冷调走廊|极近怼脸后拉

起幅是极近怼脸,镜头实拍后拉到半身,冷灰走廊配钴蓝高领毛衣,毛孔与碎发细节都保留。10 秒横版,720P,提示词可直接复制复现。

在 YouTube 上观看

原始视频提示词(Video Prompt)

EN

提示词原文(一字未改)

[SPEC] 10 seconds, 16:9 landscape, 24fps, 4K, one continuous take, no cuts.

[CHARACTER LOCK] A clearly adult 27-year-old East Asian woman with distinctly East Asian features: cool, composed look, soft oval face, almond-shaped eyes with subtle double eyelids, a fine straight nose bridge, full lips, and polished camera-ready makeup. Long jet-black hair parted near the center, falling naturally over her shoulders with a few loose strands framing her cheeks. Keep the same face, outfit, hairstyle and body proportions for the entire clip.

[SKIN & MAKEUP] Photoreal skin: visible pores, texture along the nose wings, faint freckles and uneven tone on the cheeks, real under-eye layers, fine facial peach fuzz, natural oil highlights, real lip lines. Thin luminous base, clean eye makeup, individually defined lashes, no exaggerated false lashes.

[COSTUME] Cobalt-blue ribbed turtleneck sweater, no jewelry; collar stitching and knit texture clearly visible.

[SETTING & LIGHT] Minimal modern indoor corridor, light grey walls. One defocused round white ceiling light behind each side of her head. Neutral soft key light from slightly above and in front, even exposure on the face, small realistic highlights on the nose bridge, nose tip, lips and eyes; low-contrast, slightly defocused background.

[CAMERA & PERFORMANCE]
0-3.3s: 85mm portrait lens, camera 35-50cm from her face, the face almost filling the frame. She looks straight into the lens, lips naturally closed, calm and cool. The camera physically dollies back very slowly (no digital zoom), from an extreme facial close-up to a standard head-and-shoulders shot, gradually revealing the sweater shoulder line and her long hair.
3.3-4.8s: her lips part almost imperceptibly, like one natural breath, no talking.
4.8-6.7s: the eyes move first - dark brown pupils drift slowly downward, upper lids follow naturally, and only then does the head tilt slightly down, creating a real delayed follow.
6.7-7.8s: holds the lowered gaze, quiet and thoughtful.
7.8-10s: pupils return slowly to the lens first, then the head lifts, re-establishing eye contact; the final half second holds a calm, restrained micro-expression.

[TEXTURE] Only a very slight handheld breathing motion throughout; no digital zoom, no sudden push/pull, no showy AI camera moves.

[NEGATIVE PROMPT] face swap, drifting features, influencer-template face, Western/Caucasian bone structure, non-Asian face, over-smoothed skin, plastic skin, exaggerated highlights, overly heavy false lashes, misaligned eyes, asymmetric pupils, hair clipping, knit texture flicker, background flicker, overexposure, subtitles, watermark, text.

上方徽标显示的是原始提示词的语言。模型对不同语言的理解存在差异,同一个创意换一种语言书写,成片效果可能不同。

复制这条提示词,免费在线生成同款视频

生成参数

任务类型
简单视频
模型
agnes-video-2.5-flash
时长
10s
分辨率
1280x720
提示词语言
EN
发布时间
2026-11-09

做出作品了?来分享

在 GitHub Discussions 发帖分享你的视频和背后的提示词。视频始终留在你自己的频道:我们只做官网精选展示(署名并回链你的主页),并可能把它收进官方 Community Showcase 播放列表。

去 GitHub Discussions 投稿

投稿由人工筛选;我们不会把你的视频重新上传到官方频道。

更多视频提示词灵感