
原文
GPT Image 2 · en
{
"type": "Live-action movie posters and composite character reference images",
"layout": {
"top_half": {
"type": "Cinematic Poster",
"visuals": "The foreground features a young Japanese man holding an orange tabby cat, while the background shows a young Japanese woman looking to the right, against a golden hour city skyline with bridges and a faint rainbow.",
"typography": {
"title": "{argument name=\"movie title\" default=\"The promise of the Rainbow Bridge\"}",
"subtitle": "NIJI NO HASHI NO YAKUSOKU",
"release_date": "{argument name=\"release date\" default=\"11.7 FRI roadshow\"}",
"taglines": "Multiple blocks of Japanese text arranged vertically and horizontally."
}
},
"middle_band": {
"type": "film strip",
"count": 5,
"description": "A horizontal row consisting of 5 movie stills showcasing the characters in different scenes."
},
"bottom_half": {
"type": "Character design sketches",
"sections": [
{
"subject": "female characters",
"name": "{argument name=\"female character name\" default=\"Haru Fujiki\"}",
"full_body_poses": 4,
"pose_labels": [
"front",
"Side view (left)",
"back",
"3/4 side view"
],
"facial_expressions": 3,
"detail_shots": 3,
"footwear_angles": 3,
"silhouette_postures": 2,
"color_swatches": 5
},
{
"subject": "Male characters",
"name": "{argument name=\"male character name\" default=\"Rin Segawa\"}",
"full_body_poses": 4,
"pose_labels": [
"front",
"Side view (left)",
"back",
"3/4 side view"
],
"facial_expressions": 3,
"detail_shots": 3,
"footwear_angles": 3,
"silhouette_postures": 2
},
{
"subject": "Orange tabby cat",
"name": "{argument name=\"cat name\" default=\"チェリー\"}",
"full_body_poses": 3,
"large_headshot": 1,
"small_headshots": 4,
"detail_shots": 3,
"color_swatches": 4,
"scale_reference": 1
},
{
"subject": "Additional stills",
"count": 6,
"description": "A grid consisting of 6 movie stills featuring characters and cats."
}
]
}
}
}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。





