分镜预览四格分镜展示提示词中的画面方向,实际视频效果以生成结果为准。
- 模型
- seedance-2-720p
- 时长
- 15s
- 画幅
- 16:9
- 分辨率
- 720p
- 参考图片
- 0
- 参考音视频
- — / —
怎样让镜头更好保留 reputable - 深夜拉面馆 oner (Imagine.art 原稿 + 8 要素整理)
生成后检查什么本站尚未实测。检查主体一致性、运镜连续性与动作完整性;参考原始来源核对效果。
社区来源 · 本站未实测Imagine.art - Seedance 2.0 prompt guide ↗ · cclank/lanshu-awesome-ai-video-kit
MIT (repository)
完整提示词 v1
Single continuous Steadicam shot. Subject: a salaryman at the counter and a chef behind the counter. Action: from the kitchen end the camera tracks along the counter to the customer; the chef moves with practiced economy - ladles broth, pulls noodles from the pot, places toppings; close-up: the salaryman lifts his chopsticks and takes his first bite, exhales, and his shoulders drop slightly; the camera holds. Scene: a ramen shop in Tokyo at 1am, near empty. Light: warm tungsten practicals over the counter, deep dark behind. Camera: a single continuous Steadicam push, never cutting. Style: photoreal slice-of-life cinematic. Audio: low exhaust fan hum, distant rain, faint sound of a late train, no music. Constraints: maintain both subjects' identities, no watermark, no on-screen text, no jump cuts.
来源模型:Seedance · seedance-2-720p
原始参数:{"aspect": "16:9", "duration": "15s", "resolution": "1080p"}
示例分类不代表本站已接入该模型。实际可用模型以视频片场为准;跨模型使用提示词时,请检查参数并调整。
复用会带入提示词、模型、时长、画幅和已发布的参考素材。缺少的参考文件可在片场补充;提交前会显示费用。
阅读视频创作指南