skills/ai-film-pipeline-master/references/phase-03-shooting-script/shot-sizes.md

# Shot Sizes

21 cards. Also in the other stage: a shot has a frame too. These decide a frame: what is in it, how it is lit, and what it is made of.

Every card carries what it is, what it does to the audience, where it works best, what to avoid, and the prompt written out in full.

---

### Name the Distance Before the Size

**Also called:** the shot-size selector, distance before size, how much of them does this beat need  
**What it is:** The card at the head of the `size.` shelf. It does not name a rung. It asks the three questions the shelf's own pair graph asks over and over — **is the thing in the frame a person, a thing, a place or a step between two framings; is this beat asking for pressure on a face or for room to read the situation; and is this a framing you shoot or a plan across a whole scene** — and hands back a named shortlist for each answer. The axes were not invented: they were read off the twelve `Yields to:` conditions the cards carry, and the condition that repeated in seven of them became the second question. Every card cited here is cited by id.  
**Effect on the audience:** None. This card never reaches the screen. It exists so the size that does reach the screen was chosen against the beat, rather than stepped down the ladder because the previous shot was wider.  
**Used for and where it works best:** Answer the three questions **in order and out loud, in one sentence each, before you open the shelf**. The first answer eliminates about two-thirds of the file and the second usually leaves two or three cards. **This shelf answers a different question from the named-framing shelf and does not replace it** — a size says how much of the subject the picture holds, a named framing says what is in the picture and how it is arranged, and you spend both on the same shot. The boundary is stated by the inventory itself: not one card on either shelf offers a card on the other as an alternative, in either direction, anywhere in the package.  
**Best in:** formats: all | genres: all  
**Avoid when:** **The decision in front of you is coverage, not a size.** There this map is dangerous rather than merely thin: `size.master` and `size.wide_vs_tight_master` are alternatives to a triangle of setups and to a shot-reverse pair, and those two live on the named-framing shelf with no pair field joining them to these. Ask the coverage question against both shelves at once or you will choose from half the candidates believing you saw them all - `frame.coverage_plan`, on the named-framing shelf, is the card that asks it. Also **distrust this map whenever the scene is already running** — every edge on this shelf points from a wider rung to a tighter one and not one points back out, so the map is a good guide to going in and says nothing about when to come out. And **it will not settle a tie it did not create**: two cards here both call themselves the top of the ladder in westerns, and two more disagree about how far you may jump between rungs. The map routes all four and settles none of them.  
**Prompt:** `[Selector — produces no prompt. It returns card ids. The chosen card carries the prompt.]`

---

**The axes**

**Axis 1 — Is the thing in the frame a person, a thing that is not a person, the place itself, or a step between two framings?**
Asked first because it is the only question that separates the three regions of the shelf before any of the others can bite, and because getting it wrong makes every later answer sort the wrong pile. The shelf states it as a boundary in `size.macro` (*"Ordinary detail, short of screen-filling magnification"*), and its cards mark the line in their own `Avoid when:` lines rather than their praise — `size.ews` gives the frame to the location and says so, `size.insert` is defined as *a close shot of something that is not a face*, `size.one_size_tighter` has no subject of its own at all and describes only the step between one setup and the next.

**Axis 2 — Is this beat asking for pressure on a face, or for room to read the situation?**
The condition the shelf repeats most: seven of its twelve `Yields to:` lines turn on it, and every one is stated as a failure of the wider card rather than a promise of the tighter one — *"When the viewer must bond with a face"* (`size.ews`), *"An internal emotional beat the wide cools down"* (`size.ws`), *"Quiet dialogue or an interview, too far from the face"* (`size.fs`), *"When the beat needs emotional pressure, not comfort"* (`size.mfs`), *"When the scene needs an emotional peak"* (`size.ms`), *"Shock: jump to the close-up, do not spend the step"* (`size.mcu`), *"Outside a confrontation, when the look must read as sympathy"* (`size.italian`). Two more state the same question from the other end — *"As a dialogue master, the faces disappear"* (`size.vws`) and *"Long dialogue scene carried by relationship, not geography"* (`size.master`).

**Axis 3 — Is this a framing you can shoot right now, a position in the scene's order, or a plan across the whole scene?**
Three conditions turn on it and they are the only ones on the shelf that are not about distance at all: *"When the location itself is telling the story"* (`size.wide_vs_tight_master`), its exact reverse in `size.master`, and `size.vws` handing a dialogue master upward rather than tighter. It is asked last because it is the question that catches the cards that are not sizes — and it is why this shelf does not need a second selector: the plan cards and the rung cards cite each other, which is what makes them one shelf rather than two.

