Complete Guide to Camera Terms You Can Use in AI Image Generation

You know that feeling when you're writing a prompt and thinking "I want it to look like it's shot from this angle..." but can't find the words to explain it? That frustration when you can picture it in your head but can't think of the English term.

When I first started, I didn't know how to describe "a shot looking down from above," so I just wrote something like "looking down from above" and got weird results. Turns out if you write "high angle" or "bird's eye view," the AI understands much better.

I came across this terminology guide compiled by Sinaida Cooper and thought "oh, so these are the words I needed!" so I translated it. Since these are based on cinematography terms, tools like Midjourney, Flux, and DALL-E recognize them well.

1. Camera Height (Vertical Position)

These terms determine where the camera sits on the Y-axis.

From my experience, low angle works best. When you want to make a character look cool, adding this makes a real difference. Conversely, high angle is great for creating a cute or vulnerable feeling.

2. Camera Distance / Framing

This determines how much of the subject fits in the frame. Without knowing these, it's hard to get the composition you want.

For character illustrations, I use medium shot or medium close-up the most. You can see the expression and the outfit to some extent, so it's good value. If you want to show the full body, you need to use wide shot, but then you lose some facial detail.

3. Camera Direction / Angle

This indicates how the lens rotates or aims.

Three-quarter angle is truly versatile. It looks the most natural and three-dimensional for portraits and character illustrations. Straight-on gives an ID photo feel, and profile creates a silhouette effect, so use them according to the situation.

Dutch angle is great for horror or thriller vibes, but overusing it makes things look dizzy, so use it in moderation.

4. Camera Position Relative to Subject

This represents the spatial relationship between camera and subject.

POV is great for game art or creating immersive scenes. Writing something like "POV, reaching out hand toward viewer" really gives that first-person feel.

Over-the-shoulder is essential for conversation scenes. When depicting two characters in dialogue, it looks awkward without this.

5. Lens / Perspective

This controls distortion and spatial compression. The same subject looks completely different depending on the lens.

Wide-angle is good for creating backgrounds with a sense of space, and telephoto compression is used in portraits to blur the background and focus on the subject.

Fisheye is used for skateboard videos or hip-hop music video vibes, but it's a bit much for regular illustrations.

6. Camera Position in 3D Space

These terms tell the model where the camera is positioned in the scene.

Close foreground placement is effective for creating depth. Placing something blurry in the foreground with the main subject behind really brings out the three-dimensionality.

7. Camera Movement (For Still Images)

Even in still images, the direction of movement affects perspective. Useful for creating dynamic effects.

I sometimes use handheld wobble for a realistic photo feel. A slight shake can actually look more real than something perfectly stable.

Commonly Used Combinations

Here are prompt combinations I actually use a lot.

Making a Character Look Cool

low angle, medium shot, three-quarter angle, dramatic lighting

Intimate Portrait

eye level, medium close-up, shallow depth of field, soft lighting

Figure in Majestic Landscape

extreme long shot, high vantage point, wide-angle perspective

Tense Scene

dutch angle, close-up, high contrast lighting

First-Person Immersive Scene

POV, wide-angle perspective, reaching toward camera

Summary

You don't need to memorize all of these at first. I still haven't memorized everything, and I just look them up from this list when I need them.

Just knowing eye level, low angle, high angle — these three — covers the basics. You can gradually expand from there.

Putting camera terms at the beginning of your prompt seems to help the AI understand better. Something like low angle, medium shot, cyberpunk warrior...

These terms work with Midjourney, Flux, DALL-E, and Stable Diffusion, so feel free to use them.