影片 圖片
登入
懷舊日本郊區人像

懷舊日本郊區人像

A highly detailed cinematic prompt for creating nostalgic Japanese summer film-style portraits in a suburban setting during the golden hour.

觀看4,000
按讚89
留言49
分享4
書籤25
引用0
翻譯
{
  "subject": "uploaded_image",
  "scene": {
    "location": "{argument name="location" default="窄小的花徑在郊區日本社區"}",
    "environment": "濃密的鮮豔橙色加州罌粟田,電線桿,架空電線,遠處住宅建築",
    "season": "晚春",
    "time_of_day": "金色時刻"
  },
  "composition": {
    "aspect_ratio": "2:3",
    "camera_angle": "低角度,稍微向上",
    "framing": "全身肖像坐在地上",
    "subject_position": "居中",
    "foreground": "大型失焦橙色花朵創造夢幻深度",
    "background": "柔和模糊的郊區街景",
    "depth_of_field": "非常淺"
  },
  "subject_details": {
    "pose": "隨意坐在土路上,膝蓋向外彎曲,雙手鬆散地交叉在腿間",
    "expression": "放鬆,沉思,略微夢幻",
    "gaze": "看向相機",
    "hair": "短黑鮑勃髮型,劉海蓬鬆",
    "clothing": {
      "top": "{argument name="top clothing" default="白色羅紋背心"}",
      "bottom": "寬鬆藍色牛仔褲,褲腳捲起",
      "shoes": "磨損的帆布運動鞋"
    },
    "accessories": [
      "圓形復古太陽眼鏡,帶有溫暖的橙色鏡片"
    ]
  },
  "lighting": {
    "type": "自然陽光",
    "quality": "柔和,發光,漫射",
    "direction": "低側光",
    "color_temperature": "溫暖的金橙色",
    "highlights": "皮膚和花朵周圍的溫柔光暈"
  },
  "style": {
    "photography": "電影感肖像攝影",
    "aesthetic": "懷舊日本夏季電影",
    "mood": "{argument name="mood" default="夢幻,年輕,寧靜,浪漫"}",
    "color_palette": [
      "溫暖橙色",
      "金黃色",
      "柔軟藍色",
      "奶油白"
    ]
  },
  "camera": {
    "lens": "50mm復古鏡頭",
    "aperture": "f/1.4",
    "focus": "臉部清晰,整個畫面柔軟衰減",
    "effects": [
      "鏡頭光暈",
      "光暈效果",
      "膠片顆粒",
      "微妙色差",
      "柔焦發光",
      "模擬膠片質地"
    ]
  },
  "quality": {
    "resolution": "高",
    "detail": "照片級真實",
    "rendering": "自然的皮膚質地,真實的花朵細節,電影感深度"
  },
  "negative_prompt": [
    "過度銳化",
    "HDR外觀",
    " harsh contrast",
    "塑膠皮膚",
    "額外肢體",
    "扭曲解剖",
    "低分辨率",
    "浮水印",
    "文字",
    "標誌",
    "過飽和色彩",
    "人工照明"
  ]
}

機器翻譯 — 實際產出作品的是原文 prompt。

✨ 生成 原始貼文 在 Hitboard 開啟

如何使用這個 prompt

複製 prompt 貼進 GPT Image 2。構圖、光線與風格詞請照原文保留,那是產出這個結果的關鍵;換掉主體或場景細節讓它變成你的,比例維持不變,若 prompt 提到參考圖就一併附上。多生成兩三次,最好的版本通常不是第一張。