PROMPT RECORD图像记录
我想在 Seedance 2.0 中创建一个动漫视频场景。我有上传的图片作为参考,其中包括一张少女的角色设定图、一张凤凰的设定图、一张分镜图,以及一张森林场景图。该场景是一部日本风格的冒险系列...
中文说明
我想在 Seedance 2.0 中创建一个动漫视频场景。我有上传的图片作为参考,其中包括一张少女的角色设定图、一张凤凰的设定图、一张分镜图,以及一张森林场景图。该场景是一部日本风格的冒险系列中的动作场景。请分析这些图片,并给我一个适用于 Seedance 的提示词,以便它能生成动画视频,采用场景的艺术风格作为风格,并强调严格遵循分镜的镜头构图。 测试由 Gemini/ChatGPT 给出的调整后的提示词选项,并使用其中效果最佳的一个。 最终提示词: 请使用具有严格优先级和角色分离的参考图: [REF_STORYBOARD] = "Image1" 作为镜头顺序、构图、时长、镜头构成和场景推进的主要参考。请严格遵循。 [REF_GIRL_MODEL] = "Image2" 作为该少女的本体锚点。 [REF_PHO...
原始 Prompt
I want to create an anime video scene in Seedance 2.0. I have as reference the uploaded images which include a character sheet of a teenage girl and another of a phoenix bird, an image of the storyboard and another of a forest setting. The scene is an action scene from an adventure series, Japanese style. Analyze the images and give me back a prompt for Seedance so it generates an animation video, using as style the artistic style of the setting and emphasizing on adhering to the shot composition of the storyboard.
Test the adapted prompt options given by Gemini/ChatGPT and use the one that shows the best results.
Final Prompt:
Use the references with strict priority and role separation:
[REF_STORYBOARD] = "Image1" primary guide for shot order, framing, timing, composition, and scene progression. Follow it strictly.
[REF_GIRL_MODEL] = "Image2" identity anchor for the teenage girl.
[REF_PHOENIX_MODEL] "Image 3" creature identity anchor for the phoenix.
[REF_BACKGROUND] = "Image 4" environment and forest background anchor.
Create a Japanese anime action scene of a teenage girl fighting two phoenix birds in a forest.
Absolute priority:
1. Follow the storyboard strictly for all shots and scene progression.
2. Keep the girl design locked to [REF_GIRL_MODEL].
3. Keep the phoenix design locked to [REF_PHOENIX_MODEL].
4. Keep the forest and path consistent with [REF_BACKGROUND].