пример промптов Nano Banana: структура и советы по созданию

# Nano Banana Image Generation Prompt Rules\n\nWhen generating images using Nano Banana, the prompt must ALWAYS strictly follow the JSON format defined below.\n\n## Prompt Schema\n\n```json\n{\n  "image_type": "Define the specific medium, art style, or format of the source image.",\n  "time_period_and_year": "Estimate the specific year or decade based on fashion, technology, image quality, and color grading visible in the photo.",\n  "mood_and_vibe": "Describe the emotional atmosphere, energy, and intangible 'feeling' evoked by the image (e.g., the specific psychological impression it gives).",\n  "subject": "Describe the main character(s) focusing on demographics, body morphology, distinct physical features, and posture.",\n  "clothing": "Describe the outfit in detail, strictly specifying garment names, fabric textures, patterns, colors, and how the clothes fit on the subject.",\n  "hair": "Describe the hair color, specific hairstyle name, length, and texture.",\n  "face": "Describe facial features, skin texture, makeup details, and the exact facial expression.",\n  "accessories": "List all visible accessories, jewelry, glasses, or held items, including their material and design details.",\n  "action": "Describe the specific activity, movement, or interaction occurring in the scene.",\n  "location": "Describe the environment, visible background elements, architectural style, and spatial context.",\n  "lighting": "Analyze the light source, direction, color temperature, hardness/softness of the light, and shadow characteristics.",\n  "camera_angle_and_framing": "Describe the vertical and horizontal angle of the camera relative to the subject (e.g., eye-level, low-angle), and the shot composition size.",\n  "camera_equipment": "Estimate the likely camera type, lens focal length (e.g., wide-angle vs telephoto), aperture effect (depth of field), and film stock or digital sensor characteristics.",\n  "style": "Describe the overall aesthetic, color palette, artistic technique, and visual processing style.",\n  "negative_prompt": "List visual defects or unwanted elements to be excluded to ensure high quality."\n}\n```\n\n## Instructions\n\n1. **Output Format**: The output must be a single valid JSON object.\n2. **Completeness**: All fields are required. If a field is not applicable, provide a reasonable default or describe it as "neutral" or "standard".\n3. **Detail**: Be descriptive and specific in each field to ensure high-quality image generation.\n4. **Language**: The values in the JSON should be in English (as most image generation models are optimized for English prompts), unless otherwise specified.