---

**The map**

**Axis 1 — what is in the frame**

| Answer | The shortlist, by id |
| :--- | :--- |
| **A person, and the question is how much of them** | `size.vws` · `size.ws` · `size.fs` · `size.mfs` · `size.cowboy` · `size.ms` · `size.mcu` · `size.cu` · `size.bcu` · `size.ecu` · `size.italian` |
| **A thing that is not a person** | `size.insert` · `size.macro` |
| **The place, with or without people standing in it** | `size.ews` · `size.establishing` · `size.reestablishing` · `size.master` · `size.wide_vs_tight_master` |
| **Nothing in the frame — a step between two framings** | `size.one_size_tighter` |

**Axis 2 — pressure or room**

| Answer | The shortlist, by id |
| :--- | :--- |
| **Room — the geography has to read and the feeling may cool** | `size.ews` · `size.vws` · `size.ws` · `size.establishing` · `size.master` · `size.reestablishing` · `size.wide_vs_tight_master` |
| **Neither — the natural distance, neither pressed nor cooled** | `size.fs` · `size.mfs` · `size.cowboy` · `size.ms` |
| **Pressure — the face, or less than the face** | `size.mcu` · `size.cu` · `size.bcu` · `size.ecu` · `size.italian` |
| **Neither, because the attention is on a thing** | `size.insert` · `size.macro` |
| **Whatever the last setup was, minus a rung** | `size.one_size_tighter` |

**Axis 3 — a framing, a position, or a plan**

| Answer | The shortlist, by id |
| :--- | :--- |
| **A framing you can shoot right now** | `size.ews` · `size.vws` · `size.ws` · `size.fs` · `size.mfs` · `size.cowboy` · `size.ms` · `size.mcu` · `size.cu` · `size.bcu` · `size.ecu` · `size.italian` · `size.insert` · `size.macro` |
| **A position in the scene's order, defined by where it sits rather than how wide it is** | `size.establishing` · `size.reestablishing` |
| **A plan across the whole scene** | `size.master` · `size.wide_vs_tight_master` · `size.one_size_tighter` |

*A card appears once per axis. The answer is the intersection of the three shortlists, and on this shelf it is normally one to three cards wide.*

**Two live collisions this map routes and does not settle.**

**The top of the ladder in westerns: `size.ecu` and `size.italian` both claim it.** `size.ecu` says it is *the top rung of the size ladder* in westerns and thrillers; `size.italian` says it is *the top of the size ladder* in westerns, samurai films and martial arts. Both land in the same cell: a person, pressure, a framing you shoot. Nothing in the package ranks them and no edge runs between them. What the cards themselves offer as a discriminator, and it is thinner than a rule: `size.ecu` frames **any one part** — an eye, a mouth, a hand, a trigger finger — while `size.italian` frames **one specific band across both eyes, inside a confrontation**, and its own `Yields to:` line says it does not apply outside one. So in a western standoff the two genuinely overlap and the choice is the reader's; outside a standoff only one of them is even eligible. **Record which you chose.**

**How far you may jump: `size.ladder` and `size.one_size_tighter` disagree.** `size.ladder` says that when cutting on the same axis you skip **at least two rungs**, so the cut does not read as a jump. `size.one_size_tighter` says the coverage drops **one step** on the turn, shot and reverse together. Both are about movement on the ladder and they give different numbers. The interesting part is structural: `size.ladder` is filed `additive` in `PAIR-AUDIT.md` and is therefore excluded from this map altogether, so the card that states the shelf's movement rule can never appear as an alternative to the card that contradicts it. **The two cases may not be the same case** — one is a cut on the same axis within a setup, the other is a matched pair stepping down across a turn — but neither card says so, and until one of them does the disagreement is live.

---

**The default, and the one against it**

**The default is `size.mcu`, and the measurement that picks it is the one that rules out the obvious answer.** Six cards send a reader to `size.cu`, far more than to anything else here, and it is **not** the default: all six edges fire on one condition, the beat has turned and the face is needed — `size.ews` (*"bond with a face"*), `size.ws` (*"an internal emotional beat"*), `size.mfs` (*"emotional pressure, not comfort"*), `size.ms` (*"an emotional peak"*), `size.mcu` (*"shock"*) and `size.italian` (*"the look must read as sympathy"*). A card reached six times for one reason is a door out of the shelf, not a fallback: `size.cu` is where the ladder goes when the scene stops being ordinary, and the card says so itself — *it is hard currency: the less you spend it, the more it is worth*. A default you are told to ration is not a default.

`size.mcu` is. It is the only rung with traffic arriving from below and leaving upward — `size.fs` hands it a scene that has got too far from the face, and it hands `size.cu` a scene that has turned — so it sits on the hinge of the ladder rather than at either end. It is one of only six cards on the shelf carrying no format or genre restriction at all. And it is the only card here whose own prose claims the ordinary case rather than a peak, an entrance or an event, which is what a default has to be.

