
原文
GPT Image 2 · en
Scene:
A tall 9:16 episode-cover poster inside a mirrored elevator lobby during heavy rain, with gold lift doors, water tracked across marble floor, and cool overhead light reflected in long vertical strips. The composition must remain instantly legible on mobile: a dramatic top third, a tight emotional midsection, and a restrained lower area where platform labels or episode markers could later sit as unreadable graphic blocks. Keep the palette deep charcoal, champagne gold, storm blue, and wine red.
Subject:
Using the referenced lead pair as the primary continuity source, preserve the same heiress and bodyguard identities, their height difference, face structure, hair silhouettes, and distrust-charged attraction. Move them into a rain-soaked elevator standoff: the heiress has just stepped out in a dark red dress under a soaked coat, and the bodyguard blocks the closing doors with one hand while looking directly at her. Their poses should intensify the relationship while clearly remaining the same characters from image 01. Rain droplets, coat texture, and mirrored reflections should add melodrama without cluttering the narrow frame.
Important details:
This image tests whether the poster system can evolve chapter to chapter while staying platform-clean. Keep the visual hierarchy disciplined: top rain atmosphere, center face tension, bottom quiet area for overlay. Preserve character continuity more than styling novelty. The poster should feel like a real vertical-drama app cover that could run as episode two artwork, with room for unreadable hook-badge shapes but no literal text.
Use case:
Vertical Drama Poster System series study 2/9, preserving the reference couple while translating them into a rain-soaked elevator episode cover.
Constraints:
No readable text, no watermark, no logos, preserve exact 9:16 aspect ratio, no cast swap, no extra paparazzi, no broad embrace, no empty white studio backdrop, no comic lightning effects, no overfilled lower frame, no platform UI buttons.使い方
- パラメータを含む元のプロンプト全体をコピーします。
- 被写体、文字、構図を自分の要件に置き換えます。
- 対応する外部ツールで使用し、利用条件と結果を確認します。
開始前の確認
- 参考画像の要否を確認してください。互換モデルの表示は、そのモデルでの生成実績を意味しません。
- 名前、数値、画像内の文字を確認します。モデルの版で結果は変わります。





