skill-mv/templates/storyboard.schema.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "https://git.fmode.cn/fmode-13177980648/skill-mv/templates/storyboard.schema.json",
  "title": "fmode-mv storyboard",
  "description": "分镜数据结构。产线A（音乐MV）与产线B（口播讲解）共用，差异在 shots[].type 与 narration 字段。整条链路（keyframe / video / lipsync / assemble）都消费本结构。",
  "type": "object",
  "required": ["title", "aspect", "totalDuration", "shots"],
  "additionalProperties": true,
  "properties": {
    "title": {
      "type": "string",
      "description": "标题：产线A 为歌名，产线B 为视频标题"
    },
    "aspect": {
      "type": "string",
      "enum": ["16:9", "9:16", "1:1", "4:3"],
      "default": "16:9",
      "description": "画面比例，决定关键帧尺寸与成片分辨率"
    },
    "totalDuration": {
      "type": "number",
      "description": "全片总时长（秒），由各镜头 start+duration 推导"
    },
    "appearanceAnchor": {
      "type": "string",
      "description": "★人物一致性锚点（英文）。每个含人物的镜头都必须原样复述这段描述，否则跨镜头会换脸。",
      "examples": [
        "Asian woman in her late 20s, shoulder-length black hair with soft waves, natural makeup, wearing a deep navy oversized jacket over a white tee, calm melancholic gaze"
      ]
    },
    "styleKeywords": {
      "type": "string",
      "description": "全片统一的影像风格关键词（英文），会拼进每张关键帧 prompt",
      "examples": ["cinematic, teal and orange color grade, neon city night, shallow depth of field, 35mm film grain"]
    },
    "shots": {
      "type": "array",
      "minItems": 1,
      "description": "镜头列表，按时间顺序",
      "items": {
        "type": "object",
        "required": ["id", "type", "start", "duration", "description", "imagePrompt", "videoPrompt"],
        "additionalProperties": true,
        "properties": {
          "id": {
            "type": "integer",
            "minimum": 1,
            "description": "镜头序号，从 1 开始。对应产物文件名 shot-XX.png / shot-XX.mp4"
          },
          "type": {
            "type": "string",
            "enum": ["broll", "perf", "talking"],
            "description": "broll=空镜/场景（无人物主体）；perf=产线A 人物演唱镜头；talking=产线B 讲者镜头"
          },
          "start": {
            "type": "number",
            "minimum": 0,
            "description": "起始时间码（秒）"
          },
          "duration": {
            "type": "number",
            "minimum": 1,
            "description": "镜头时长（秒）。AI 图生视频单段建议 5~8 秒。"
          },
          "section": {
            "type": "string",
            "description": "所属段落：产线A 为 [Verse 1]/[Chorus] 等；产线B 为「开场钩子」等"
          },
          "narration": {
            "type": "string",
            "description": "产线B 专用：该镜头对应的口播原文"
          },
          "subtitle": {
            "type": "string",
            "description": "该镜头字幕文本（建议 ≤20 字，单行显示）"
          },
          "description": {
            "type": "string",
            "description": "中文画面描述（给人看的分镜表正文）"
          },
          "imagePrompt": {
            "type": "string",
            "description": "★英文提示词，必须可独立投喂文生图模型：主体+动作+环境+光线+镜头语言+风格+画质词；含人物时须包含 appearanceAnchor 的英文版"
          },
          "videoPrompt": {
            "type": "string",
            "description": "★英文运镜描述，供图生视频：主体动作 + 镜头运动（push in / pull back / pan / orbit / handheld）+ 节奏。不含画面内容（内容来自关键帧）。"
          },
          "lipSync": {
            "type": "boolean",
            "default": false,
            "description": "是否对该镜头做唇形驱动。仅对 perf/talking 有效；全片建议 ≤4 个（成本和口型风险）"
          },
          "notes": {
            "type": "string",
            "description": "中文备注（可选）"
          }
        }
      }
    },
    "_meta": {
      "type": "object",
      "description": "生成元信息（自动写入，勿手工编辑）",
      "properties": {
        "createdBy": { "type": "string" },
        "generatedAt": { "type": "string", "format": "date-time" }
      }
    }
  }
}