**The one against it is `size.reestablishing`.** Nothing routes to it, it routes to nothing, it carries no pair field, and it is the only card on the shelf whose content is a **return** rather than an approach. Every other card answers *how close*; this one answers *go back out*, and it is the sole card that argues against the shelf's own one-way traffic. It is also the only one that treats a wide as a thing the audience needs rather than a thing they are being denied — *it lets the audience breathe*, *it prevents confusion before it starts*. Its stated cost is precision, not expense: return on a transition beat rather than on a climax, or the wide breaks the tension it was meant to release. Its near cousin against the grain is `size.vws`, which refuses to choose between the person and the place and is likewise the target of nothing.

---

**The unasked**

**Thirteen of the nineteen reachable cards are the target of no `Yields to:` field anywhere on this shelf.** They can be chosen but nothing ever sends you to them. By id:

`size.ews` · `size.vws` · `size.ws` · `size.fs` · `size.mfs` · `size.cowboy` · `size.bcu` · `size.italian` · `size.insert` · `size.macro` · `size.establishing` · `size.reestablishing` · `size.one_size_tighter`

**And the shape of that list is the finding, not its length. Every rung wider than the face is on it.** `size.ews`, `size.vws`, `size.ws`, `size.fs`, `size.mfs` and `size.cowboy` — the entire wide half of the ladder — is the target of nothing at all. Every one of them sends a reader inward and not one of them is ever sent to. The graph runs strictly downward: wide yields to tight, tight yields to tighter, and the only pair on the shelf that points in both directions, `size.master` and `size.wide_vs_tight_master`, is not a pair of rungs. **This shelf can tell you when to get closer and cannot tell you when to back off.**

Which is why the two cards whose whole job is backing off are the most invisible things on it. `size.reestablishing` — *it lets the audience breathe*, *the moment you feel lost in the positions while watching, that is where it belongs* — carries no pair field, is filed `execution-note`, and is the target of no edge. `size.establishing` is in the same position. The readers who wrote these fields were each holding one shot and asking what to use **instead**; nobody was holding a scene and asking what to use **next**. A size you cut back out to is not an alternative to the size you cut in from — it is a different kind of answer, and the pair field cannot express it.

The second thing the list shows: **the two cards that describe motion on the ladder are both unreachable from any rung.** `size.one_size_tighter` is the target of nothing and `size.ladder` is excluded from the map entirely. So the shelf holds nine rungs that each know one place to go, and two cards that know how the whole staircase works, and no edge connects the two kinds. That is the mechanism behind the second collision above, and it will not close by adding an edge — **the card this shelf wants is the one that asks what the scene has to climb back to**, and the nearest thing that exists, `size.reestablishing`, states that question in its own prose without anything routing a reader there to read it.

`size.distance_before_size`

---

### Extreme Wide Shot / Extreme Long Shot

**Also called:** EWS, ELS, XLS  
**What it is:** The location is the protagonist. A person is a small dot in the frame or absent altogether, and the horizon, the architecture or the landscape dictates the entire image.  
**Effect on the audience:** It shrinks the human being and tells the viewer that the world is bigger than they are. This is a shot of judgement from a distance, not a shot of sympathy.  
**Used for and where it works best:** Opening a film or a chapter, the moment of defeat, the epic set piece, the final image after a death. In heritage documentary, the ziggurat, the city wall or the desert with one small figure so the viewer feels the scale of the civilisation. Strongest in features, heritage documentary and trailers.  
**Best in:** formats: Feature Film, Documentary Feature, Heritage / Historical Documentary, Trailer / Teaser, Travel / Real Estate / Hospitality, 3D CG Animation, Anime (Japanese 2D), AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Fantasy / Epic, Mythology / Ancient Epic, Heritage (civilisation-focused), Historical / Biopic, War, Western, Adventure, Dystopia / Post-Apocalyptic, Survival, Tragedy  
**Avoid when:** When you want the viewer to bond with a face. In vertical 9:16 you lose the horizon completely, so only reach for it when the subject itself is scale.  
**Prompt:** `extreme wide establishing shot, a single tiny figure lost in a vast landscape, horizon dominating the frame, static camera`  
**Example:** Lawrence of Arabia, the desert wides where riders are specks against the sand.  
**Yields to:** `size.cu` — When the viewer must bond with a face.  
`size.ews`

