
原始文档
GPT Image 2 · en
Scene:
A square promo still in the same floating-bridge world, now cropped tighter and energized by drifting haze ribbons, faint motion blur in fabric edges, and stronger light pulses under the performer’s feet. The environment must remain unmistakably tied to the reference hero frame, just optimized for a looping surface mood.
Subject:
Preserve the exact same performer from the reference image, including face shape, hair base, body proportions, and long coat silhouette. The musician now turns slightly within the frame as if caught mid-loop, while the light bridge and mist layers remain the signature spatial anchors of the streaming campaign.
Important details:
This should feel like a believable canvas still from the same release system. Keep the color world, bridge geometry, and crop discipline consistent, but introduce subtle motion implication through fabric, haze, and light sweep. The image must remain clean enough for platform use and recognizably identical to the reference release identity.
Use case:
Streaming Promo series study 2/9 for the Streaming Banner and Canvas Promo Set Prompt Atlas theme.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, preserve the exact reference face, hair base, and body proportions; no extra characters, no fake playback controls, no chaotic particles, no lens warping, no bright daylight使用方法
- 复制完整原始提示词,并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用,核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型,不代表该示例由这个模型生成。
- 核对名称、数字与画面文字;不同模型版本的结果可能不同。





