PromptweetPromptweet
Library / Fashion / Korean Fashion MV: Red Collage Character Lock

Korean Fashion MV: Red Collage Character Lock

FashionMusic videoMiniMax H3

A 15-second fashion music video featuring one Korean-style woman moving through bold red graphic collage layouts, magazine typography, and geometric patterns while showcasing eight distinct high-end fashion looks.

Prompt461 copies
Create a 15-second, 16:9 horizontal, 24fps high-quality fashion MV based on the single uploaded character reference. CHARACTER LOCK — HIGHEST PRIORITY Use @Image1 as the only identity reference. Keep one young adult Korean-style woman only throughout the entire video. Strictly preserve her exact face, facial proportions, skin tone, age, long dark-brown hair, hairstyle and bangs, body proportions, limb proportions, and overall identity. She must always remain the same person. No face swap, identity drift, Westernized face, male features, second character, background people, or multiple people. Her vibe: Seoul fashion muse, high-end Asian beauty, sweet but sharp, photogenic, trendy, campaign-ready. STYLE RED PATTERN COLLAGE POP × KOREAN FASHION EDITORIAL × FASHION MV. She moves through a bold visual world built from red graphic patterns, fashion posters, magazine typography, geometric layouts, and collage compositions. Keep the woman photorealistic. Background graphics, typography, blocks, and decorative elements should be clean, sharp, premium 2D design — like Korean fashion magazines and moving campaign posters, not handmade scrapbook collage. Palette: bright red, cherry red, burgundy, deep red, cool white, black, small touches of silver-gray. Keep it modern, chic, controlled, and high-fashion. GRAPHIC SYSTEM Use checkerboards, stripes, dots, grids, frames, diagonal cuts, geometric blocks, cropped shapes, pattern bands, oversized typography, repeated keywords, magazine-cover layouts, and graphic strips crossing the frame. Graphics should actively interact with composition, not sit passively in the background. Suggested text: NAME / RED SIGNAL / OWN THE LOOK / AFTER GLOW / ICON GIRL / TREND SHIFT / SEOUL CUT / BOLD MOOD / FASHION LINE Typography must stay readable, never cover the face, and vary in scale, direction, and placement while maintaining one unified visual system. FASHION Do not copy the reference pose or background. Extend her identity into Korean fashion styling: fitted black/white/red tops, mini skirts, leather details, cropped jackets, tall boots, silver waist chains, earrings, and small bags. Looks may vary slightly, but she must remain one continuous character. MOTION & CAMERA This must feel like a real fashion MV, not a slideshow. Each section: start pose → one clear fashion action → stronger ending pose. Use natural editorial gestures: walking forward, turning, looking back, touching hair, adjusting accessories, hand on waist, leaning, raising eyes, or holding a hero pose. Expression: cool, confident, stylish, slightly distant. Use fast but clean cuts, subtle push-ins/pull-outs, lateral camera movement, and alternating close-up / medium / full-body framing. TIMELINE 0:00–0:01.6 — LOOK 01 Close or medium shot. She slowly raises her eyes toward camera. Bold red pattern + oversized title. Strong first-second hook. 0:01.6–0:03.2 — LOOK 02 Medium/full shot. Half-step forward, then look back. Red-white blocks, checkerboard, geometric frames. "RED SIGNAL / OWN THE LOOK." 0:03.2–0:04.8 — LOOK 03 Medium-close. Sitting or leaning pose, hand near face. Diagonal poster modules + repeated small typography. 0:04.8–0:06.4 — LOOK 04 Turn slightly, touch hair, return gaze to camera. Strong magazine-cover layout with enlarged graphic elements. 0:06.4–0:08.0 — LOOK 05 Slight low angle. Lean toward camera or stand with hand on waist. Oversized text and graphic bands cross the frame. 0:08.0–0:09.6 — LOOK 06 Close/medium. Lower gaze, then look up confidently. Cleaner background with "ICON GIRL / TREND SHIFT." 0:09.6–0:11.2 — LOOK 07 Full-body. Turn, shift weight, finish in a strong cover pose. Slight lateral camera move. 0:11.2–0:12.8 — LOOK 08 Medium shot. Adjust sleeve or accessory. Mature red-white-black poster composition. 0:12.8–0:15.0 — FINAL HERO Strong central or off-center hero pose. Recombine earlier red patterns, typography, and geometric modules into one final campaign cover. Hold the final frame for about 0.5s. MUSIC Sharp K-pop / fashion-pop instrumental, 115–125 BPM, no vocals. Clear beat cuts. Optional subtle whoosh, paper-slide, click, or fabric-swish sounds. NEGATIVE No extra people, identity drift, face change, body-proportion change, Westernization, slideshow feel, stiff motion, one-shot framing only, cluttered realistic backgrounds, unreadable text, face obstruction, childish collage, cheap filters, logos, watermarks, subtitles, exaggerated expressions, or excessive dancing.
Includes a consistency lock: a trait the prompt asks the model to hold constant.
How to use this prompt
  1. 1
    Copy the prompt

    Use the copy button above to grab the text exactly as written.

  2. 2
    Paste it into MiniMax H3

    Drop it in unchanged since the camera, lighting and duration are already specified.

  3. 3
    Swap the subject, keep the structure

    Change only the subject or setting for the most predictable result; the same wording carries over to other models too.

Aspect ratio
16:9
Camera
close-up · medium shot · full-body · push-in · pull-back · low-angle · lateral tracking · static/fixed
Lighting
high-key · studio lighting · cinematic · bright
Mood
confident · editorial · stylish
Duration
15s
Difficulty
Advanced
Model
MiniMax H3
Includes
Subject lock

More from MiniMax H3

What makes this prompt work

Every claim below is measured against the 627 prompts in this library, not asserted.

Camera

‘close-up’ is used by 482 prompts in the library, rank #1 of the camera terms seen here.

Lighting

‘high-key’ appears in 71 prompts library-wide, rank #13.

Aspect ratio

16:9 is used by 472 of the 627 prompts in the library. Naming it at all matters more than which one is picked.

Consistency

This prompt declares a consistency lock, which is what keeps a subject recognisable across a clip.

Model

MiniMax H3 accounts for 11% of the library (69 of 627 prompts).

Difficulty

Marked advanced, based on how many specs it stacks (camera, lighting, duration, lock) at once.