### Very Wide Shot / Very Long Shot

**Also called:** VWS, VLS  
**What it is:** One step in from the extreme wide: the figure reads clearly as a full body but stays small, and the location still holds most of the frame.  
**Effect on the audience:** A balance tipped toward place. The viewer knows who is there and where they are, but the environment carries more weight than the person.  
**Used for and where it works best:** A character walking a road, a market or a ruin when you need to sell the place and keep the person legible at the same time. In westerns, road movies and heritage documentary, the figure moving between old walls. In action, the full body plus the distance they still have to cover.  
**Best in:** formats: Feature Film, Drama Series Episode, Documentary Feature, Heritage / Historical Documentary, Travel / Real Estate / Hospitality, Trailer / Teaser, 3D CG Animation, AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Western, Road Movie, Adventure, Heritage (civilisation-focused), Historical / Biopic, Survival, Drama, War  
**Avoid when:** When the face is the subject. Do not lean on it as a dialogue master either, because the faces disappear.  
**Prompt:** `very wide shot, full figure small in frame walking through the ruins, environment dominating, camera static`  
**Yields to:** `size.wide_vs_tight_master` — As a dialogue master, the faces disappear.  
`size.vws`

### Wide Shot / Long Shot

**Also called:** WS, LS  
**What it is:** The full body head to toe, with enough air around it that you can read where the character is and what surrounds them.  
**Effect on the audience:** A balance between character and place. It is neutral: it hands you geography and imposes no feeling on you.  
**Used for and where it works best:** The first shot of any new scene (the master), full body action, dance, physical fighting, physical comedy that needs the whole body. In dance film, action and sitcom it is the shot that preserves the complete performance.  
**Best in:** formats: all | genres: all  
**Avoid when:** When the beat is internal and emotional, the wide cools the feeling down. Do not open every scene with it or it turns into a template.  
**Prompt:** `wide shot, full body visible head to toe, character grounded in the environment with room around them, eye level, static camera`  
**Yields to:** `size.cu` — An internal emotional beat the wide cools down.  
`size.ws`

### Full Shot

**Also called:** FS, full body shot  
**What it is:** The body exactly fills the frame height, with a little air above the head and below the feet.  
**Effect on the audience:** It introduces the character as a body: the clothes, the stance, the walk, the confidence or the defeat.  
**Used for and where it works best:** A character's first appearance, the fashion shot, period costume in heritage documentary, the commercial selling clothing or footwear. In anime it is the standing pose where a character declares themselves. It works in Reels because a standing body fills 9:16 naturally.  
**Best in:** formats: Feature Film, Drama Series Episode, Fashion Film, Commercial / TVC, Heritage / Historical Documentary, Vertical Short-Form (Reels, Shorts, TikTok), Anime (Japanese 2D), 3D CG Animation, Game Cinematic / Gameplay, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography | genres: Heritage (civilisation-focused), Historical / Biopic, Fantasy / Epic, Superhero, Shonen Battle (anime), Action, Drama, Musical  
**Avoid when:** In a quiet dialogue scene it keeps you too far from the face. In an interview it makes the subject look small and isolated.  
**Prompt:** `full shot, head to toe, character fills the frame height, standing still, minimal headroom, static camera`  
**Yields to:** `size.mcu` — Quiet dialogue or an interview, too far from the face.  
`size.fs`

### Medium Full Shot / Medium Long Shot

**Also called:** MFS, MLS, knee shot, three-quarter shot, medium wide shot  
**What it is:** Knees to head. Close enough to read the face, wide enough to keep the hands and the whole torso.  
**Effect on the audience:** The golden middle: expression and body language at the same time. Natural, with no pressure on the viewer.  
**Used for and where it works best:** Walking and talking, any scene with hand business or an object being carried, a small group of two or three. In sitcom and series it is the shot most scenes live on. In documentary, the craftsman working with his hands.  
**Best in:** formats: Drama Series Episode, Sitcom / Multi-Cam Comedy, Feature Film, Documentary Feature, Documentary Series, Vlog / YouTube Long-Form, Corporate / Training, Anime (Japanese 2D), Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Drama, Comedy, Rom-Com, Slice of Life, Family, Crime / Procedural, Verite / Observational  
**Avoid when:** When you need emotional pressure, this is a comfortable shot, not a tight one. Repeat it across every setup and the scene goes flat.  
**Prompt:** `medium full shot, framed from the knees up, hands and torso visible, natural eye level, static camera`  
**Yields to:** `size.cu` — When the beat needs emotional pressure, not comfort.  
`size.mfs`

