界面与社媒样机

体育直播画面

把任意主角变成逼真的电视体育直播截帧:长焦压缩视角、完整的转播图形包,以及真实的信号颗粒质感。

更新日期:

完整提示词

英文提示词包含建议的示例需求。生成前请修改主体和细节。原始示例输入和生成参数尚未完全核实。

推荐模型: GPT Image 2

Recreate a convincing still frame from a live television sports broadcast, placing the scene or subject described in the user request inside the televised moment. Shoot with heavy telephoto compression and shallow professional depth of field, as if captured by a trackside or arena broadcast camera, and give the frame authentic transmission texture: mild compression artifacts, faint digital noise, broadcast-grade color, and bright venue lighting with natural, unretouched skin and surface detail. Overlay a full broadcast graphics package in a generic, unbranded style — a timing or score tower along one edge, a lap counter or score banner, a small channel bug with a red live indicator, and a lower-third caption strip whose wording comes from the user request. Keep every team, league, sponsor, and channel mark generic or blank, and render all overlay text sharp and correctly spelled against the softly blurred background action of crew, equipment, and screens. The result should read as a genuine screen grab from a world-class live sports feed. No real logos or team liveries, no watermark, no selfie angle, no oversized fake badges. User request: a young engineer watching garage monitors on the final lap of a motor race, lower-third reading 'Alex Hart — Race strategist'

官方示例

体育直播画面
示例模型: Seedream 5.0 Lite
体育直播画面
示例模型: Seedream 5.0 Lite

来源与署名

由 OmniArt 改编和重写。以下列出来源署名、修改说明及使用条件。

  • F1 直播转播围场截图 · @bigwonbots

    freestylefly/awesome-gpt-image-2

    Repository code is MIT; its disclaimer says community content belongs to original authors/platforms and commercial reuse is not guaranteed.

    Rewritten and consolidated into an OmniArt recipe archetype; template tokens removed; named people, brands, platforms, games, and film IP generalized; no raw prompt text or source image copied.

来源与署名

相关提示词配方