AI image generated from the “Video - Three Kingdoms ARPG - Lyu Bu Yuanmen Archery (Seedance 2.0)” prompt
Result produced with GPT Image 2 · open-design

Video - Three Kingdoms ARPG - Lyu Bu Yuanmen Archery (Seedance 2.0)

视频模板 - 电影叙事

The prompt4550 chars
A ~10 second in-engine cinematic ARPG action sequence, photoreal, Unreal-Engine-5-grade render quality, desaturated filmic color grading, shallow depth of field on the background. Strict in-game camera discipline — one continuous beat, no cuts, no swish pans, no overlays drawn on top of the frame.

# Scene (reference image: {argument name="reference_image" default="game-screenshot-three-kingdoms-lyubu-yuanmen-archery"})
- Setting: {argument name="environment" default="a dusty late-Han military encampment (辕门) between two facing armies, earthen ramparts and wooden palisades framing a wide parade ground, tattered red and black war banners fluttering on tall poles, distant rows of spearmen in lacquered leather armor standing in formation, low golden-hour sun cutting horizontal god rays through drifting dust"}.
- Hero (center of frame, stays identical from first to last frame): {argument name="hero" default="Lyu Bu (吕布), a tall broad-shouldered general in ornate red-lacquered lamellar armor with a phoenix-crest helmet and a purple-and-gold sash, long tied-back black hair, thick sideburns, serious piercing eyes, powerful warrior frame"}.
- Prop: {argument name="weapon" default="a tall red-lacquered recurve longbow with gold-filigree grip, a single long-shafted bronze-tipped war arrow nocked on the string"}.
- Target in the deep background, visible through haze: {argument name="target" default="a single upright halberd planted head-down in the sand roughly 150 paces down the range, a small red ribbon tied near its blade"}.

# Action (three micro-beats across ~10s, single continuous shot)
1. 0.0 - 3.5s  Draw + hold. Hero is already in the frame as the shot opens. He pulls the bowstring smoothly back to anchor next to his cheek; the bow arches into a tight half-moon; his stance widens slightly; the ribbon on the arrow shaft flutters in the wind; his breath is held, eyes locked on the distant target; dust whirls slowly around his boots. Camera is locked-off or breathes with a micro-dolly of less than 5cm — no full camera move.
2. 3.5 - 6.5s  Charge. A faint {argument name="qi_color" default="warm amber-gold"} qi glow builds along the arrow shaft and up the bowstring; tiny embers of qi drift off the arrowhead; the bow creaks subtly; a banner behind the hero snaps once in the wind; the two armies in the far background remain still and silent, watching; no muzzle flash, no lens flare, no UI text pops.
3. 6.5 - 10.0s  Release + follow. A crisp bowstring snap; the arrow streaks cleanly across the frame in a single golden trail, leaving a very short motion trail (less than a third of the frame); camera stays locked on the hero's stance recovery — shoulders drop, drawing hand recoils, the bow returns toward a neutral hold; in the deep background the halberd's red ribbon reacts almost imperceptibly (a single pluck of the ribbon is enough — do not show the arrow striking).

# Technical direction
- Framing: hero occupies roughly the center-left third of the frame for the entire shot; horizon line low; target visible as a small silhouette far right-deep.
- Lighting: cinematic golden-hour key from the left, cool teal fill on the right, volumetric god rays + fine dust particles drifting left-to-right, ~5fps dust parallax in the deep field, subtle bloom on the qi glow only.
- Motion: no camera shake on release, no match-cuts, no slow-motion ramps; the release is in real time at normal speed. Physics plausible — bowstring tension, arrow flex, armor plates click, sash flutters.
- Sound in spec (even if the exported file has no audio): draw creak, one inhale, bowstring snap, arrow whistle fading into wind. Do not render any on-screen sound-effect text.
- Color: desaturated filmic grade, deep warm shadows, slight filmic grain, subtle chromatic aberration at frame edges only.
- HUD: none rendered in this shot (the companion image template carries the HUD; this video is a pure in-engine action beat that can be overlaid with HUD in post or paired with the still).

# Negative prompt
no anime cel-shading, no comic speed lines, no particle fireworks, no explosion, no blood spray, no arrow impacting the target on-screen, no slow-motion ramp, no whip-pan, no lens flare, no watermark, no on-screen text or UI, no warped Chinese characters or garbled glyphs on banners, no modern clothing, no firearms, no Western medieval armor, no duplicated hero, no crowd extras wandering in the foreground, no extra fingers on the drawing hand, no broken bow geometry, no floating arrow without proper nock, no jump cut.

How to use it

game-cinematic, arpg, three-kingdoms, ancient-china, combat, archery, lyubu, key-visual, hud-safe, companion-to-image

Swap these placeholders

  • argument name="qi_color" default="warm amber-gold"

Replace each bracketed token with your own subject before generating — leaving them in place makes the model render the literal words.

Credit

Collected from awesome-gpt-image2-prompts · originally shared by open-design.