### Cowboy Shot / American Shot

**Also called:** plan américain, hollywood shot, mid-thigh shot  
**What it is:** Framed from mid-thigh up. The name comes from westerns because it keeps the gun and the belt in frame, and the French called it the American shot.  
**Effect on the audience:** Readiness and contained threat. This person is about to do something. The hips and hands are present, so action is possible at any second.  
**Used for and where it works best:** Westerns and standoffs, the villain's entrance, the hero before the battle, the sports commercial. In samurai and martial arts, the hand resting on the hilt. In battle anime, the stance before the charge.  
**Best in:** formats: Feature Film, Drama Series Episode, Trailer / Teaser, Commercial / TVC, Anime (Japanese 2D), Game Cinematic / Gameplay, 3D CG Animation, AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Western, Action, Samurai / Chanbara, Martial Arts / Wuxia, Gangster, Superhero, Shonen Battle (anime), Thriller / Suspense  
**Avoid when:** In a scene of weakness or collapse, because this is a shot of power. Do not use it for a documentary interview, it makes the guest look ready to fight.  
**Prompt:** `cowboy shot, framed from mid-thigh up, hands resting near the hips, holster and belt visible, low eye level, static camera`  
**Example:** Sergio Leone westerns, the standoffs where the hand hovers over the holster.  
**Yields to:** `size.ms` — A documentary interview, where cowboy reads combative; and `size.bcu` — a scene of weakness or collapse.  
`size.cowboy`

### Medium Shot

**Also called:** MS, mid shot, waist shot  
**What it is:** Waist to head. The face is clear and the hands are usually in frame.  
**Effect on the audience:** The natural distance we talk to people at. Completely neutral, which is why it is the default talking shot.  
**Used for and where it works best:** Dialogue, the documentary interview, news, explainers, any scene where you want the viewer to listen rather than feel. In podcasts and educational video it is the workhorse because gestures stay readable without crowding the face.  
**Best in:** formats: all | genres: all  
**Avoid when:** Three or four in a row and it starts to look like cheap television. When you want an emotional peak, this shot will not get you there.  
**Prompt:** `medium shot, framed from the waist up, subject centered, eye level, natural light, static camera`  
**Yields to:** `size.cu` — When the scene needs an emotional peak.  
`size.ms`

### Medium Close-Up

**Also called:** MCU, head and shoulders, bust shot  
**What it is:** From the upper chest to the head. The single most used framing in all of film and television.  
**Effect on the audience:** The start of the move inward. The face is now the subject, but the body is still present so nothing feels trapped.  
**Used for and where it works best:** Sixty percent of any dialogue scene. The reaction shot, the serious interview, documentary testimony, the presenter addressing camera. In vertical 9:16 it is the new wide: the face fills the width and the shoulders anchor the frame.  
**Best in:** formats: all | genres: all  
**Avoid when:** When the scene needs body or place. And when you want shock, jump to the close-up instead so you do not spend the step for nothing.  
**Prompt:** `medium close-up, chest up, face is the subject, shoulders framing the bottom edge, shallow depth of field, eye level`  
**Yields to:** `size.cu` — Shock: jump to the close-up, do not spend the step.  
`size.mcu`

### Close-Up

**Also called:** CU  
**What it is:** The face fills the frame, from the top of the head (or slightly cropped) down to the chin or upper chest.  
**Effect on the audience:** Forced empathy. You cannot escape the face. It is also hard currency: the less you spend it, the more it is worth.  
**Used for and where it works best:** The moment of decision, the lie, the realisation, the confession, the tear, the truth. In drama and tragedy it is the peak the whole scene is built toward. In Reels it works as a first second hook because a close face stops the thumb.  
**Best in:** formats: all | genres: all  
**Avoid when:** Before the moment has arrived. Open a scene on it and you have nowhere left to go.  
**Prompt:** `close-up, face fills the frame, eyes clearly readable, soft background, minimal head movement, static camera`  
**Example:** The Passion of Joan of Arc (1928), a whole film built on close-ups of Falconetti's face.  
`size.cu`

### Big Close-Up / Choker

**Also called:** BCU, choker shot, tight close-up  
**What it is:** Forehead to chin. The top of the head is cropped off, no shoulders and no air.  
**Effect on the audience:** A siege. The viewer is now closer than comfortable and can feel the character breathing.  
**Used for and where it works best:** The interrogation, the breakdown, the beat before the emotional explosion, the face that threatens quietly. In psychological horror and thriller it lands when the character speaks softly and you want the viewer short of air. In a trailer, the villain's face for one second.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Trailer / Teaser, Short Film, Anime (Japanese 2D), AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography, Wedding Photography | genres: Thriller / Suspense, Psychological Horror, Crime / Procedural, Drama, Tragedy, Noir / Neo-Noir, Horror, Mystery / Whodunit  
**Avoid when:** More than once or twice in a scene, because the siege becomes routine and loses its grip. Wrong for light comedy and for kids content.  
**Prompt:** `big close-up, forehead to chin, top of head cropped, uncomfortably close, sharp focus on the eyes, static camera`  
**Example:** The Silence of the Lambs, Lecter's face filling the frame as he speaks to Clarice.  
`size.bcu`

