TelosCompass · note AI视频创作实践·· 2026-10-13AI score36
How to use prompts to set camera movement in AI video generation and lock in the final frame
動画生成のカメラワーク、最後の画面まで決める。
AI briefing
When generating AI video from a single photo, camera movement only requires choosing one of three options—static, dolly in, or dolly out—and specifying the stopping position to determine what the final frame will show. If you only write "dolly in," the frame may stop on a close-up of a vase pattern without revealing it's a vase; dolly out may cause the AI to generate sheds and windows not present in the photo, so real-world scenes should use a wide-angle photo with "static." DomoAI's official blog recommends using only one camera movement per scene, while Runway's official documentation advises adding speed or duration to camera movement terms and distinguishing between camera push-in and zoom.
Open original note.com