A wordless vertical video: a beauty in ancient Chinese attire with a gold floral crown in a waterside pavilion conveys emotion with her eyes alone

Prompt

Image-to-video · Seedance 2.0 · 4K · Vertical 9:16 · 15 seconds · Purely eye and facial expression acting · No dialogue · Exceptionally nuanced live-action performance  Character reference CHARACTER REF（=upload the character design image and lock the character’s appearance）： a beautiful woman in an ancient Chinese style, black hair arranged in a high bun with long curled locks falling down; a gold-and-gemstone floral crown with dangling pearl tassels; a long strapless gown with bare shoulders, a pink floral-embroidered bodice, and a sheer pale-pink outer layer; long tassel earrings. Throughout the entire video, her facial features, makeup, hair ornaments, and clothing must match the reference image 100%, with no distortion or face swapping.  Vertical composition: 9:16 portrait frame, character’s face centered slightly above the middle, with breathing room; framed for close-ups and macro shots of both eyes, with the top and bottom of the background blurred.  Scene SCENE（generate a new scene）： a quiet waterside pavilion in a secluded courtyard; semi-transparent gauze curtains behind her, blurred flower silhouettes in the distance; warm dusk light slants in, dust motes drift through the beams, and a light breeze stirs the curtains and strands of hair. Quiet, elegant, with ample negative space.  Performance core: purely eye and facial expression acting, with no dialogue, voice-over, or subtitles throughout — every emotion conveyed solely through the eyes, eyelashes, brows, nostrils, and the tiniest movements of the lips. Award-winning actress-level live-action performance: all the drama is on her face, all the subtext in her eyes. Realistic fine facial hair and the subtle rise and fall of breathing; irregular blinking; the eyes move first, then the head follows. Microexpressions with an “imperfectly real” quality, vivid and never stiff; absolutely no plastic AI look.  Storyboard（emotions build layer by layer, not a single line of dialogue, primarily vertical close-ups）：  SHOT 1｜0–2.5 seconds｜medium close-up of face（vertical composition） Her gaze is slightly lowered, fixed on the empty space before her; her eyes are still and distant. Her eyelashes flutter almost imperceptibly, her lips part slightly and then gently press together; her breathing is visible.  SHOT 2｜2.5–5.5 seconds｜macro close-up of both eyes Her gaze shifts from stillness to movement, pupils drifting subtly as if recalling a memory. A trace of tender warmth first rises in her eyes, and the corners of her mouth soften almost imperceptibly — the emotion begins only in her eyes, while her face barely moves.  SHOT 3｜5.5–8.5 seconds｜close-up of face A flicker of pain suddenly passes through the warmth; her gaze trembles slightly, her eyes grow moist with a fine sheen of tears. Her brows knit almost imperceptibly, then relax; her nostrils gently flare as she struggles to keep her feelings from spilling over.  SHOT 4｜8.5–11.5 seconds｜macro close-up of both eyes Tears well and turn in her eyes but never fall. She slowly closes her eyes for a brief moment. When she opens them again, she has already tucked her vulnerability away in their depths — the most emotionally restrained climax of the entire video.  SHOT 5｜11.5–15 seconds｜return to medium close-up of face（hold frame） She slowly lifts her eyes and looks offscreen; her gaze settles into softness and warmth, with a faint glimmer. The corners of her lips form the faintest smile, relieved yet tinged with wistfulness. Her lips move ever so slightly, then still at once（as if she wants to speak but holds back）. The video ends on a still frame of cool, gentle, slightly melancholy emotion — her eyes are still “speaking” as the image falls silent.  Camera CAMERA： portrait handheld with an almost imperceptible breathing quality; multiple shots, alternating between medium close-ups of the face and macro shots of both eyes; cuts follow the emotional rhythm; shallow depth of field, gauzy curtains and flower silhouettes blurred in the background; occasional subtle, breath-like refocusing, always locked on the eyes and face.  Atmosphere · lighting： warm dusk light, soft warm tones, dust motes in the light beams, the grain of live-action film, soft highlights with no overexposure; curtains, loose strands of hair, and pearl tassels move slightly in the faintest airflow, keeping a subtle sense of movement in the frame at all times and avoiding a lifeless stillness.  CONSTRAINTS（constraints）： - 4K ultra-high definition, vertical 9:16, 15 seconds; character matches the reference image（no face or outfit changes between shots） - No dialogue, voice-over, or subtitles throughout — purely eye and facial expression acting - Emotions build layer by layer（distant and lost in thought → tender recollection → pain and tearful eyes → restrained composure → relieved, as if wanting to speak but holding back） - Award-winning actress-level live-action performance: realistic skin, irregular blinking, eyes move before the head turns, tears well but do not fall - Remove the AI look: no excessive skin smoothing, excessive symmetry, excessive fluidity, plastic fake face, or stiff, vacant expression - No glowing eyes or abnormal pupils; keep the tear shine realistic and restrained - Cuts follow the emotional rhythm, with no dropped frames or flickering - End with a natural still frame as if she wants to speak but holds back; no black screen or transition - No subtitles, no watermark, no symbols of any kind, no text