### Extreme Close-Up

**Also called:** ECU, XCU  
**What it is:** One part only of the face or the body: an eye, a mouth, a hand, a finger on the trigger.  
**Effect on the audience:** It tells the viewer that this matters, pay attention. It also builds tension because it hides everything else.  
**Used for and where it works best:** Mystery and suspense, the beat before violence, the eye that flickers before a decision, the trembling lip. In westerns and thrillers it is the top rung of the size ladder (`size.ladder`); the Italian or Leone eyes shot (`size.italian`) is a horizontal-band composition inside this rung, not a tighter rung above it. In horror, an eye opening in the dark.  
**Best in:** formats: Feature Film, Drama Series Episode, Trailer / Teaser, Short Film, Anime (Japanese 2D), Music Video, Commercial / TVC, AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Thriller / Suspense, Western, Horror, Mystery / Whodunit, Crime / Procedural, Psychological Horror, Noir / Neo-Noir, Action  
**Avoid when:** Without context. The viewer has to know where we are first. In light comedy it feels overblown.  
**Prompt:** `extreme close-up on the eyes only, iris and lashes sharp, everything else out of frame, locked camera`  
**Example:** Psycho, the shower scene ending on the drain dissolving into Marion's eye.  
**Yields to:** `size.ws` — The viewer does not yet know where we are.  
`size.ecu`

### Detail Shot / Insert Shot

**Also called:** insert, detail, cut-in insert  
**What it is:** A close shot of something that is not a face: a phone screen, a letter, a watch hand, a key turning in a lock. Cut inside the scene to deliver information.  
**Effect on the audience:** It carries information without dialogue and gives the object weight. It also hands the editor an exit that rescues a bad cut.  
**Used for and where it works best:** Every scene where something matters in the character's hand: the letter, the weapon, the medicine, the artefact in heritage documentary. In commercials, the product detail. In cooking and craft, the step the viewer has to see. Shoot them even when they are not planned, this is the edit's safety net.  
**Best in:** formats: all | genres: all  
**Avoid when:** When it adds no information and no meaning it is filler. Do not open a scene on it before we know who and where.  
**Prompt:** `insert shot, close detail of the object on the table, hands entering frame, shallow depth of field, locked off camera`  
`size.insert`

### Macro Shot

**Also called:** macro, extreme magnification, probe lens shot  
**What it is:** Extreme magnification of something very small: an insect eye, the weave of a cloth, ink on clay, a drop of water. It differs from the extreme close-up by the degree of magnification: the tiny thing becomes screen sized.  
**Effect on the audience:** It opens a world the eye cannot see, which delivers wonder and discovery. It also makes the material itself, the fabric, the ink, the stone, something you can almost feel with your fingers.  
**Used for and where it works best:** Nature and science documentary, product commercials that sell texture (leather, cloth, a watch), heritage documentary on cuneiform or a cylinder seal impression. In Reels, open on an unreadable macro then pull out to reveal what it is.  
**Best in:** formats: Documentary Feature, Documentary Series, Heritage / Historical Documentary, Commercial / TVC, Product Video / Motion Design, Vertical Short-Form (Reels, Shorts, TikTok), Explainer, AI-Generated Clip, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography, Event Photography | genres: Nature / Wildlife, Science / Educational, Heritage (civilisation-focused), Historical / Biopic, Surreal / Absurdist, Horror, Body Horror  
**Avoid when:** On human skin unless you want body horror, because pores and hair turn alien. Do not hold it long either, the wonder is over in seconds.  
**Prompt:** `macro shot, extreme magnification of the cuneiform wedges pressed into clay, razor-thin depth of field, slow focus pull across the surface`  
**Example:** Microcosmos (1996), insect life filmed entirely in macro.  
**Yields to:** `size.ecu` — Ordinary detail, short of screen-filling magnification.  
`size.macro`

