GPT Image 2.0 プロンプト - ローアングル夜街の動作スナップ肖像
- アスペクト比
- auto
- 画質
- 1K
プロンプト
[Location]: subway exit / vending machine / bus stop / convenience store
[Outfit]: jacket + fitted top + mini skirt/shorts + knee-high socks + sneakers
[Action]: crouching, sitting low, bending down, fixing hair/shoes, caught mid-motion
Create a 9:16 ultra-realistic night portrait of an adult East Asian woman in her early 20s. Use a close 28–35mm perspective with a slightly low camera angle, letting one knee or leg enter the foreground for stronger depth.
The body should stay in an unfinished everyday action rather than a perfect model pose — low center of gravity, natural hip/waist twist, loose shoulders, hair slightly messy from movement, then suddenly looking directly into the camera.
Use harsh on-camera direct flash as the main light. Keep the subject much brighter than the dark environment, with realistic highlights on skin, hair and fabric, subtle high-ISO grain, imperfect nighttime exposure, and natural background lights.
Prioritize real skin texture, visible pores, natural flyaway hair, believable clothing folds and anatomy. Avoid plastic skin, beauty filters, studio lighting, CGI look, stiff posing, extreme wide-angle distortion, or overly cinematic grading.
この画像プロンプトを使う
- 01
プロンプトから始める
プロンプトをコピーするか、生成ボタンで開きます。
- 02
自分のアイデアを加える
被写体や背景、細部を調整し、必要に応じて参考素材を追加します。
- 03
生成前に確認する
モデルと出力設定を確認してください。生成結果は毎回異なる場合があります。







































