Camera and lens terms for AI photo prompts: focal length, aperture and framing

Four decisions behind every photograph

Every photograph rests on four camera decisions: where the camera stands, how wide the lens sees, how much of the scene is sharp, and how much of the subject fills the frame. Writing those decisions into a prompt gives an image generator concrete, visual instructions instead of vague words such as “cinematic” or “stunning”. This guide explains the terms that matter and what each one changes in a real photograph. No generator follows them exactly, so treat them as strong hints and check every result.

Focal length, in plain terms

Focal length, written in millimetres, describes how wide or narrow a lens sees. The reference is the full-frame 35 mm format. For that format, manufacturers have by custom adopted 50 mm as the normal lens, and a lens of 35 mm or less is considered wide-angle. At the other end, short or portrait telephoto lenses are usually placed between 85 and 135 mm. The numbers below follow that convention.

Wide-angle: 35 mm and below

A wide-angle lens takes in a lot of the scene, which suits interiors, architecture, landscapes and a person shown in their surroundings. In a prompt, pair the focal length with what the width is for: “24 mm, the whole kitchen visible, doorway on the left”. Be careful with faces. Perspective distortion is particularly noticeable in portraits taken with wide-angle lenses at short distances, so a close portrait “at 16 mm” is likely to produce stretched features.

Normal: around 35 to 50 mm

A normal lens reproduces a field of view that appears natural to a human observer. Such lenses suit street scenes, documentary-style moments, food on a table and everyday objects. When you want an image to feel like an ordinary photograph rather than a styled one, “35 mm” or “50 mm” is a sensible starting point, combined with a plain description of where the camera is.

Portrait telephoto: 85 to 135 mm

Longer lenses show a narrower slice of the scene, and the 85 to 135 mm range is the one classed as portrait telephoto. In a prompt, “85 mm portrait” pairs well with a head-and-shoulders or half-length framing with the background reduced to a small area behind the subject. Longer telephoto terms, such as 200 mm or 400 mm, ask for an even narrower slice, suited to distant subjects and small details far from the camera.

Perspective comes from distance, not from the lens

This point is easy to miss. Changes in linear perspective are caused by the distance between camera and subject, not by the lens itself: two photographs of the same scene taken from the same spot show identical perspective, whatever the lens. The lens mostly decides how much of the scene is included. So say where the camera is: “photographed from across the street” or “from arm’s length” changes the look more reliably than a focal length alone.

Phone focal lengths are equivalents

Phone cameras have small sensors, so their lenses are described by a 35 mm equivalent: the actual focal length multiplied by the sensor’s crop factor. Apple, for example, lists the main camera of the iPhone 15 Pro as “24 mm, ƒ/1.78”. Equivalence also applies to depth of field: a 50 mm f/2 lens on a camera with a 2× crop factor behaves like a 100 mm f/4 lens on full frame. A phone’s f/1.78 therefore does not blur a background the way f/1.8 does on a full-frame camera.

Aperture and depth of field

Depth of field is the distance between the nearest and farthest things that look acceptably sharp. It depends on focal length, distance to the subject and aperture. The aperture is written as an f-number, and a smaller aperture, meaning a higher f-number, increases depth of field. In a prompt, describe the effect as well as the number: “f/1.8, shallow depth of field, background softly out of focus” or “f/11, everything sharp from the foreground to the horizon”.

Shot size: how much of the subject

Film vocabulary gives precise names for how much of a subject fills the frame. A long or wide shot shows the environment around the subject. A full shot shows the whole subject. A medium-long shot ends near the knees, and a medium shot stops just above or below the waist. A medium close-up shows more of the shoulders than a close-up, a close-up still shows the shoulder line, and an extreme close-up stops at the chin and the forehead. These names are clearer than “zoomed in” or “far away”.

Camera angle: where the camera looks from

The angle changes how a subject reads. At eye level, the camera faces the subject straight on. In a high-angle shot, the camera is physically higher than the subject and looks down on it; in a low-angle shot, it looks up. A bird’s-eye view is taken from directly above the scene, and a worm’s-eye view looks up from the ground. A point-of-view shot shows the image through the subject’s eyes, and the over-the-shoulder shot is another common viewpoint.

Combine terms without contradictions

Camera terms fail most often when they contradict each other. “Extreme close-up” and “the whole mountain range visible” cannot both be true. “f/1.4” and “everything sharp” pull in opposite directions, and so do “24 mm” and “85 mm portrait” in the same sentence. Choose one value for each decision, write it in its own short sentence, and make sure the rest of the prompt describes a scene that camera could actually capture.

Brand names are weaker than numbers

Prompts often name a camera model or a phone. That can nudge a style, but the effect varies between generators and is hard to predict. A focal length, an aperture, a shot size and a plain sentence about what should be sharp give instructions you can check on the result. If you use a model name, keep it as a secondary hint after those, and never rely on it to fix framing or focus.

Test one camera term at a time

To learn what a term does in your generator, change one thing and keep the rest of the prompt identical. Generate the same scene at 24 mm, 50 mm and 85 mm, then at f/1.8 and f/11, and compare the images side by side. Keep a note of which wording gave the effect you wanted. Results vary between attempts, so look at several images for each version before concluding.

Where LarpGPT fits

LarpGPT is a library of photo prompts for travel, architecture, cars and lifestyle scenes. Almost every prompt carries a camera line, usually a phone’s equivalent focal length and aperture such as 24 mm at f/1.78, alongside its framing, lighting and subject. You can copy a prompt, change only its lens or framing using the terms above, and paste it into the image generator you already use. Its customiser also lets you swap the lighting among several presets. LarpGPT does not generate images itself, and results vary between attempts.

Further reading

• Wikipedia · Wide-angle lens
• Wikipedia · Telephoto lens
• Wikipedia · Normal lens
• Wikipedia · Perspective distortion
• Wikipedia · 35 mm equivalent focal length
• Wikipedia · Depth of field
• Wikipedia · Shot (filmmaking)
• Wikipedia · Camera angle
• Apple Support · iPhone 15 Pro, Tech Specs