### Italian Shot / Leone Eyes Shot

**Also called:** Sergio Leone shot, eyes-only shot  
**What it is:** An extreme close-up cut as a horizontal band across the eyes only, with the forehead and the mouth cropped off. It is called the Italian shot because Sergio Leone made it a fixture of the spaghetti western.  
**Effect on the audience:** It turns a look into a weapon. The viewer reads intent from the eyes alone and the waiting becomes unbearable.  
**Used for and where it works best:** The peak of a standoff before the first shot, the beat where both sides measure each other. In westerns, samurai films and martial arts it is the top of the size ladder — the extreme close-up rung (`size.ecu`) cut as a band across the eyes, not a rung above it. In battle anime it lands just as hard because a drawn eye carries the scale.  
**Best in:** formats: Feature Film, Drama Series Episode, Trailer / Teaser, Anime (Japanese 2D), Game Cinematic / Gameplay, Short Film, AI-Generated Clip | genres: Western, Samurai / Chanbara, Martial Arts / Wuxia, Thriller / Suspense, Action, Shonen Battle (anime), Crime / Procedural  
**Avoid when:** Outside a confrontation, because it is a shot of challenge, not of sympathy. Do not use it in an ordinary dialogue scene.  
**Prompt:** `Italian shot, extreme close-up strip framing only the eyes, forehead and mouth cropped out, sweat on the brow, eyes locked forward, camera locked off`  
**Example:** The Good, the Bad and the Ugly, the cemetery standoff cutting between eyes.  
**Yields to:** `size.cu` — Outside a confrontation, when the look must read as sympathy.  
`size.italian`

### Master Shot

**Also called:** master, cover shot  
**What it is:** One shot that covers the whole scene from start to finish, usually wide or medium wide, with every character in it. The rest of the coverage is cut inside it.  
**Effect on the audience:** It gives the viewer the map: who stands where and what the distances are. Coming back to it after the close-ups reads as relief or as coldness, depending on the beat.  
**Used for and where it works best:** The first shot of any dialogue or group scene, and the safety net the editor returns to when the coverage does not cut. Sitcom and filmed theatre live on it. In AI video, generate the master first and lock the identity of the space from it before you generate the coverage.  
**Best in:** formats: Feature Film, Drama Series Episode, Sitcom / Multi-Cam Comedy, Soap Opera / Telenovela, Short Film, 3D CG Animation, Anime (Japanese 2D), AI-Generated Clip | genres: all  
**Avoid when:** Do not leave a scene on the master with no coverage unless that is a deliberate stylistic choice. And do not shoot it from an angle that does not work, because it will end up in the cut anyway.  
**Prompt:** `wide master shot covering the whole room, all three characters visible with clear spacing, eye level, static camera holding for the full scene`  
**Yields to:** `size.wide_vs_tight_master` — Long dialogue scene carried by relationship, not geography.  
`size.master`

### Establishing Shot

**Also called:** establisher, geography shot  
**What it is:** A wide or extreme wide at the head of a scene or a sequence that says where we are and when: the city, the house, the desert, night or day.  
**Effect on the audience:** It settles the viewer because they know where they stand, and it sets the mood before a word is spoken.  
**Used for and where it works best:** The first beat of every new location in a film or a series, and the head of every documentary chapter. In heritage documentary, the site from the air or from a distance before going in for detail. In a trailer, one shot that delivers the whole world in a second. In Reels cut it to one second or drop it entirely.  
**Best in:** formats: Feature Film, Drama Series Episode, Documentary Feature, Documentary Series, Heritage / Historical Documentary, Travel / Real Estate / Hospitality, Trailer / Teaser, 3D CG Animation, Game Cinematic / Gameplay, Generated Keyframe Feeding Video, Poster / Static Design, Portrait Photography, Street Photography, Documentary Photography, Sports Photography | genres: all  
**Avoid when:** When you want the viewer disoriented or ambushed by the place, dropping it is the tool. Do not give every scene an establisher or you get television monotony.  
**Prompt:** `establishing shot, wide exterior of the walled city at dusk, slow aerial approach settling on the main gate`  
**Example:** The Shining, the aerial opening following the car up the mountain road to the Overlook Hotel.  
`size.establishing`

### Re-Establishing Shot

**Also called:** re-establisher, return to the wide, reorientation cut  
**What it is:** A return to a wide in the middle or at the end of a scene, after a run of close-ups or after the geography has changed: someone stood, someone moved, someone new came in. The shot redraws the map.  
**Effect on the audience:** It lets the audience breathe and releases the tension, or cools the moment down deliberately after a peak, and it reminds the viewer of the distance between people after they had forgotten it. It also prevents confusion before it starts: the viewer sees the new arrangement once and then accepts every close-up that follows without questioning it.  
**Used for and where it works best:** When someone moves, enters or leaves, after an emotional peak you want to come down from, and just before the scene exits. It is essential in table scenes, meetings, fights, courtrooms and any long scene, and the cue is simple: the moment you feel lost in the positions while watching, that is where it belongs. In drama and series the wide after a confession that shows two people far apart says everything without a line, and in war and heritage work you need it after every cluster of fight shots.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Short Film, Sitcom / Multi-Cam Comedy, Soap Opera / Telenovela, Documentary Feature, Sports Broadcast / Sports Film, Anime (Japanese 2D), 3D CG Animation, Game Cinematic / Gameplay | genres: Drama, Crime / Procedural, Legal Drama, Medical Drama, Thriller / Suspense, Action, Comedy, Historical / Biopic, Heritage (civilisation-focused), War, Family, Tragedy  
**Avoid when:** When the tension is at its highest and you want to hold it, the wide breaks it and cools the scene, so return on a transition beat rather than on a climax. Do not go back to it after every line either, and a scene you want to feel suffocating usually holds better with no re-establisher, unless the new wide is itself the trap and reveals that the space is tighter or emptier than the viewer assumed.  
**Prompt:** `re-establishing wide shot after the close-ups, both characters small and far apart in the empty room, static camera holding the silence`  
`size.reestablishing`

