15 Camera Shots & Movements Every AI Filmmaker Should Know
By Richard Garland · 2026-09-05
Generating a beautiful image is one thing. Directing a shot is another.
One of the fastest ways to make AI-generated video feel more intentional is to stop treating the camera as an invisible observer. Where is the camera? How close is it to the subject? Is it moving? Why is it moving? What should the audience learn or feel because of that movement?
Modern AI video models increasingly understand the language of cinematography: close-ups, wide shots, tracking shots, push-ins, crane movements, handheld cameras, POV shots and more. But simply stuffing cinematic terminology into a prompt isn't enough. The goal isn't to make the camera move. The goal is to give the camera a reason to move.
Here are 15 shots and camera movements every AI filmmaker should have in their toolkit. If you want the wider framework first, it sits alongside our prompting guide.
First: Shot Size and Camera Movement Aren't the Same Thing
Before we start, there's an important distinction. A shot size describes how the subject is framed:
- Extreme wide shot
- Wide shot
- Medium shot
- Close-up
- Extreme close-up
A camera movement describes what the camera does during the shot:
- Pan
- Tilt
- Dolly
- Tracking
- Crane
- Orbit
You can combine the two. For example:
Now we know both how the shot begins and how the camera behaves. That's much more useful than simply writing "cinematic shot of a detective."
1. Establishing Shot
An establishing shot introduces the audience to a location and establishes geography before the story moves closer. These are often wide or extreme-wide compositions.
Best for
Opening scenes, introducing new locations, establishing scale, transitions between locations and giving the audience spatial context.
Prompt example
Directing tip
Don't use an establishing shot simply because films have them. Ask what the environment tells us. A tiny character surrounded by a massive landscape can communicate isolation before the character says a word.
2. Close-Up
A close-up brings the audience directly into a character's emotional space. Faces are the obvious use, but close-ups can also focus on meaningful objects.
Best for
Emotion, reactions, important details, dialogue, suspense and reveals.
Prompt example
Notice that the camera doesn't need to move. Sometimes stillness is the direction.
3. Extreme Close-Up
An extreme close-up isolates a tiny detail and makes it important. An eye. A trigger finger. A drop of sweat. A key turning inside a lock.
Prompt example
Extreme close-ups are particularly effective as inserts between wider shots.
4. Low-Angle Shot
Place the camera below the subject and point upward. The result can make a character or object appear powerful, imposing, heroic or threatening.
Prompt example
The angle itself tells part of the story.
5. High-Angle Shot
Now reverse the relationship. Place the camera above the subject and look downward. A high angle can emphasize vulnerability, geography, isolation or simply reveal information unavailable from eye level.
Prompt example
Camera angle is not merely decoration. It changes the audience's relationship with the subject.
6. Dolly Push-In
A push-in physically moves the camera closer to the subject. It's one of the most useful movements in cinematic storytelling.
Best for
Realization, tension, emotional emphasis, discovery and drawing attention to an important moment.
Prompt example
The important part isn't "camera pushes in." It's why it pushes in at that moment. Here, the camera reacts to the character's realization.
7. Dolly Pull-Out
A pull-out does the opposite. The camera moves away from the subject. This can reveal new information, increase scale or make someone appear increasingly alone.
Prompt example
The movement becomes a reveal. We begin with one piece of information and end with another.
8. Tracking Shot
A tracking shot moves with the subject. Instead of watching someone cross the frame, the camera travels alongside, behind or ahead of them.
Best for
Walking scenes, running, action, exploration, entrances and immersive sequences.
Prompt example
The subject and camera now have a relationship, so specify that relationship. Don't just prompt "tracking shot" — try "the camera tracks beside her at matching speed." There are more prompt examples to adapt in the library.
9. Pan
A pan rotates the camera horizontally from a fixed position. Think of turning your head left or right.
Best for
Following movement, revealing information beside the subject, surveying environments and transitioning attention between characters.
Prompt example
Again, the movement has a destination. The pan isn't there to look cinematic. It's there to reveal the monitor.
10. Tilt
A tilt rotates the camera vertically, up or down.
Best for
Revealing height, introducing characters, architecture, scale and dramatic entrances.
Prompt example
This is essentially controlled information delivery. The audience doesn't see everything immediately. You decide when they see it.
11. Orbit Shot
The camera moves in an arc around the subject. Orbit shots can make a character feel important, create dramatic dimensionality, reveal the environment around them or heighten a pivotal moment.
Prompt example
One warning: more camera movement isn't automatically better camera movement. If a gentle partial orbit communicates the shot, you may not need an enormous 360-degree rotation.
12. Crane / Jib Shot
A crane-style movement changes the camera's vertical position through space. It can begin close to the subject and rise to reveal the world around them, or descend from a large environment into an intimate scene.
Prompt example
This is a great scale-reveal movement. Start small. End big.
13. Handheld Shot
Not every shot should glide perfectly through space. Handheld movement introduces small imperfections that can make a scene feel immediate, documentary-like, chaotic or intimate.
Prompt example
The important word here is not just "handheld." Describe the behavior of the handheld camera. Subtle? Nervous? Aggressive? Following someone? Trying to keep up? Those choices affect the feeling of the shot.