参考图生视频 · Seedance 2.0 · 4K · 竖屏 9:16 · 15秒 · 纯眼神面部戏 · 零台词 · 极致真人表演  人物参考 CHARACTER REF（=上传设定图，锁定人物形象）： 古风美人，乌发高髻挽卷垂落长发，金镶宝石花冠+珍珠流苏垂坠， 裸肩粉色花卉刺绣抹胸长裙、外罩淡粉薄纱、长流苏耳坠。 全程五官、妆容、发饰、服饰与参考图 100% 一致，不得变形换脸。  竖屏构图：9:16 竖画幅，人物面部居中偏上，留出呼吸空间， 适合特写与双眼 macro，背景上下虚化。  场景 SCENE（新生成）： 幽静庭院水榭，身后半透纱幔、远处花影朦胧，黄昏暖光斜入， 光柱里浮尘漂动，微风拂动纱幔与发丝。安静、清雅、留白。  表演核心：纯眼神+面部表情戏，全程零台词、零旁白、零字幕—— 所有情绪只靠眼睛、睫毛、眉梢、鼻翼、唇部极微动作完成。 影后级真人实拍表演：戏全在脸上、潜台词在眼里。 真实皮肤细绒与呼吸起伏，眨眼节奏不规律，眼神先动、头才跟， 微表情带"不完美的真实感"，灵动不僵，绝无塑料 AI 味。  分镜脚本（情绪层层递进，无一句台词，竖屏特写为主）：  SHOT 1｜0-2.5秒｜面部中近景（竖构图） 微微垂眸、目光落于眼前虚空，眼神静而出神， 睫毛极轻一颤，唇瓣微启又轻抿，呼吸可见。  SHOT 2｜2.5-5.5秒｜双眼特写 macro 眼波由静转动，瞳孔细微游移像在回忆， 眼底先浮起一缕温柔暖意，嘴角几不可察地软下来—— 情绪只在眼里起，脸几乎不动。  SHOT 3｜5.5-8.5秒｜面部特写 暖意里忽然掠过一丝刺痛，眼神微微一颤、眼眶泛起薄水光， 眉梢极轻一蹙又松开，鼻翼轻翕，强忍着不让情绪溢出。  SHOT 4｜8.5-11.5秒｜双眼特写 macro 泪意在眼眶里打转却始终不落，她缓缓阖眼一瞬， 再睁眼时已把脆弱收进眼底——这是全片情绪最克制的爆点。  SHOT 5｜11.5-15秒｜回到面部中近景（定格） 缓缓抬眼望向画外，眼神归于柔和、温润、微亮， 唇角牵起一抹极淡、释然又略带怅然的浅笑， 唇瓣极轻一动旋即收住（欲语还休）， 最终定格在清冷、温柔、略带忧郁的情绪里——眼神仍在"说话"，画面已静。  镜头 CAMERA： 竖屏手持极轻微呼吸感，多分镜——面部中近景／双眼 macro 交替； 切换贴合情绪节拍，浅景深，背景纱幔花影虚化， 焦点偶有细微呼吸式重对焦，始终咬住眼睛与面部。  氛围·光线： 黄昏暖光，暖调柔和，光柱浮尘，胶片实拍颗粒感，高光柔不过曝； 纱幔、碎发、珍珠流苏随极轻气流微动，画面始终有微弱流动，避免死寂。  CONSTRAINTS（约束）： - 4K 超清，竖屏 9:16，15秒，人物与参考图一致（换镜不换脸不换装） - 全程零台词、零旁白、零字幕——纯眼神与面部表情戏 - 情绪层层递进（出神→温柔回忆→刺痛泛泪→克制收住→释然欲语还休） - 影后级真人实拍：真实皮肤、不规律眨眼、眼神先到头再转、泪在眶中不落 - 去 AI 味：禁过度磨皮、过度对称、过度顺滑、塑料假脸、僵硬呆滞 - 禁止眼睛发光、瞳孔异常；泪光真实克制 - 分镜切换贴合情绪节拍，不跳帧不闪烁 - 结尾欲语还休自然定格，无黑屏无转场 - 无字幕、无水印、无任何标记符号、无文字

• Needs your photo: attach an image to the request — the prompt works with it.