PROMPT RADAR / VIDEO视频叙事
Just a slow morning, good coffee, and a little exploring. Made with Seed
Just a slow morning, good coffee, and a little exploring. Made with Seedance 2.5. Prompt: Create a 15-second ultra-photorealistic travel vlog featuring the SAME young European woman throughout the entire video. She has n
@SyntheSarah 1.3k 热度 2.2 万 浏览
编辑先看
它为什么值得进入灵感库
这条内容值得留下的,不是表面的风格词,而是它提供了一条可以继续实验的画面路径。 把一段描述拆成起始状态、动作变化、镜头响应和结束画面,连续性比形容词更重要。
01 / 原始方法
动作与镜头 Prompt
Just a slow morning, good coffee, and a little exploring. Made with Seedance 2.5. Prompt: Create a 15-second ultra-photorealistic travel vlog featuring the SAME young European woman throughout the entire video. She has natural skin texture, subtle facial imperfections, realistic hair movement, authentic body proportions, and completely natural expressions. The video must look like genuine footage recorded by a real travel creator, NOT AI-generated or cinematic CGI. SHOT 1 — 0–3 seconds: Selfie-style handheld footage. The woman walks naturally down a beautiful European old-town street in the morning. She holds a small vlog camera at arm's length, looks directly into the lens and gives a subtle genuine smile. Natural walking bounce, slight handheld shake, realistic breathing and hair movement. Background includes pedestrians casually walking, parked bicycles, small cafés, textured historic buildings and soft morning sunlight. SHOT 2 — 3–6 seconds: She turns the camera away from herself toward the street. A natural handheld pan reveals detailed European architecture, outdoor café tables, people having coffee, bicycles passing and a small bakery. The camera movement is imperfect and human, with slight exposure adjustment as it moves from shadow into sunlight. SHOT 3 — 6–9 seconds: Cut naturally back to selfie footage. She continues walking and briefly says something casually to the camera with natural lip movement and relaxed facial expressions. A light breeze moves a few strands of her hair. People pass behind her naturally without looking at the camera. Realistic depth, authentic street ambience and subtle camera shake. SHOT 4 — 9–12 seconds: POV-style footage from her camera as she enters a small local café. The camera moves naturally through the doorway toward a wooden table with a freshly prepared coffee and a small pastry. Realistic reflections on glass, steam rising subtly from the coffee, natural indoor lighting, background conversations and realistic movement. SHOT 5 — 12–15 seconds: She sits outside the café and turns the camera back toward herself. She takes a small sip of coffee, looks into the lens and smiles naturally. The camera remains handheld and slightly imperfect. Background pedestrians, bicycles and café activity continue moving naturally. End with a subtle realistic focus shift from her face toward the street behind her. VISUAL REQUIREMENTS: Absolute photorealism, real-world physics, realistic human motion, authentic handheld smartphone/vlog camera behavior, natural skin pores and texture, realistic eyes and hair, physically accurate lighting and shadows, natural exposure changes, realistic reflections, detailed European environment, believable background pedestrians, consistent face and clothing across every shot. Avoid overly smooth camera movements, artificial beauty filters, plastic skin, exaggerated expressions, perfect symmetry, CGI appearance, dramatic movie-style transitions, unrealistic background motion, floating objects, warped hands, distorted faces or AI-looking details. The final result should feel like an actual 15-second travel vlog filmed spontaneously by a real European woman on a high-quality handheld camera.
02 / 复刻路径
不要一次改完所有变量
用 Seedance 复刻时,建议先单独生成稳定首帧,再把动作拆成“开始、变化、收束”三段;每段只保留一个主要运动。
- 主体:Just a slow morning, good coffee, and a little exploring. Made with Seedance 2.5. Prompt: Create a 15-second ultra-photorealistic travel vlog featuring the SAME young European woman throughout the entire video. She has n
- 视觉方向:实时 AI 视频案例;查看原帖媒体与作者说明
- 镜头与构图:按原帖镜头流程执行;查看原帖媒体与工作流
- 交付约束:auto,保留原作者方法与归属;复刻时避免水印、低清、畸变、主体漂移和无关文字。
负向约束
保留原作者方法与归属;复刻时避免水印、低清、畸变、主体漂移和无关文字。
模型建议
Seedance / Veo / Kling。先用原始 Prompt 建立基准结果,再根据模型对自然语言、镜头运动或文字排版的能力做局部改写。
来源与编辑原则