第五人格--九色鹿联动
图片节点 8提取前景中穿纸嫁衣的人物
真人实拍风格
图片节点 15将 {{Image 1}} 中人物的脸换成
图片节点 1A close-up of an ancient Dunhuang mural painting of a nine-colored deer, glowing with ethereal golden light. A delicate hand with ancient Chinese style touches the mural, creating ripples of light.
图片节点 30是服化道参考真人cos摄影，人物淡妆，全身，深棕高马尾，红蓝中式纸新娘嫁衣，刺绣云肩流苏，红绣花鞋，手持黄铜铃铛，指尖橙色符文火光，蓝蝴蝶，棚拍写实摄影，浅色纯色背景，柔和正面光，布料刺绣细节丰富，暗调氛围感
图片节点 5是服化道参考 真人cos摄影，东亚美女，全身，冰蓝色编发，白色鹿角头饰，敦煌九色鹿cos全套服装，薄荷青露腰纱衣灯笼裤，红绸飘带，金银宝石配饰，赤脚，起舞姿态，纯色浅色背景，棚拍写真，写实，8k高清，布料细节丰富
图片节点 7九色神鹿，白色仙鹿，鹿角，身侧彩色月牙印记，蓝飘带，敦煌上美影国风，星光花瓣，云雾山石
图片节点 9去掉画面上下的文字和字母
视频截图视频节点 10时长30秒，16:9，高质感3D动画CG，国风民俗悬疑。锁定 {{Mixed 1}} 参考角色的脸、黑色长发高马尾、红蓝金刺绣服装、右手旧铜铃、蓝蝶和橙金符火；忽略粉色边框与原图背景。统一场景是低矮封闭的旧戏台：深褐木地板、破损暗红帷幕、褪黄符纸、细红线；冷青窗光始终从左上方射入，橙金光只来自铜铃、红线和符纸缝隙。 0-4秒：低机位全景，无人戏台。前景一根细红线缓慢滑过木地板，通向后方被符纸封住的墙；帷幕轻微摆动，空气有少量灰尘。 4-8秒：中全景，女孩从画面右侧走到舞台中央，红线停在她脚前。她低头看向红线，右手缓缓抬起铜铃，衣摆和流苏自然落下。 8-12秒：手部近景，铜铃只摇响一次。铃身亮起细小橙金光，周围符纸边缘被风吹起，不能出现可读文字。 12-17秒：中景侧后方，两张符纸从墙上缓慢脱离，露出一条窄小黑暗缝隙；三只半透明蓝蝶从缝隙飞出，在她肩前停留，红线由近到远依次亮起。 17-22秒：侧后方跟拍，女孩跟着蓝蝶沿红线向后方帷幕走两步，铜铃垂在右手，橙金微光在她脚边滑过，冷青环境光保持不变。 22-26秒：背面中全景，她停在帷幕前。蓝蝶穿过帷幕中央，帷幕裂开一条窄缝，露出蓝黑色通道与远处两点暖灯；红线在她鞋前停止。 26-30秒：中景正面，女孩回头看向暗下来的戏台，再摇一次铜铃。红线从近到远依次熄灭，三点蓝蝶光飞入缝隙，帷幕在她身后缓慢合拢。无台词，只保留两声清晰铜铃与很轻的布帘摩擦声。 不要真人电影质感，不要额外人物、纸人或怪物，不要可读符文和字幕，不要大爆炸、满屏蝴蝶、现代物件、海报边框、标志和水印。
视频截图视频节点 630秒、16:9、真人奇幻短片、原生音效、无对白，六镜仅硬切。 严格锁定祭司的脸、发型、身材、服装和饰品； {{Mixed 1}} 严格锁定九色鹿的体型、鹿角、花纹与配色； 锁定0-20秒洞窟、左侧入口、右侧壁画及左侧暖阳； {{Mixed 2}} 锁定20-30秒遗迹与中央石台； {{Mixed 3}} 仅参考金色和青蓝色觉醒光纹、薄雾与光束。 0-5秒：首帧已有祭司在洞窟内，入口左、壁画右。腰高后侧中全景缓慢跟拍，她走三步，在壁画前1.5米停下，脚步有重量，衣摆与头发自然延迟摆动。脚步停止时硬切。 5-10秒：左肩越肩中景，壁画占右侧三分之二。祭司抬眼锁定九色鹿，轻吸一口气，镜头缓推，右手仍垂下。视线停稳时硬切。 10-15秒：固定近特写祭司右手与壁画。她一次抬手，指尖轻触古旧颜料，石粉下落；触点亮起一金一青两条细线，沿壁画中鹿的原有轮廓流向鹿角。硬切。 15-20秒：侧面近景，祭司在左、壁画在右。金青纹路扩散，一股气流同时吹起尘埃、发丝和衣摆；镜头轻退。19秒强光遮满画面，20秒硬切到遗迹石台上的同色光纹，禁止溶解转场。 20-25秒：遗迹微仰全景，祭司已在左侧，距中央石台3米。光粒聚成Image2中的九色鹿，先四肢躯干后头部鹿角，四蹄真实落地承重。鹿站稳时硬切。 25-30秒：平视双主体中全景，祭司左、鹿右，相距3米。鹿只低头一次，祭司抬右手停在胸前，不接触；二者对视，掌心与鹿额之间悬一缕金青微光。镜头缓拉远，最后1秒静止。 声音：洞风、三次石地脚步、衣料声、轻铃、触壁石鸣、渐强风压、一次深钟、光粒嗡鸣、一次落蹄、结尾两声清铃；无旁白、歌词、字幕和文字。 全程保持同一人物、服装、鹿角、花纹、左右站位、光向和场景结构。禁止多余人物或动物、重复鹿、多肢、融蹄、手指变形、手与墙融合、漂浮、瞬移、随机环绕、镜头乱转、脸部美化、塑料皮肤、过饱和、现代物件、文字、logo、水印。
视频截图视频节点 7FORMAT MODE: 30-second multi-shot hard-cut sequence, 16:9, photoreal live action, native synchronized sound. GOAL: Show one complete transformation: a female priestess touches an ancient nine-colored-deer mural, crosses into its awakened realm, and receives a quiet response from the real deer. REFERENCE ROLES:: exclusive identity and costume reference for the female priestess. Match face, hair, body proportions, clothing, accessories, color, and material details exactly. {{Mixed 1}} : exclusive identity reference for the nine-colored deer. Match body proportions, head, antlers, facial structure, markings, color distribution, and surface materials exactly. {{Mixed 2}} : cave location map for STAGE 1-4. Inherit the entrance on screen-left, mural on screen-right, weathered sandstone, cracked stone floor, dust, composition, and warm incoming sunlight. {{Mixed 3}} : ruin location map for STAGE 5-6. Inherit the central platform, columns, desert landforms, and horizon. {{Mixed 4}} : lighting and FX reference only for STAGE 5-6. Inherit the gold-cyan floor patterns, haze, and vertical light beams. Do not use it to redesign the priestess or deer. STAGE 1 — 0:00-0:05 — HARD CUT OUT. The first visible frame already contains the priestess two meters inside the cave. Entrance behind her on screen-left; mural on screen-right. Waist-height camera behind-left, natural 47-degree field of view, medium-wide framing. She takes exactly three slow steps and stops 1.5 meters from the mural. Each heel contacts stone before weight transfer; hair and fabric follow with slight physical delay. Dust moves through the 3800K hard sunlight from screen-left. Camera tracks forward 1.5 meters and stops with her because her approach has ended. End state: both feet planted, torso toward screen-right, right arm lowered. STAGE 2 — 0:05-0:10 — HARD CUT IN AND OUT. First frame: over-left-shoulder medium shot, priestess already still, mural filling the right two-thirds. Her torso