By PayPerVideo Editorial Team, October 10, 2026

AI Video Shot Sizes: Choose Wide, Medium, or Close-Up

AI video shot sizes describe how much of the subject and its surroundings appear in the frame. A wide shot shows more context, a medium shot balances subject and setting, and a close-up draws attention to a detail. Shot size is different from frame shape, camera movement, or output resolution.

Quick answer: Decide what the viewer needs to understand, then choose the framing that shows it clearly. Describe the subject, the visible surroundings, and what must stay inside the frame. Inspect the full generated clip; a prompt saying "close-up" does not guarantee a stable crop or accurate details.

This guide covers text-to-video scene writing. It does not establish exact focal-length controls, locked framing coordinates, or an in-product shot-size selector in PayPerVideo. Check the selected model's actual options before generation.

Shot size is about what the viewer sees

A wide shot can show where an action happens. A close-up can show a texture or small gesture. Neither is automatically better. Choose the amount of context the message needs, not the most dramatic term.

Adobe's camera shots and angles guide explains framing vocabulary. These film concepts are useful when describing an intended scene, but an AI model may interpret them differently from a real camera operator.

Also separate framing from motion. "Close-up" describes the view, while "slow camera push" describes a change over time. For a first attempt, a static view makes it easier to check whether the shot size serves the idea. Add movement only when it contributes something useful.

Wide shot: make the setting understandable

A wide shot leaves more room around the subject. It can establish a generic workspace, show the relationship between objects, or explain where an action happens. More visible surroundings also mean more geometry and background details to inspect.

A generic quiet workspace with one desk, one chair, and a plain lamp. Wide composition showing the desk in the context of the room. The main desk remains easy to identify. A locked-off camera, steady daylight, neutral colors, and a simple background. Furniture and room geometry stay unchanged throughout one continuous shot.

Check whether the subject is large enough to understand on the destination screen. A room that looks clear on a desktop preview may make the useful detail tiny on a phone. Inspect doors, windows, and furniture edges for changes through the clip.

Use this as an illustration, not a view of a real company's office. If the setting itself is a factual claim, use accurate footage rather than an invented room.

Medium shot: balance detail and context

A medium shot shows enough of the subject to understand its action while retaining some surroundings. The exact boundary depends on the subject and intended use, so describe what should be visible instead of relying only on the label.

Adobe's medium-shot guide discusses this balance. For an object scene, you can say that the whole object and a small area around it should remain in frame.

A plain ceramic mug and a closed blank notebook on a neutral table. Medium composition showing both objects fully, with a little table space around them. Soft daylight from the left and a static camera. The objects remain still, unbranded, and unchanged during one continuous shot.

Check the relationship between the mug and notebook. If one object disappears or changes size, the framing instruction did not produce a reliable result. If the scene needs an exact product, use real media rather than accepting a plausible substitute.

A medium frame can be useful for an educational illustration because it gives enough context without asking the viewer to inspect an entire room. It is still a creative choice, not a guaranteed learning or performance advantage.

Close-up: make one detail the point

A close-up directs attention toward a small part of the scene. Adobe's close-up guide describes its use in film. In a generated object shot, name the detail and any edges that should remain visible.

A close-up of the corner of an open blank notebook on a neutral desk. The page texture and notebook edge are the focal details. Soft steady light, a fixed viewpoint, and a simple background. The paper remains blank and the notebook keeps the same shape throughout one continuous shot.

Inspect small marks, edges, and textures closely. A tight frame can hide unrelated background errors, but it can also make a defect in the focal subject more obvious. Do not assume that less background means every detail is correct.

If a real item's material, stitching, controls, or construction matters, film it or use accurate source media. A generated close-up can illustrate a concept, but it cannot prove the exact features of a product being sold.

Describe the boundaries, not just the label

Instead of writing only "wide shot," say what the setting includes and how prominent the subject should be. Instead of only "close-up," identify the detail and whether the entire object or only part of it should appear.

Useful additions include "the whole mug stays visible," "show the desk and chair together," or "focus on the blank page corner." These are instructions to inspect, not guaranteed constraints. A model can still cut off an edge, add an object, or change the composition.

Google's Veo prompt guide and Runway's text-to-video guide provide scoped scene-writing examples. They do not establish identical camera controls across all PayPerVideo models.

Avoid adding technical lens numbers you do not need or cannot verify. A visual description often gives a clearer review target than a stack of camera specifications. The prompt library offers other scene ideas.

Plan the edit before ordering several views

A sequence might use a wide view to establish a setting, a medium view to explain an action, and a close-up to show a detail. It does not need all three merely because they are common film terms.

Use the AI video storyboard guide to give each shot a purpose. Independent generations may change the room, object, light, or identity between views. Repeating the same description helps planning but does not lock continuity.

If the edit must show the same real item from several angles, use accurate footage or confirm suitable reference controls in the chosen workflow before relying on them. This article does not establish those controls in PayPerVideo.

A framing review checklist

1. Purpose: The amount of context matches the idea the viewer should understand.
2. Visibility: Important objects, details, and action remain inside the frame through the whole clip.
3. Scale: The subject is readable on the intended screen, not only in a large preview.
4. Accuracy: The frame does not imply a false product feature, setting, or event.
5. Text space: Captions added later do not cover the useful subject or sit behind platform controls.
6. Export: Actual dimensions and full playback suit the destination after editing.

Use the AI video artifacts guide to inspect changing geometry and motion. A good opening frame is not enough. Watch the beginning, middle, and end, then play the full clip.

Crop only when the remaining view still works

An editor can turn a wider view into a tighter crop, but it removes information and may enlarge defects. It does not create a different physical perspective or restore accurate details the model never generated.

Review the crop through the full motion. If a subject drifts out of the retained area, choose another segment or composition. Do not stretch the picture to fit a destination. Exact titles, prices, and labels belong in editable text added separately.

Review current pricing and set an attempt budget before generating additional views. A still image, real close-up, or simple diagram may serve the message better than another paid shot. No prompt guarantees a usable frame or fewer attempts.

Frequently asked questions

Is a close-up the same as a zoom?

No. A close-up describes the framing. A zoom describes a change in framing over time from a fixed viewpoint. An AI result may not follow that distinction precisely.

Does a wide shot mean a 16:9 video?

No. Shot size describes how much context is visible, while aspect ratio describes the frame's width-to-height relationship. Check both separately.

Can PayPerVideo lock an exact crop or lens setting?

This guide does not verify exact framing coordinates or focal-length controls. It covers scene descriptions whose results must be inspected.

Should every video use wide, medium, and close-up shots?

No. Use only the views that help communicate the message. A single clear shot may be enough.

Can cropping fix an inaccurate generated product?

No. Cropping can remove surrounding content, but it cannot establish the real item's features or repair invented construction. Use accurate media when those details matter.

Next step: Choose one framing purpose, write a clear scene brief, and check the options in the AI video generator. Inspect the full result before adding it to the edit.