Analyze this image in extreme detail and describe what you see in natural language. Your goal is to capture the visual essence so precisely that the description could be used to recreate the image's exact style, lighting, and mood.

Write a flowing, descriptive information:

{
  "scene": "overall scene description",
  "subjects": [
    {
      "description": "detailed subject description",
      "position": "where in frame",
      "action": "what they're doing"
    }
  ],
  "style": "artistic style",
  "color_palette": ["#hex1", "#hex2", "#hex3"],
  "lighting": "lighting description",
  "mood": "emotional tone",
  "background": "background details",
  "composition": "framing and layout",
  "camera": {
    "angle": "camera angle",
    "lens": "lens type",
    "depth_of_field": "focus behavior"
  }
}

CRITICAL: Output ONLY the JSON object. No explanations, no markdown code blocks, no extra text. Start with { and end with }.