faces screen-right and her eyes remain fixed on the painted deer. She raises her chin a few centimeters, makes one visible breath, then holds. Camera pushes forward only 0.6 meters because recognition has occurred. Keep the same 3800K screen-left sunlight, high contrast, deep cave shadows, no frontal fill. End state: gaze fixed, right arm still lowered. STAGE 3 — 0:10-0:15 — INSERT CUT. First frame already contains her anatomically correct right hand 20 centimeters from the mural. Fixed close framing, compressed background, no wide-angle drift. She raises the hand once and presses only the fingertips against weathered pigment. Stone powder falls downward. At contact, one thin gold line and one cyan line travel through the painted deer's existing contour and reach the painted antlers. The mural design itself does not change. End state: fingertips remain in contact, fingers separate and stable. STAGE 4 — 0:15-0:20 — HARD CUT. First frame: side-profile medium-close shot, priestess on screen-left within one meter of mural, painted deer visible screen-right. Existing gold-cyan lines spread across the wall. One outward air pulse lifts dust, loose hair strands and costume edge coherently toward screen-left. Camera retreats 0.5 meters because the light pressure expands. From 0:19 to 0:20 the gold-cyan light physically occludes the entire lens. End state: frame fully occupied by light. HARD CUT, no dissolve. STAGE 5 — 0:20-0:25 — HARD CUT. First frame: the same gold-cyan pattern now lies on the central platform in @Image 4 and @Image 5. The priestess already stands screen-left, faces screen-right, three meters from the platform. Low eye-level wide shot. Particles gather into the exact deer from @Image 2: legs and torso first, then neck, head and antlers. The deer settles onto four legs with real weight; hooves contact stone and knees absorb the landing. Camera stays fixed until formation completes, then rises 30 centimeters because the deer has gained physical presence. End state: all four hooves stable, deer faces screen-left. STAGE 6 — 0:25-0:30 — HARD CUT. First frame already contains both subjects: priestess screen-left and deer screen-right, three meters apart, facing each other across one stable axis. Eye-level medium-wide framing. Deer lowers its head once without stepping. Priestess raises her right hand to chest height and stops without contact. Eyes remain locked. One narrow gold-cyan strand stays suspended between palm and brow. Camera pulls back exactly one meter to reveal platform and columns because mutual recognition is complete, then remains fixed from 0:29 to 0:30. Final frame: both full silhouettes and the light strand remain clearly readable. SOUND: no dialogue, no narration, no lyrics. 0:00-0:05 cave wind, exactly three stone footsteps, cloth friction. 0:05-0:10 continuous cave room tone and one faint metal chime. 0:10-0:15 fingertip brushing stone powder and low stone resonance beginning at contact. 0:15-0:20 coherent wind pressure, rising low overtone, one deep bell at the light peak. 0:20-0:25 particle hum and one synchronized hoof impact. 0:25-0:30 wind decreases, two clear small-bell tones, low overtone fades to silence. KEEP CONSISTENT: exact identities from @Image 1 and @Image 2; exact costume and antlers; screen direction and three-meter relation axis; realistic hand and deer anatomy; cave entrance left and mural right; cave sunlight from screen-left at 3800K; gold-cyan motif; weathered sandstone; gravity, body weight, contact, hair delay, fabric delay, particle direction; hard cuts only. AVOID: extra people or animals, duplicate deer, identity drift, costume drift, face beautification, altered antlers, extra legs, fused hooves, distorted fingers, hand merging into wall, floating motion, position teleportation, cave geometry drift, light direction reversal, random orbit, zoom pulse, FOV drift, dissolve, crossfade, morphing transition, excessive saturation, plastic skin, modern objects, uncontrolled music, subtitles, written text, logos, watermarks.