### Wide Master vs Tight Master

**Also called:** tight master, close master  
**What it is:** The wide master shows the whole geography. The tight master is a closer framing, a medium two shot for instance, that still covers the entire scene with every character in it, minus the extra room.  
**Effect on the audience:** The wide gives you place, the tight gives you faces and relationship. An editor holding both can live in the tight and go back to the wide only when the scene needs it.  
**Used for and where it works best:** The long dialogue scene: shoot a wide master for geography and a tight master for relationship, and cut between them instead of leaning entirely on singles. In series and sitcom it saves shooting days. In AI video, generate two versions of the same scene with identical location and lighting.  
**Best in:** formats: Feature Film, Drama Series Episode, Sitcom / Multi-Cam Comedy, Soap Opera / Telenovela, Limited Series, Short Film, AI-Generated Clip | genres: Drama, Comedy, Rom-Com, Crime / Procedural, Legal Drama, Family, Slice of Life  
**Avoid when:** When the location itself is telling the story (a palace, a temple, a ruin), the tight master throws it away. And do not let the tight clip the edges of your characters.  
**Prompt:** `tight master, medium two shot holding both characters for the whole conversation, background softly out of focus, static camera`  
**Yields to:** `size.master` — When the location itself is telling the story.  
`size.wide_vs_tight_master`

### The Shot Size Ladder as a System

**Also called:** size progression, shot scale escalation  
**What it is:** Sizes are not a list, they are a ladder: extreme wide, wide, full, medium full, medium, medium close-up, close-up, big close-up, extreme close-up. Named compositions such as `size.italian` live inside a rung; they do not add a new size above ECU. A scene starts high on the ladder and steps down as the emotion rises, then climbs back up when the tension releases.  
**Effect on the audience:** The viewer feels the approach without knowing why, because every step down removes a piece of the location and leaves the face. Going back wide afterwards reads as a breath.  
**Used for and where it works best:** Plan every dialogue scene as a ladder: wide master, two shot, over the shoulder, medium close-up single, close-up on the peak line, then a wide to close. In thriller and drama this progression is the tension structure itself. When cutting on the same axis, skip at least two rungs (wide to medium close-up) so the cut does not read as a jump.  
**Best in:** formats: all | genres: all  
**Avoid when:** Do not start on the last rung, you will have nowhere to go. And do not step down the ladder without an emotional reason, an empty approach reads as manipulation.  
**Prompt:** `shot progression for the scene: open on a wide master, cut to a medium two shot, then medium close-ups, land on a close-up at the peak line, release to a wide`  
`size.ladder`

### The One Size Tighter Rule

**Also called:** tighten on the turn, escalate one step  
**What it is:** On the turn, the coverage of the character drops one step closer than the setup before it: medium to medium close-up, then medium close-up to close-up. The whole pair, shot and reverse, steps down together, and the step is taken across the cut to the other character, so no cut lands one rung tighter on the same axis — that cut needs the two-rung skip of `size.ladder`.  
**Effect on the audience:** Every turn in the scene registers in the viewer's body because the face grows a little. It is hidden escalation, unnoticed and effective.  
**Used for and where it works best:** When the dialogue turns from ordinary to dangerous, when the lie is exposed, when the question becomes personal. In interrogations, courtroom scenes and family confrontations it is the main engine. In documentary, a second camera on the interview gives you the tighter size ready for the beat where the subject breaks.  
**Best in:** formats: Feature Film, Drama Series Episode, Limited Series, Interview / Video Podcast, Documentary Feature, Short Film, Anime (Japanese 2D) | genres: Drama, Thriller / Suspense, Crime / Procedural, Legal Drama, Tragedy, Melodrama, Mystery / Whodunit, Political Thriller  
**Avoid when:** Do not step down on every line or the scene runs out of close-ups halfway through. And do not tighten one character while leaving the other where they were, unless the point is that only one of them is affected.  
**Prompt:** `as the accusation lands, cut one size tighter on both characters, from medium shot to medium close-up, same angles and lighting, each step taken across the cut to the other character, never on one character's own axis`  
`size.one_size_tighter`