Use keywords like: Fixed perspective, fixed camera, static shot, unmoving camera, no shaking, stable image
2.1 Number of Characters:
Default is one person, unless otherwise specified.
2.2 Character Identity:
Describe clearly: man/woman, child/elderly, chef/worker, etc.
2.3 Appearance Description:
Examples:
-Wearing simple clothes
-Short hair / long hair / bun hairstyle
-Wearing a light blue dress / beige casual short-sleeve top / white shirt / yellow long-sleeve sweatshirt
2.4 Ethnicity:
Chinese / Asian / European ...
2.5 Framing:
Specify whether the character is:
-Full body
-Half body
-Full body shown in frame
3.1 How does the character enter the scene?
If not specified, the default is usually "walking into frame."
However, there may be unexpected behavior — for example, the character floats in from the sky. To avoid this, clarify how the character enters.
3.2 What does the character do?
For example:
-Sits on a stool
-Sits on the bed
Note: It’s important to anchor the action to a specific object (e.g., "sits on the black stool") — otherwise, the AI may generate a character sitting on a self-generated stool.
A complete prompt: "[Number of Character] + [Character Identity] + [Apperance] [Action] + [Camera Perspective/Movement] "
-The camera uses a fixed perspective. A young Chinese man in a simple T-shirt walks into the room. Full body shown. He sits in front of a computer.
-The camera uses a fixed perspective. A young Chinese woman in simple clothing walks into the room. Full body shown. She sits on the round stool shown in the image and is applying makeup.