps. Adjust duration and speed to the tool you use. You can search below by move, effect or prompt wording, and open all 25 bilingual prompts, path diagrams and mix-up notes. This atlas is original teaching material. The prompts have not been tested in any video-generation model. The diagrams explain intended camera movement; they are not model output or proof of an effect. Different tools and settings can give different results. Check the features and limits of your tool before use, and test a short clip first.
How to write a camera-movement prompt
When you write a camera-movement prompt, finish one move in one sentence. Do not mix several paths in the same sentence. Each sentence should make four things clear: where the camera travels or turns (direction), how fast or slow it goes (speed), the starting framing, the ending framing, and what must stay still, for example whether the subject stays put, whether camera height is fixed, and whether focal length may change. Writing only “steady”, “zoom in”, or “move the lens right” can be read as a slow push-in, a zoom, or a trucking move of the whole camera. The beginner examples in this atlas describe intent with the same scene, one continuous shot, and a target duration of 6 seconds. The words “6 seconds” do not guarantee output length. Replace the red chair, setting, speed and duration with what you need. These lines are a starting point for describing camera intent, not commands for any one tool. Different models may ignore or misread camera movement.
How push-in and zoom differ
This page lists separately the three “make it larger” methods that are easiest to mix up. Push-in moves the whole camera toward the subject with focal length unchanged. The shooting distance shortens, and perspective changes with it. Zoom-in leaves the camera in place and only increases focal length, so the field of view narrows and the subject looks larger. That is a lens-parameter change, not a camera move, so the perspective relationship between subject and background stays the same. Dolly zoom combines a camera move with a zoom in the opposite direction. In this example the camera pulls back while focal length increases, keeping the red chair about the same size in the frame as the white columns behind it grow. A zoom without a camera move will not create that effect. The retreat and the zoom must stay in step, and you must name the subject whose size should stay constant. Do not write only “enlarge the chair”, and do not describe a zoom as “the camera moves closer”.
A pull-out is the counterpart: the whole camera moves back from the subject with focal length unchanged, revealing more of the surrounding space. If the camera does not retreat and the frame only shrinks, that is a zoom-out, not a pull-out.
A long take is not a camera move
This page also treats “long take / one-shot” as a related shot concept. It is not a 26th camera move. A long take is about the duration and continuity of a single shot. It can be filmed from a fixed camera, or combined with tracking, trucking or orbiting. A long shot is a wide framing, and a long lens is a focal length; neither is a long take. An apparent one-shot sequence may also use hidden cuts. A prompt does not guarantee an uncut result. Duration and continuity depend on the model, so check the actual output yourself.
Common questions
How do I write a camera-movement prompt?
Finish one move in one sentence. Do not mix several paths in the same sentence. Make direction, speed, starting framing, ending framing, and what stays still all clear, for example whether the subject stays put, whether camera height is fixed, and whether focal length may change. Writing only “steady”, “zoom in”, or “move the lens right” can be read as a slow push-in, a zoom, or a trucking move of the whole camera. The beginner examples describe intent with the same scene and one continuous shot. Replace the red chair, setting, speed and duration with what you need. These lines are a starting point for describing camera intent, not commands for any one tool.
How do push-in, zoom and dolly zoom differ?
Push-in moves the whole camera toward the subject with focal length unchanged. The shooting distance shortens, and perspective changes with it. Zoom-in leaves the camera in place and only increases focal length, so the field of view narrows and the subject looks larger. That is a lens-parameter change, not a camera move, so the perspective relationship between subject and background stays the same. Dolly zoom combines a camera move with a zoom in the opposite direction. In this example the camera pulls back while focal length increases, keeping the red chair about the same size in the frame as the white columns behind it grow. A zoom without a camera move will not create that effect.
Is a long take a camera move? How is it different from a long shot or a long lens?
A long take is not a camera move. It is about the duration and continuity of a single shot. It can be filmed from a fixed camera, or combined with tracking, trucking or orbiting. A long shot is a wide framing, and a long lens is a focal length; neither is a long take. An apparent one-shot sequence may also use hidden cuts. A prompt does not guarantee an uncut result. Duration and continuity depend on the model, so check the actual output yourself.
Have these prompts been tested in a model? Are the diagrams generated results?
No. This atlas is original teaching material. The prompts have not been tested in any video-generation model. The diagrams explain intended camera movement; they are not model output or proof of an effect, and they are not AI-generated video or filmed results. Different tools and settings can give different results. Check current features and limits before use, and test a short clip first.
Do a static shot, a zoom and a bird’s-eye view count as camera moves?
Not every listed item is a physical move. A static shot is a state: camera position, orientation and focal length stay unchanged, and the page states that this is a fixed way of shooting, with no camera move. A zoom is a change of focal length, with the camera left in place. A straight-down bird’s-eye view describes a viewing angle; this example is a fixed overhead viewpoint, not a move. The levels are only this atlas’s learning order, not an industry-wide difficulty scale.
This is original educational content. These prompts have not been tested in any video-generation model. Diagrams explain the intended shot; they are not model outputs or evidence of performance. Results may vary by tool and settings.
Original schematic, not an AI-generated video or filmed result. Left: camera/subject path or viewpoint. Right: framing from A to B. Dashed lines and faded shapes show the starting state.
No matching techniques or concepts. Try arc, long take, an English name, or clear the search.
