Tang Dynasty Realistic Performance Video
Full prompt
FIRST FRAME (= upload reference image, only for the starting posture, character moves naturally afterwards): Tang Dynasty beauty, black hair in a high bun, a red peony pinned next to a golden hair ornament, floral decoration on the forehead, Cream-white thin silk shawl over a crimson orange gold-painted strapless ruqun, hands naturally folded in front of her; Peony bushes beside her, Forbidden City towers behind, warm golden twilight light, floating dust in the air. Character features, makeup, and clothing consistent with the first frame, body naturally relaxed throughout, not stiff. Core requirement: Real-life documentary performance quality, removing all 'AI flavor'— Real skin pores and fine hair, oiliness and breathing fluctuations, Micro-expressions with imperfect realism (occasional wandering gaze, lips moving before smiling, Irregular blinking rhythm, occasional subtle movements like swallowing/pursing lips/adjusting center of gravity). Storyboard script (few lines, focus on real detail performance): SHOT 1 | 0-4s | Medium-close shot (handheld camera · very slight breathing shake) She looks down at her folded hands, fingers unconsciously fiddling with her belt, Weight shifting from one foot to the other, shoulders relaxing; Then her attention is drawn to the peonies beside her, eyes naturally lifting and glancing sideways, eyes lead, then head follows. No lines, only natural breathing and ambient sound (wind, distant birds). SHOT 2 | 4-7s | Facial close-up (eye acting) Camera slightly pushes into the face: irregular blinking, subtle focus shifts in pupils, Nostrils fluttering slightly with breath, lips pursing once before breaking into a real faint smile, The smile starts from the eyes—not a formulaic standard smile. [Almost breath voice · one line] "These flowers... bloom so well." SHOT 3 | 7-11s | Half-body dynamic She raises her hand to lightly touch the peony petals, fingertips sinking realistically into the softness, Wrists rotating naturally, silk sleeves slipping down a bit, she subconsciously pulls the sleeves back with the other hand, Body turning slightly, loose hair blown by the wind sticking to her cheek, she ignores it. The frame is full of real, non-deliberate small movements. No lines. SHOT 4 | 11-15s | Return to medium-close shot (conclusion · natural freeze) She retracts her hand, gaze falling forward again, exhaling lightly, Eyes carrying a trace of real melancholy and relief, lips moving as if to say something but ending in silence, Finally, a smile that isn't perfect but very real, naturally stopping at that moment. [Breath voice · fading · like self-talk] "...It's fine." Audio setting: female voice, natural and lifelike, more breathy, speed uneven with emotion, Not overly articulate, with realistic breathing pauses; only two very short whispers throughout. CAMERA: Handheld texture (extremely slight natural shaking and breathing sensation), multi-shot: Medium-close / Facial close-up / Half-body / Return to medium-close, natural switching without showing off; Real shallow depth of field, focus occasionally has subtle breathing-style refocusing, background palace and peonies blurred. Atmosphere · Lighting: Real natural twilight light, side-backlight edging hair in gold, warm tones; Floating dust, petals occasionally falling, silk scarf and hair moving with real breeze, Overall like film photography, fine grain, soft highlights without overexposure. CONSTRAINTS: - 4K Ultra HD, 15 seconds, character makeup and clothing consistent with the first frame (changing shots not faces or clothes). - Core: Real-life documentary performance—real skin texture, irregular blinking, natural weight shifting, micro-expressions with 'imperfect realism', strictly no stiffness or plastic fake faces. - De-AI: Avoid over-smoothing, excessive symmetry, over-fluid movement, and perfect smiles. - Two lines of dialogue strictly synchronized with mouth movements, lifelike breathy quality, no lip-sync errors. - Wind, falling flowers, silk sleeves, and loose hair follow real physics, natural and not exaggerated. - Eyes have emotion and real focus shifts; no glowing eyes or abnormal pupils. - Ends naturally on the resonance of small movements, no black screen, no transition. - No subtitles, no watermark, no symbols, no text.
