Camera language for AI video: the 8 angles and 4 apertures that make AI footage look real

Most AI video looks like AI video. The lighting is fine, the subject is fine, and yet something feels off. Nine times out of ten the problem is not the model. It is the prompt. It describes what is in the shot but says nothing about how the shot was filmed.

Real footage always comes from a real camera. That camera sits somewhere, points somewhere, and has a lens set a certain way. When you leave those choices out, the AI fills them in with a flat, safe middle. When you put them in, the footage starts to read like something a person actually shot.

You do not need film school for this. You need two small vocabularies: four apertures and eight angles. That is the whole kit.

Part one: the 4 apertures

Aperture is how wide the lens opens. It controls how much of the frame is in focus. The rule is simple: a low number means a soft, dreamy background. A high number means everything stays sharp.

01

f/1.4

Only the eyes are sharp and the background melts into soft color. Portraits, emotion, anything intimate.

02

f/2.8

The subject pops but you can still read the room. Product shots, interviews, a person in their space.

03

f/8

Sharp from front to back. Streets, storefronts, a team standing together.

04

f/16

Everything sharp all the way to the horizon. Landscapes, skylines, wide establishing shots.

Part two: the 8 angles

Angle is where the camera sits and which way it points. Each one tells the viewer something before a single word is said.

01

Low angle, tilted up

The camera sits below and looks up. It adds drama and makes people or products feel powerful.

02

Close up

Tight on a face or detail. Your reaction shot: the moment someone tastes, sees, or feels something.

03

Floor level fisheye

The lens sits on the ground with a wide, curved view. It exaggerates scale and makes small spaces feel big.

04

Over the shoulder

We look past one person toward what they see. It pulls the viewer in, as if standing right there.

05

Overhead

Straight down from above. It shows layout clearly: a table spread, a workspace, a plated meal.

06

High angle

The camera looks down at a slight angle. It shows space and context, and can make a subject feel small.

07

Foreground framing

Something sits close to the lens with the subject behind it. It adds depth and makes the shot feel observed.

08

Circular fisheye

A round, heavily curved image. Playful and a little goofy, which makes it a great ending shot.

You do not need all eight in every video. A short promo might use three: a low angle to open, a close up for the reaction, and a circular fisheye to close. The point is to choose on purpose instead of letting the model choose for you.

How to write this into a prompt

A simple order works well. Start with the angle, then the subject and action, then the setting, then the lens and aperture, then the light. Keep it in plain sentences. The model reads camera words the same way a cinematographer would.

Here are three examples you can copy and adapt.

A few habits that help

• Match the aperture to the angle. Close ups love f/1.4. Overheads and wide shots usually want f/8 or higher.
• Name a lens length when you can. 24mm feels wide and energetic, 50mm feels natural, 85mm feels flattering on faces.
• Describe the light like a real source: window light, neon sign, golden hour, overhead fluorescent. Real sources make real footage.
• Generate a few versions and keep the best. Small wording changes can move the result a lot.

How we use this at Bourvic Studio

We are a design and development studio in Kansas City. We build websites, brand kits, content and the systems behind them for businesses of every size. AI video has become one more tool on the bench, and camera language is how we keep it looking like it belongs next to real work.

When we plan a motion promo, we write a shot list first, the same way we would for a live shoot. Each shot gets an angle, an aperture, and a reason. Then we generate those shots in Higgsfield, pick the strongest takes, and finish them with real editing, type, and sound. The camera choices are what make the pieces cut together like one film instead of a pile of random clips.

Credit where it is due: this framework was sparked by Promptwise, a creator who shares sharp, practical AI video tips on Facebook and Instagram. If this post helped, go follow them. And if you would rather have someone handle the whole thing for your brand, that is what we are here for.