
原文
GPT Image 2 · en
Scene:
A square premium key-art frame built for a reference-dependent production series. Stage the same mission at an objective capture or activation moment with clear focal hierarchy with controlled lighting, clear silhouette separation, and a camera setup that feels deliberate enough for campaign reuse. The composition should carry strong focal hierarchy, believable depth, and polished surface behavior while still staying stable enough to support downstream dependent variations.
Subject:
Center the image on the same mission at an objective capture or activation moment with clear focal hierarchy. The result should feel like a concrete, shippable asset for Mission Gameplay Screenshot Sequence, not a generic fantasy render. Build the subject with layered materials, stable anatomy or engineering logic, and enough specificity that later images can inherit the same underlying identity. Preserve the same mission environment, player-facing viewpoint logic, HUD language, and objective flow from the reference gameplay screen from the reference image while changing only what this variation needs. The viewer should immediately understand what is being preserved across the sequence and what is changing in this particular beat.
Important details:
capture progress should be implied through world cues and interface shapes without readable text. Make continuity explicit through repeatable shape language, proportion control, and a coherent palette system. Changes in pose, angle, mood, rarity, weather, or layout should still feel downstream of one stable source. Avoid clutter that would confuse reference use. Any interface-like blocks or graphic surfaces must remain abstract and unreadable. Keep the image polished, original, and suitable for both gallery display and production review.
Use case:
Mission Gameplay Screenshot Sequence series study 3/4, focused on objective capture screen.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no giant completion text, no impossible camera, no broken continuity, no off-model redesign, no accidental franchise resemblance.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。