Original prompt
首帧 FIRST FRAME(=上传参考图,仅作起手姿态,之后人物自然活动): 唐风美人,乌发高髻、金步摇旁簪一朵红牡丹,眉心花钿, 米白薄纱外披罩绛橙描金抹胸襦裙,双手在身前自然交叠; 身畔牡丹丛、身后紫禁城楼台、黄昏暖金光、空气微尘浮动。 人物五官妆容服饰与首帧一致,肢体全程自然松弛,不僵直。 核心要求:真人纪实表演质感,去掉一切「AI 味」—— 皮肤有真实毛孔与细绒、油光与呼吸起伏, 微表情有不完美的真实感(眼神偶尔游移、嘴角先动一下再笑、 眨眼节奏不规律、偶有极轻的吞咽/抿唇/调整重心的小动作)。 分镜脚本(少台词·重真实细节表演): SHOT 1|0-4秒|中近景(手持镜头·极轻微呼吸晃动) 她低头看着交叠的双手,指尖无意识地轻捻衣带, 重心从一只脚换到另一只,肩膀松一下; 随后被身旁牡丹引去注意,自然抬眼侧望,眼神先到、头才转过去。 无台词,仅自然呼吸与环境声(风、远处鸟鸣)。 SHOT 2|4-7秒|面部特写(眼神戏) 镜头轻推到面部:眼睫不规律地眨动,瞳孔有细微对焦的游移, 鼻翼随呼吸极轻翕动,唇角先抿了一下、才漾开一个真实的浅笑, 笑意先从眼睛起来——不是程式化的标准微笑。 〔几乎气声·一句〕"这花……开得真好。" SHOT 3|7-11秒|半身动态 她抬手轻触身旁牡丹花瓣,指腹真实地陷进花瓣的软, 手腕自然转动,纱袖滑下一截,她下意識用另一只手把袖子撩回去, 身子随之微微側轉,碎髮被風吹亂、貼在臉頰,她不去理。 滿畫面都是真實的、不刻意的小動作。無台詞。 SHOT 4|11-15秒|回到中近景(收束·自然定格) 她收回手,目光重新落回前方,輕輕吐了口氣, 眼神裡帶著一點真實的悵然又釋然,嘴唇微動想說什麼、終究沒說, 最後是一個沒那麼完美、卻很真實的淺笑,自然停在那一刻。 〔氣聲·漸弱·像自語〕"……也好。" 聲音設定:女聲,自然生活化,氣聲偏多,語速隨情緒不勻, 不字正腔圓、略帶真實的呼吸停頓;全程僅兩處極短輕語。 鏡頭 CAMERA: 手持質感(極輕微自然晃動與呼吸感),多分鏡: 中近景/面部特寫/半身/回中近景,切換自然不炫技; 真實淺景深、焦點偶有極細微的呼吸式重對焦,背景宮闕牡丹虛化。 氛圍·光線: 黃昏真實自然光,側逆光鑲髮絲金邊,暖調; 空氣微塵、花瓣偶爾飄落、紗披與碎髮隨真實微風動, 整體像膠片實拍,顆粒細膩、高光柔和不過曝。 CONSTRAINTS(约束): - 4K 超清,15秒,人物妝容服飾與首幀一致(換鏡不換臉不換裝) - 核心:真人紀實表演——真實皮膚質感、不規律眨眼、自然重心轉移、 微表情有「不完美的真實感」,嚴禁僵硬呆滯或塑料假臉 - 去 AI 味:避免過度磨皮、過度對稱、過度順滑的運動與完美微笑 - 兩處台詞與口型嚴格同步,氣聲生活化,不假唱、不字正腔圓 - 風動、落花、紗袖、碎髮符合真實物理,自然不誇張 - 眼神有戲、有真實焦點游移;禁止眼睛發光、瞳孔異常 - 結尾在自然的小動作餘韻裡定格,無黑屏無轉場 - 無字幕、無水印、無任何標記符號、無文字
Generate this with Seedance 2 on VdoBloom.
Opens the image-to-video creator with this prompt and Seedance 2 pre-selected.