首尔日常：超真实手持Vlog提示词

当真实感开始有人情味时，AI 生成的劣质内容就消亡了。 使用 Seedance 2.5 制作 制作一段 30 秒、16:9 超真实手持生活 Vlog，跟随参考图中同一位年轻韩国女性度过在首尔的平凡一天。始终保持她的面部、身份、发型、身材、服装、配饰和比例完全一致。 她全程穿着浅蓝色宽松衬衫，内搭白色合身 T 恤，深灰色直筒牛仔裤，白色运动鞋，棕色斜挎包，佩戴银色手表。 风格：由亲密朋友用智能手机/紧凑型相机随意拍摄的真实素材。手持抖动、不完美的构图、自动对焦/曝光变化、自然的运动模糊、轻微数字噪点、真实的皮肤质感、偶尔滞后的相机反应。没有稳定器、无人机、电影级运镜、慢动作、美颜滤镜或商业感。 0-5 秒：安静的首尔街区。她离开公寓，查看手机，注意到镜头，睡眼惺忪地微笑，整理头发并自然地说：“今天也一起走吧。” 5-11 秒：进入一家小型社区面包店，注意到新鲜糕点并说：“这个闻起来太香了。”买下它，在外面咬了一口，对着镜头微笑。 11-17 秒：登上当地公交车，坐在窗边。首尔街道从窗外掠过，带有真实的反射和阳光。她轻声说：“这种日子就想走走。” 17-23 秒：在河边小径附近下车。从小推车买了一杯水果。一辆自行车意外地靠近经过；她闪到一边，笑着说：“刚才有点吓到了。” 23-27 秒：金色时刻的河边。她坐在长椅上，把水果杯放在旁边，取下一只耳机，望着水面，风吹动她的头发时露出一个真诚的小微笑。 27-30 秒：她站起来走开。走了几步后，回头说：“下次你也一起来吧。”她继续走着，镜头在后面跟随。录制在约 29.5 秒自然结束。 真实感：准确的人体运动学、脚部着地、物体连续、真实的公交车运动、自然的行走/进食/坐姿、一致的建筑和服装。没有瞬移、变形、身份漂移、物体重复/消失、手或肢体扭曲。 音频：自然的韩语对话，口型同步准确，手机麦克风音质，脚步声、交通声、鸟鸣、面包店声音、公交车引擎、自行车、风声和城市环境声。没有音乐、旁白或电影级音效。 绝对没有字幕、说明文字、屏幕文字、标志、水印或用户界面。

AI slop dies when realism starts feeling human. Made with seedance 2.5 Create a 30-second, 16:9 ultra-realistic handheld lifestyle vlog following the same young Korean woman from the reference image through an ordinary day in Seoul. Keep her face, identity, hairstyle, body, outfit, accessories, and proportions perfectly consistent. She wears a light-blue oversized shirt over a white fitted T-shirt, dark-gray straight jeans, white sneakers, brown crossbody bag, and silver watch throughout. Style: Genuine footage filmed casually by a close friend on a smartphone/compact camera. Handheld shake, imperfect framing, autofocus/exposure shifts, natural motion blur, mild digital noise, realistic skin texture, occasional late camera reactions. No gimbal, drone, cinematic choreography, slow motion, beauty filter, or commercial look. 0–5s: Quiet Seoul neighborhood. She leaves her apartment, checks her phone, notices the camera, smiles sleepily, fixes her hair and says naturally: “오늘도 같이 가자.” 5–11s: Enters a tiny neighborhood bakery, notices a fresh pastry and says: “이거 냄새 너무 좋다.” Buys it, takes a bite outside, and smiles at the camera. 11–17s: Boards a local bus and sits by the window. Seoul streets pass outside with realistic reflections and sunlight. She quietly says: “이런 날은 그냥 걷고 싶어.” 17–23s: Gets off near a riverside path. Buys a fruit cup from a small cart. A bicycle passes unexpectedly close; she steps aside, laughs, and says: “방금 좀 놀랐어.” 23–27s: Golden-hour riverside. She sits on a bench, places the fruit cup beside her, removes one earbud, watches the water, and gives a small genuine smile as the wind moves her hair. 27–30s: She stands and walks away. After a few steps, looks back and says: “다음엔 너도 같이 와.” She continues walking as the camera follows behind. Recording naturally ends around 29.5s. REALISM: Accurate human biomechanics, grounded feet, continuous objects, realistic bus movement, natural walking/eating/sitting, consistent architecture and wardrobe. No teleportation, morphing, identity drift, duplicated/disappearing objects, distorted hands or limbs. AUDIO: Natural Korean dialogue with accurate lip-sync, phone-mic quality, footsteps, traffic, birds, bakery sounds, bus engine, bicycle, wind and city ambience. No music, narrator, or cinematic SFX. Absolutely no subtitles, captions, on-screen text, logos, watermarks, or UI.