【Subject】
Prompts to optimize camera movement when generating text-to-video (T2V) and images-to-video (I2V). (App where you can enter prompts)
【The purpose】
Based on the situational information you provide, create detailed prompts (moving camera work (but not still images)) that enhance the story.
[Recommended for creating prompts]
Language: English
Length: Approximately 100 words, including additional words
Structure: Chronologically structured using two or three cameraworks
Form of expression :
“Start with, ..., then ..., finally ...”
However, given the SeaArt specifications, it would be difficult to incorporate three stages of camera movement in the first place….
[Situation and Story Information]
・Video summary:
・Characters:
・Emotions and mood:
・Additional information:
[Example] (T2V)
A slow zoom out from darkness reveals a beautiful Japanese woman standing still in a vast digital void. Her translucent body is traced with softly glowing wire-like light patterns that gently pulse. She wears a sheer, skin-tight dress and glasses, which project streaming data. As the scene subtly shifts sideways, countless translucent data windows pass by in rigid alignment.
Then the camera pans left, She lifts one delicate hand and touches a window — it shatters into glowing particles that dissolve into the air.
Finally the camera Push in, Wireframe lights rhythmically pulse across her calm, unreadable face, partially obscured by the glint of her glasses.
Cinematic lighting, cyberpunk atmosphere, high-detail, soft shadows, volumetric glow, digital neon aura.
[Video generated by the above prompt]
https://www.seaart.ai/ja/explore/detail/d02m4r5e878c73dtutrg
https://www.seaart.ai/ja/explore/detail/d02m50te878c73e9krs0
https://www.seaart.ai/ja/explore/detail/d02mnsle878c73b69itg
https://www.seaart.ai/ja/explore/detail/d02nfu5e878c73b5e4m0
As expected, only the first instruction is output. SeaArt needs to work harder on this. :-x
By the way, it seems that one operation takes about 5 seconds.
[Supplement]
Below are some negative prompts that I often use.
If the app you use doesn’t have “Negative Prompt”, include words related to image quality in your prompt.
worst quality, low quality, out of focus, bad ?, jpeg artifacts, lowres, error, missing limb, bad hands, missing fingers extra digit, fewer digits, malformed hands, disconnected limbs, inaccurate limb, extra fingers, bad fingers, poorly drawn fingers, liquid fingers, distorted gaze, character is not moving, ????,
thanks.
