Yt Storyboard

1. Each narration sentence = potential cut point
2. Group 1-3 sentences into one scene if they share visual context
3. Scene duration: min 2s, max 8s for shorts / max 15s for long-form
4. First scene must be visually impactful (matches hook)
5. Scene transitions should feel natural — avoid jarring visual jumps between related content

Duration Estimation

• Korean narration speed: ~3.5 syllables/second (shorts), ~3.0 syllables/second (long-form)
• Auto-calculate scene duration from narration text length
• Flag scenes that exceed recommended duration for their type
• Total duration must match target format length

Pipeline Steps

1. Scene Split: Break script into scenes based on content boundaries
2. Visual Assignment: Assign visual type/layout per scene
3. Image Prompt Generation: Create AI image prompts (English) for scenes needing generated images
4. Subtitle Generation: Create timed subtitle entries with highlight words
5. Transition Planning: Assign transitions between scenes
6. Audio Planning: BGM cues, SFX placement
7. Camera Movement: Assign camera movements per scene
8. Asset List Compilation: List all needed images, b-roll, icons, charts
9. SRT Preview: Generate subtitle file preview

Scene Type Templates

Hook Scene

• Bold text overlay, dynamic background, zoom-in camera
• Short duration (2-4s)
• Transition: cut or zoom

Explanation Scene

• Diagram/infographic, text bullets, static camera
• Medium duration (4-8s)
• Transition: fade or slide

Example Scene

• Real-world image/b-roll, minimal text, ken_burns camera
• Medium duration (3-6s)
• Transition: cut

Comparison Scene

• Split screen layout, side-by-side elements
• Medium duration (4-8s)
• Transition: slide_left

Chart/Data Scene

• Chart/graph visualization, number emphasis
• Medium duration (3-5s)
• Transition: fade

Quote Scene

• Text-centered, subtle background, fade transition
• Short duration (3-5s)
• Transition: fade

CTA Scene

• Subscribe/like graphics, channel branding
• Short duration (2-4s)
• Transition: slide_up

Visual Style Presets

Each channel can define visual presets in channels/{channel_id}/config.json:

{
  "visual_style": {
    "preset": "minimal|cinematic|infographic|meme|documentary",
    "color_palette": ["#1a1a2e", "#e94560"],
    "subtitle_style": "modern_bold|classic|typewriter|handwritten",
    "default_transitions": "cut|fade|slide",
    "background_preference": "dark|light|gradient|image_heavy",
    "text_density": "low|medium|high"
  }
}

When no channel config exists, use defaults: minimal preset, dark background, modern_bold subtitles, cut transitions.

Output Format

Generate a storyboard JSON with this structure:

{
  "storyboard_id": "sb_20260316_001",
  "script_id": "script_20260316_001",
  "channel_id": "tech-shorts",
  "format": "shorts",
  "aspect_ratio": "9:16",
  "total_duration_seconds": 52,
  "scenes": [
    {
      "scene_id": "scene_001",
      "scene_number": 1,
      "type": "hook|explanation|example|comparison|chart|quote|cta",
      "time_range": {
        "start": 0.0,
        "end": 3.2
      },
      "narration": {
        "text": "나레이션 텍스트",
        "word_count": 12,
        "estimated_duration_seconds": 3.2
      },
      "visual": {
        "description": "화면 설명 (한국어)",
        "layout": "full_image|split_screen|text_overlay|chart|comparison_lr|zoom_in|zoom_out",
        "background": {
          "type": "solid_color|gradient|image|video_loop|b_roll",
          "value": "#1a1a2e",
          "image_prompt": "AI 이미지 생성용 프롬프트 (English)",
          "b_roll_keywords": ["keyword1", "keyword2"]
        },
        "text_overlay": {
          "main_text": "화면에 표시할 핵심 텍스트",
          "sub_text": "",
          "position": "center|top|bottom",
          "style": "highlight|normal|emphasis|question"
        },
        "elements": [
          {
            "type": "icon|chart|arrow|number|image",
            "description": "요소 설명",
            "position": "center|left|right|top|bottom"
          }
        ]
      },
      "subtitle": {
        "text": "자막 텍스트",
        "highlight_words": ["강조할", "단어"],
        "style": "default|emphasis|whisper|shout"
      },
      "transition": {
        "type": "cut|fade|slide_left|slide_up|zoom|none",
        "duration_ms": 300
      },
      "audio": {
        "bgm_action": "continue|fade_in|fade_out|change|silence",
        "sfx": "whoosh|pop|ding|none",
        "sfx_timing": "scene_start|with_text|scene_end"
      },
      "camera": {
        "movement": "static|slow_zoom_in|slow_zoom_out|pan_left|pan_right|ken_burns",
        "focus_point": "center|text|subject"
      }
    }
  ],
  "global_settings": {
    "subtitle_style": {
      "font": "Pretendard Bold",
      "size": 48,
      "color": "#FFFFFF",
      "stroke": "#000000",
      "highlight_color": "#FFD700",
      "position": "bottom_third"
    },
    "bgm": {
      "track": "bgm_name_or_path",
      "volume": 0.15,
      "fade_in_seconds": 1.0,
      "fade_out_seconds": 2.0
    },
    "color_palette": ["#1a1a2e", "#16213e", "#0f3460", "#e94560"]
  },
  "asset_list": {
    "images_needed": [
      {
        "scene_id": "scene_001",
        "prompt": "AI generation prompt",
        "style": "realistic|illustration|diagram|infographic",
        "dimensions": "1080x1920"
      }
    ],
    "b_roll_needed": [
      {
        "scene_id": "scene_003",
        "keywords": ["technology", "server room"],
        "duration_seconds": 4
      }
    ],
    "icons_needed": ["chart_up", "warning", "lightbulb"],
    "charts_needed": [
      {
        "scene_id": "scene_004",
        "type": "bar|line|pie|comparison",
        "data_description": "차트 데이터 설명"
      }
    ]
  },
  "srt_preview": "1\n00:00:00,000 --> 00:00:03,200\n첫 번째 자막\n\n2\n..."
}

Commands

• yt-storyboard {script_id} — generate storyboard from script
• yt-storyboard "대본 텍스트" --channel tech-shorts — from raw text
• yt-storyboard adjust {storyboard_id} --scene 3 --visual "change to comparison layout" — adjust specific scene
• yt-storyboard assets {storyboard_id} — show asset checklist
• yt-storyboard srt {storyboard_id} — export SRT only
• yt-storyboard preview {storyboard_id} — show visual preview (text-based scene descriptions)
• yt-storyboard list — show recent storyboards

Output Location

• Storyboard JSON: storyboards/{channel_id}/{topic_slug}/storyboard.json
• Human-readable visual: storyboards/{channel_id}/{topic_slug}/storyboard_visual.md
• Subtitle file: storyboards/{channel_id}/{topic_slug}/subtitles.srt

Integration

• Reads from: yt-script output (scripts/{channel_id}/{topic_slug}/script.json)
• Feeds into: yt-compose (video assembly), image generation skills
• Channel config: channels/{channel_id}/config.json for visual style presets
• Image generation: Image prompts can feed into maple-cover/maple-background skills or generic image generation