每种取景都看得见，提示词可直接复制
看清每一种景别到底在什么位置切开画面，从消失在风景里的一个人影，到占满整块屏幕的一双眼睛，然后复制那段能在 Veo、Sora、Runway、Kling 或任何图像模型里做出这个取景的提示词。
取景是电影语言的第一个词：同一个瞬间，在远景里是地理，在中景里是对话，在近景里是告白。下面每一张卡片都用同一个对象呈现经典景别序列中的一种，让差别被看见，而不是被描述。
取景演示图在你的浏览器里绘制。复制提示词不会向任何服务器发送内容。
The subject is a small element in a vast environment; the location is the real subject of the frame.
什么时候用： Openings, establishing geography, isolation, and scale - the world before the person.
extreme wide shot, the subject tiny in a vast landscape
A traveler crosses a salt flat at dawn, extreme wide shot, the subject tiny in a vast landscape, pastel sky.
The full subject is visible with generous room around them; body and environment share the frame.
什么时候用： Establishing a character in a place, blocking-heavy scenes, physical comedy.
wide shot showing the full subject and their surroundings
A mechanic stands in a cluttered garage doorway, wide shot showing the full subject and their surroundings.
The frame cuts just above the head and just below the feet: the body fills the height of the frame.
什么时候用： Costumes, posture, and movement - showing exactly how a character carries themselves.
full shot framing the subject from head to toe
A dancer in rehearsal clothes holds a final pose, full shot framing the subject from head to toe, studio light.
The frame cuts at mid-thigh - historically to keep a gunslinger's holster visible.
什么时候用： Confrontations and swagger; a power framing that keeps hands and hips in play.
cowboy shot framing the subject from mid-thigh up
A sheriff squares up in the saloon doorway, cowboy shot framing the subject from mid-thigh up, backlit dust.
The frame cuts at the waist: close enough for expression, wide enough for gesture.
什么时候用： The workhorse of dialogue - conversations, interviews, most scene coverage.
medium shot framing the subject from the waist up
A journalist leans into the interview question, medium shot framing the subject from the waist up, office bokeh.
The frame cuts at the chest: the face leads, but shoulders keep body language alive.
什么时候用： Emotional dialogue beats, reactions that still need a hint of posture.
medium close-up framing the subject from the chest up
A witness hesitates before answering, medium close-up framing the subject from the chest up, courtroom light.
The face fills the frame, cut around the shoulders; every micro-expression reads.
什么时候用： Key emotional moments, decisions, the audience inside a character's head.
close-up on the subject's face
Tears she refuses to release, close-up on the subject's face, candlelight flickering.
A fragment of the face - usually the eyes - fills the entire frame.
什么时候用： Maximum intensity: realization, terror, obsession, the western standoff stare.
extreme close-up on the subject's eyes
The moment the plan clicks into place, extreme close-up on the subject's eyes, reflected neon.
A tight detail of an object or action - hands, a letter, a key turning - not a face.
什么时候用： Story-critical props, cutaways that carry plot information, tactile texture.
insert shot, extreme close-up detail of the object
A brass key slides into an old lock, insert shot, extreme close-up detail of the object, shallow focus.
找到和你这个瞬间所需信息量相匹配的取景，复制它的提示词片段，然后放在 AI 提示词的最前面，在主体和动作之前，因为模型给靠前的词更高权重。每个片段都用了行业标准术语加上一句大白话锚定（「从腰部往上取景」），即使模型对术语的理解偏松，也能把它拉回来。每张卡片上的示例展示了这个片段放进完整提示词后的样子。
景别序列是电影行业最古老的共同词汇：大远景、远景、全景、牛仔景、中景、中近景、近景、大特写，再加上用于物件的插入镜头。全世界的剧组都用这些词来画分镜、列镜头表、指导拍摄，这也正是那些在有描述的素材上训练出来的 AI 模型能听懂它们的原因。
相邻两种景别之间真正的差别，在于什么还留在画面里。中景保住手势；中近景用手势换来表情；近景用体态换来微表情。要的景别比这个瞬间所需的宽一档或窄一档，正是 AI 生成的对白镜头常常显得太远或太逼人的原因。
当景别有意识地交替时，一段序列才读得出专业感：用远景建立空间，用中景演这场戏，用近景落下转折，当道具推动情节时切到插入镜头。带着这种纪律去生成 AI 镜头，剪出来的片段会像真正的现场覆盖，而不是一堆互不相连的画面。
It visualizes the complete framing scale, extreme wide to extreme close-up plus the insert shot, on the same subject, each with the copy-paste prompt fragment for AI video and image generation.
For this workflow, the central problem is clear: shot sizes are learned from written definitions, so creators prompt 'medium shot' expecting a close-up and storyboard frames that do not match what the terms mean on set. Left unresolved, this creates downstream friction and slower decisions. The practical target is you see exactly where every framing cuts the body, and copy the prompt phrasing that produces that exact framing in AI models.
Limitation to keep in mind: Framings are conventions, not laws: a 'medium shot' cut point varies slightly between productions, and AI models interpret the terms with the same small variance.
Advanced workflow: Directors plan coverage by alternating sizes deliberately: wide to establish, medium for dialogue, close-up for the turn. Prompting AI shots with the same discipline produces sequences that cut together like real coverage.
Treat this tool output as a decision support layer, not a replacement for authorship. Great scripts are remembered for specific choices, emotional precision, and clarity of dramatic movement. Tools help by removing noise so your energy can go where it matters: character, conflict, escalation, and payoff. If you review outcomes after each pass and keep an explicit log of accepted changes, your workflow becomes faster and more predictable from draft to draft. That consistency is exactly what professional collaborators value: fewer surprises, clearer rationale, and a script that evolves with intent.
The classic scale runs extreme wide shot, wide shot, full shot, cowboy (medium wide), medium shot, medium close-up, close-up, and extreme close-up, plus the insert shot for object details. Each one is visualized on this page with its AI prompt.
A medium shot cuts at the waist and keeps gesture and body language in frame; a close-up fills the frame with the face and cuts around the shoulders. Between them sits the medium close-up, cut at the chest.
A framing that cuts at mid-thigh, named for westerns that needed the gun holster visible. It reads as a power framing: taller than a medium shot, with hands and hips still in play.
Yes, all major models (Veo, Sora, Runway, Kling, Seedance) understand standard framing vocabulary. Lead the prompt with the size term for the most reliable results, since models weight early words more heavily.
A tight detail shot of an object or action rather than a face: a key turning, a letter, hands on a keyboard. Inserts carry plot information and give the edit room to breathe.
Ask how much context the moment needs. Environment and geography want wides, dialogue wants mediums, emotion wants close-ups, and story-critical objects want inserts. Sequences work best when sizes alternate deliberately, like real coverage.
大远景、远景、全景、牛仔景、中景、中近景、近景和大特写，再加上用于物件细节的插入镜头。上面这九种都在同一个对象上做了呈现。
中景切在腰部，手和手势都还在画面里。中近景切在胸口：脸是主角，只留下一点肩膀，让肢体语言还活着。
管用。Veo、Sora、Runway、Kling 和 Seedance 都能响应标准的取景词汇。把景别词放在提示词开头，并带上那句大白话锚定，就像这里可复制的片段所做的那样。
对峙和气场。它切在大腿中段，既让手和胯留在画面里，又比中景显得更挺拔，所以西部片发明了它，惊悚片留下了它。
让取景匹配这个瞬间所需的信息量：地理要远景，对话要中景，情绪要近景，推动情节的物件要插入镜头。然后像真正的现场覆盖那样，在整段序列里交替使用不同景别。
