Cinematic Language Visual Notes: Image-to-Video Shot Language Demonstration

Starting from a single static storyboard frame, this piece uses Kling 3.0 Omni image-to-video to "act out" common camera movement vocabulary one by one: push in, pull out, pan, truck, follow, crane up/down, and orbit. The core idea is to turn abstract shot terminology into side-by-side dynamic demonstrations—once the same composition is locked in the first frame, motion is driven solely by the shot-language description in the prompt, letting developers intuitively see what kind of camera position changes and pacing "a single line of camera movement description" corresponds to on screen. In terms of composition, a single first-frame image is bound to image_urls[0], and the output aspect ratio follows the first frame; the prompt focuses on camera movement direction, speed, and the boundaries of subject motion, duration controls the length of each demonstration, and resolution and prefer_multi_shots can be used as needed to decide whether the model does its own shot breakdown. Native audio is not enabled, keeping it a silent-film-style pure visual reference. Use cases: pre-production storyboard previews for film and animation, shot design communication, visual proofreading of short-video storyboard scripts, and building an internal reference library of "terminology—image" correspondences for teams.