Kizomba forehead-to-forehead slow dance — Mandarin Chinese R1

Video reference

Draft title: Kizomba forehead-to-forehead slow dance — Mandarin Chinese R1

Watch external video ↗ · Playback is available at the source.

Reconstructed prompt

Reconstructed from the video, not the author's prompt. This draft has not been generation-tested. Results vary by model and photo.

FORMAT: 10s vertical 9:16 video.
REFERENCE ROLES: @Image1 is the lead dancer in a navy silk slip dress. Synthesized generic dance partner in a black fitted shirt. If non-human, use stylized jaw motion.
SOUNDTRACK SPECIFICATION: 72 BPM, 4/4 sensual bachata/kizomba rap track. Nylon-string guitar, soft percussion, deep sub-bass, and smooth, intimate Mandarin rap vocals performed by @Image1. Performer's lips and jaw precisely synchronize to the Mandarin delivery.
SCENE AND CAMERA: Dimly lit hall, soft blue rim light, warm background bokeh. Medium-close framing, slow creeping push-in maintaining continuous intimate composition.
ORIGINAL LYRICS:
0.0-2.5s: 贴近心跳的律动
2.5-5.0s: 舞步缓慢而从容
5.0-7.5s: 呼吸交织在夜色
7.5-10.0s: 沉醉在这段温柔
PERFORMANCE TARGETS:
0.0-2.5s: Couple in close embrace; gentle side-to-side torso sway on the downbeat while softly rapping first line with synchronized lip movement.
2.5-5.0s: Partner's hand slides slightly across upper back; subtle turn of shoulders while delivering second line.
5.0-7.5s: Downward shift of partner's hand to waist; fluid hip sway aligned to 72 BPM pulse as third line is articulated.
7.5-10.0s: Foreheads draw closer, settling into a deep, gentle sway; delivering final lyric smoothly, easing into close.
PRIORITY: Preserve @Image1 facial features and skin tone; precise lip sync with Mandarin rap; realistic anatomy during tactile contact.
EXCLUSIONS: Out-of-sync mouth shapes, extra limbs, rapid jumps, text, watermarks.

INPUT AND AUDIO BOUNDARY: @Image1 is the sole uploaded photo; no other image, audio or video reference. Preserve the uploaded person/pet identity, hair or fur markings and body proportions, or the reference object’s geometry and materials; do not copy another performer’s identifying appearance. Generic supporting figures are distinct. All BPM, instrumentation and lyrics are newly authored targets, not measurements or quotations of the source. Generate original audio, never clone source music, melody or a named voice. Lips match the generated vocal exactly, never precede it; during instrumental sections the mouth stays relaxed. If the input is a pet use stylized jaw articulation only, never human lips or human teeth. No recognizable copyrighted character appearance, text, logos or watermarks. Do not change the input subject’s gender or age. No generation has been tested.

Copy prompt