seedance-storyboard

When Not to Use This Skill

• Single shot with a single action → use seedance-prompter
• Compact 5/10/15-second short videos (the primary use case for HappyHorse) → use happyhorse-prompter
• Creating a loop video → single shot +

  loop-ready motion with seamless start and end frames

Core Rules

1. Use identifiers such as

   Shot 1

   ,

   Shot 2

   , and

   Shot 3

   , and organize them in the order the events occur (primary events first)
2. Do not impose a fixed duration on each segment (notations such as "00-03s" are unstable; let the model generate the pacing naturally)
3. Seedance supports a maximum of 15 seconds, so keeping the number of shots between 3 and 5 is the safest approach
4. Organize each shot along four dimensions:
   • Camera movement or shot transition method (e.g., "slow push-in from a wide shot," "locked-off camera," "cut to...")
   • Subject action and expression (key actions and expression changes of the main character or object)
   • Position or spatial change (the scene, position, or spatial relationship in which the subject is located)
   • Audio information (sound effects, voices, or background music)

Workflow

Step 1: Analyze the Plot

• What is the most important image at this moment?
• What action or expression serves as the emotional anchor at this moment?
• How do consecutive shots connect (action continuity, scene transition, or time jump)?

Step 2: Determine Asset Roles

Step 3: Assemble with the Template

[素材分配 + 主体定义]，参考 @视频X 的运镜方式（如有）。

镜头 1：[运镜] + [主体@素材N 的动作与表情] + [位置/空间] + [音频]。
镜头 2：[运镜切换] + [主体@素材N 的动作变化] + [位置变化] + [音频]。
镜头 3：[运镜] + [主体动作] + [位置] + [音频]。

[整体约束：风格/色调/质量/稳定性]

Step 4: Check

• Specify only one type of camera movement per shot (do not combine a push-in and a pan)
• In multi-subject scenes, refer to each character with a stable label ("police"/"thief") and avoid ambiguous references
• When a subject is involved, explicitly identify it ("@Image N" or "<Subject N>"); do not omit the reference
• Make actions specific to body movements (slowly raising a hand, forcefully pushing off the ground), and avoid abstract verbs such as "run" or "fight"
• Externalize emotions concretely (see the Emotion Externalization Table)

Step 5: Output Format

## 分镜脚本

\`\`\`
[完整分镜提示词]
\`\`\`

**镜头数**：N 个 · **推荐时长**：10-15 秒 · **比例**：[X:Y]

**镜头节拍分析**：
- 镜头 1（建立）：[简述]
- 镜头 2（发展）：[简述]
- 镜头 3（结尾）：[简述]

**素材建议**：[告诉用户准备哪些参考素材]

Practical Example

1. Calm before answering the phone (setup)
2. Freezing at the moment she receives the news (turn)
3. Breaking down in private (release)

将@图片1的中年女性定义为主角，故事场景设定在@图片2的医院走廊。整体运镜参考@视频1
的中景跟拍。

镜头 1：医院走廊白炽灯下，主角@图片1 靠在墙边低头看手机，肩膀放松，嘴角带着浅
        浅的微笑回复信息。中景平稳跟拍，从侧面缓慢推近。窗外傍晚柔光，画面
        色调温暖。<远处护士站的低语声和脚步回声>。

镜头 2：手机响起，她抬手接听，镜头切到面部特写。瞳孔瞬间放大，下颌线绷紧，呼吸
        屏住三秒。她的手指无意识地攥紧手机壳。固定机位，焦点锁定在眼神变化上。
        背景虚化，色调变冷。<电话听筒里隐约的机械嗓音 + 心跳声渐强>。

镜头 3：她快步走进走廊尽头的洗手间，背靠门缓慢滑坐到地上，双手抱住膝盖，肩膀
        开始无声颤抖。镜头从上往下缓慢拉远，最后定格在她蜷缩成一团的全景。
        荧光灯白光与瓷砖冷调。【背景音】只剩水龙头滴水声和压抑的抽泣。

全程画面高清电影纪实风，色调由暖转冷，光影柔和；人物面部稳定不变形，动作自然
流畅，无卡顿无闪烁；不要生成水印、Logo、字幕。

• Shot 1 (Setup): Warm and relaxed, creating contrast as setup
• Shot 2 (Turn): Locked-off camera + facial close-up + sudden sound change, locking onto the change in her eyes
• Shot 3 (Release): Pull-back + curled-up posture + dripping sound, making sadness "visible" rather than merely narrated

Reference Resources

• Storyboard Timing: methodology/03-分镜时序.md
• Emotion Externalization Table: methodology/04-情绪外化表.md
• Special Character Conventions: methodology/07-特殊字符规范.md (music () / SFX <> / dialogue {} / subtitles 【】)
• Practical Examples: Emotional Dormitory Short Drama and Historical-Style Cliffside Confrontation → methodology/02-进阶公式.md