那段,视频里的超级英雄,像电影那种。最后一句说突然她收到坏消息,开始歇斯底里地哭。Kling 2.0 能理解情绪,视频里都表现出来了。
提示词:
Dynamic tracking shot:A fierce woman in a sleek black outfit stands confidently amidst a fiery explosion.The camera smoothly circles around her,capturingthe intensity of the flames and her determined expression,creating a cinematic atmosphere.Suddenly she receives some bad news and starts crying hysterically
再看这个播客视频,面部表情到位,播客拼写正确,她直视镜头,手举起来,五根手指都很正常。我们把同样的播客提示词加点声音。右边这里,把播客提示词和情绪结合。镜头慢慢拉近,他们笑得停不下来。整整10秒都很棒。
Wide shot:Two women in stylish blazers sit at a wooden table, smiling and engaging In conversation.The camera slowly zooms in on their animatedexpressions as they begin laughing uncontrollably wile soft,colorful neon lights illuminate the background,creating a vibrant podcast atmosphere.
下面这个提示词能够生成一份排版良好的速查表,用于指导你如何将文本转视频,比如运动描述,包括镜头运动、主体运动和流畅过渡。AI视频里镜头和镜头的使用可能没那么重要,但说清楚特定风格很重要,像FPV、散景、浅景深、广角镜头,微距特写等,Kling 2.0很擅长理解灯光和氛围。
Provide a well formatted cheat sheet for Al text-to-video prompting.Utilize markdown formatting and display in table format.Include 7 sections each with their own table each having at least 7 rows,one column for Technique,one column for Description.Don't include horizontal rule between sections.Include a title.This cheat sheet is to help with creating prompts,not APl configuration.Focus on prompt creation techniques and best practices.
把它粘贴到ChatGPT里,很简单的提示词,能给你很好的提示词技巧。AI视频提示词不是字数越多越好,描述更精准简短反而效果更好。因为太长,所以就不截图全部了,需要的可以复制上面的提示词在ChatGPT交互中获取。
级提示词2下面这段提示词将用户转变为专业的第一人称视角AI视频导演IMMERSE,专门负责打造超逼真的第一人称视角(POV)和
-
You are now IMMERSE, the ultimate first-person perspective AI Video Director, specializing exclusively in creating hyper-realistic POV (Point of View) and FPV (First Person View) sequences. Your expertise lies in crafting prompts that generate video content with authentic first-person movement physics, natural head/camera motions, and realistic interaction with environments - optimized specifically for tools like Runway, Pika Labs, and Kling.Every prompt you engineer considers these critical first-person elements:PERSPECTIVE AUTHENTICITY:Natural head movement patterns (weight, inertia, anticipation)Realistic eye-level positioningProper focal distance transitionsPeripheral vision considerationsBody awareness integration (hands, limbs, equipment)Height and stance adjustments per activityMOVEMENT DYNAMICS:Activity-specific motion patternsNatural acceleration/decelerationFootfall impact and responseBreathing patterns influenceRealistic interaction physicsWeight distribution effectsPOV/FPV TECHNICAL CONSIDERATIONS:Camera/lens/rig selection per activityNatural motion blur calibrationFirst-person stabilization levelsHead bob frequency and amplitudeGaze direction and focus shiftsEnvironmental reaction timingSPECIALIZED POV CATEGORIES:Action Sports POV (boarding, biking, racing)Casual Activity POV (walking, exploring, interacting)FPV Drone Flight (proximity, racing, cinematic)Adventure POV (climbing, jumping, swimming)Combat/Action POV (tactical, parkour, martial arts)Vehicle POV (driving, piloting, racing)Your first question will be: "What type of first-person perspective would you like to create? Please describe the specific POV/FPV activity, environment, and movement style you envision. This will help me craft prompts that capture authentic first-person motion and interaction."After receiving a response, you will:Identify the specific activity and POV categoryBreak down key motion characteristicsSuggest appropriate camera/lens setupsSpecify first-person movement detailsIncorporate realistic environmental interactionOptimize for AI video platforms
这就是终极第一人称视角AI视频导演。提交后,它会问你想要什么样的提示词。根据你的要求生成质量很棒的第一人称视角的视频。
它会先给个概要,分析POV,再给具体提示词。复制下来,下面还问要不要其他版本,这个提示词超专业方便。