视频 · Seedance 2.5
MiniDV质感的韩国村落午后日常
生成一段 30 秒 4:3 家庭录像风格短片,记录一位成年女性在传统村落中与孩子玩球、购买杂货并回家的温暖午后,并保持人物、动作和声音的连贯性。
Prompt
创作一段30秒的超写实真人生活方式视频博客,跟随同一位25岁的韩国女性,度过宁静传统韩国村落里一个温暖的午后时光。她离开瓦顶住家,看到当地孩子们在玩球,兴致所至加入他们,随后想起还要买杂货,便去了一家小小的街坊商店,然后步行回家。
始终保持面部身份、发型、服装和写实比例一致。呈现真实的 MiniDV 家庭录像美学:4:3消费级影像、自然的手持晃动、不完美的取景、自动对焦来回搜寻、轻微曝光变化、柔和对比度、褪色感、细微的 DV 压缩痕迹,并且不带现代电影感。
所有动作都要符合物理规律——包括逼真的行走、奔跑、球的轨迹、头发和衣物的动态、门和门铃的运作、杂货袋的重量、风、重力、摩擦,以及自然的人体运动。仅使用村落现场音:孩子的笑声、脚步声、球弹跳声、鸟鸣、风声、钥匙声、木门声、商店门铃声、杂货声和远处的街坊环境音。不要音乐或旁白。
最后,她走进家门并轻轻关上门;温暖的阳光洒满宁静的村巷,孩子们的笑声渐渐消散在远处。
不要 CGI、动画、美颜滤镜、塑料质感皮肤、现代智能手机影像风格、电影级稳定器镜头、不可能的物理效果、重复人物、面孔或服装变化、人工音效、字幕、标志或水印。Create a 30-second ultra-photorealistic live-action lifestyle vlog following the same 25-year-old Korean woman through a warm late-afternoon in a quiet traditional Korean village. She leaves her tiled-roof home, discovers local children playing with a ball, spontaneously joins them, then remembers her groceries and visits a tiny neighborhood shop before walking home.
Maintain consistent facial identity, hairstyle, outfit, and realistic proportions throughout. Show authentic MiniDV home-video aesthetics: 4:3 consumer footage, natural handheld shake, imperfect framing, autofocus hunting, mild exposure changes, soft contrast, faded colors, subtle DV compression, and no modern cinematic look.
Keep all movement physically believable—realistic walking, running, ball trajectories, hair and clothing motion, door and bell mechanics, grocery-bag weight, wind, gravity, friction, and natural human biomechanics. Use only diegetic village sounds: children laughing, footsteps, ball bouncing, birds, wind, keys, wooden doors, shop bell, groceries, and distant neighborhood ambience. No music or narration.
End with her entering home and gently closing the door as warm sunlight fills the quiet village lane and the children's laughter fades into the distance.
No CGI, animation, beauty filters, plastic skin, modern smartphone aesthetic, cinematic gimbal shots, impossible physics, duplicated people, changing faces or outfits, artificial sound effects, subtitles, logos, or watermark.
使用建议
使用建议:在支持较长时长和角色一致性的视频模型中使用;如模型支持分镜或参考图,可为人物、房屋和杂货店分别提供参考,以提高跨镜头稳定性。