14. POV Shot
A point-of-view shot places the camera approximately where a character's eyes would be. The audience experiences the scene from that character's perspective.
Prompt example
POV works especially well when camera behavior reflects character behavior. Notice the camera hesitates. That's not traditional camera terminology. It's direction.
15. Aerial / Drone Shot
An aerial shot places the camera high above the environment. It can establish geography, follow movement, reveal patterns or create tremendous scale.
Prompt example
Aerial shots become much more interesting when they do something besides simply say "drone shot of a desert." Give the shot a beginning and an end.
The Most Important Technique: Give Camera Movement a Purpose
Here's a useful rule. Don't ask what camera movement looks coolest. Ask what the audience should know or feel by the end of this shot that they didn't know at the beginning. Then choose the camera movement.
- Reveal information — use a pan, tilt, pull-out or crane movement.
- Increase emotional intensity — try a slow push-in.
- Stay connected to a moving character — use a tracking shot.
- Communicate instability — consider handheld movement.
- Show scale — use a pull-out, crane or aerial movement.
- Make the audience inhabit the character's experience — use POV.
Camera movement becomes storytelling instead of decoration.
Combine Subject Motion, Environmental Motion and Camera Motion
One of the biggest mistakes in writing AI video prompts is treating all motion as one thing. Instead, think in layers.
Subject motion
What does the character or object do? "A woman walks toward the elevator."
Environmental motion
What moves around the subject? "Her coat shifts in the wind while steam drifts from a nearby vent."
Camera motion
What does the camera do? "The camera tracks backward in front of her at walking speed."
Put them together:
Now the model receives three different kinds of movement instead of a vague request for a "dynamic cinematic shot."
Don't Over-Direct Every Shot
There's another trap. Once filmmakers discover camera terminology, prompts can turn into this:
Congratulations. You've invented a camera operator's nightmare.
More instructions don't necessarily produce more control. Start with the one movement that matters most, then iterate. If the shot works but feels too static, add another carefully chosen element. If the camera is doing something strange, simplify. Direction is partly knowing what to leave out.
Camera Movement for Image-to-Video
Image-to-video deserves special attention. Your starting image already establishes much of the subject, composition, environment and visual style. That means the prompt can often spend more of its attention on what changes after frame one.
Instead of repeating everything visible in the image:
Try directing the motion:
The image handles much of the what. The prompt directs much of the what happens next — the same is true whether you generate the shot from a still or from text.
A Reusable Camera Prompt Formula
When you're unsure how to structure a shot, try:
For example:
Notice the last sentence. The shot has an ending. That's useful because filmmaking isn't just composition. It's change over time.
If assembling all of that by hand feels like a lot, the Video Prompt Builder scaffolds each part for you.
Before You Generate: Ask These Five Questions
Before sending your next AI video prompt, ask yourself:
- Where is the camera?
- How is the subject framed?
- Does the camera move?
- Why does it move?
- Where should the shot end?
If you can answer those five questions, you're no longer merely describing an image. You're beginning to direct a shot.
Final Thought: Think Like a Director, Not a Prompt Engineer
The terminology matters. But memorizing words like dolly, crane and tracking shot isn't the real skill. The real skill is understanding why a filmmaker would choose one over another.
A push-in can tell us: "Pay attention." A pull-out can tell us: "There's more here than you realized." A tracking shot can tell us: "Come with this character." A locked camera can tell us: "You aren't allowed to look away."
The best AI video prompts aren't necessarily the ones with the most technical terminology. They're the ones where the camera, subject, environment and story are all trying to accomplish the same thing. Shot choice is one piece of a larger AI filmmaking workflow — so don't just tell the model what the camera does. Decide what the shot means.
Have the idea but don't want to assemble all the filmmaking language yourself? Use the RevaultAI Video Prompt Builder to turn a rough concept into a structured video prompt with camera direction, lighting, action, setting, visual style and more. Free. No account required.
Build a Video PromptFrequently Asked Questions
What camera movements work well in AI video prompts?
Common movements include push-ins and pull-outs, tracking shots, pans, tilts, orbit shots, crane movements, aerial shots and handheld camera movement. The best choice depends on what the shot needs to communicate.
Should every AI-generated shot have camera movement?
No. A locked camera can be more effective than movement when you want stillness, tension, symmetry or uninterrupted attention on a performance.
How do I describe camera movement in an AI video prompt?
Be direct. Describe where the camera begins, how it moves relative to the subject and, when important, where it ends. For example: "The camera slowly tracks backward in front of the subject at walking speed."
What's the difference between a pan and a tracking shot?
A pan rotates the camera horizontally from its position. A tracking shot moves the camera through physical space with or around the subject.
How should I prompt camera movement for image-to-video?
Focus primarily on what should change after the starting frame: subject movement, environmental movement and camera behavior. Avoid unnecessarily redescribing visual information already established by the source image.
Can I combine multiple camera movements?
Yes, but use compound movements deliberately. A simple, well-defined movement is often easier to control than several competing camera instructions.
AI video models evolve quickly, and their handling of camera language changes over time. The techniques here are deliberately model-agnostic — always check the current capabilities of the tools in your workflow.