Camera.
The language of the shot: how we frame, where we look from, how the camera moves, and what optics and sensor change in the image.
From the wide shot to the extreme close-up: each shot size frames a portion of the subject and carries an intention.
Shot sizes
From the wide shot to the extreme close-up: each shot size frames a portion of the subject and carries an intention. Select a shot to locate it on the image.
Wide shot
Full shot
Italian shot
American shot
Medium shot
Medium close-up
Close-up
Extreme close-up
Types of shots
Establishing shot : A wide shot opening the sequence: it shows the place, the time of day and where the characters are before moving closer. The viewer knows where they are.
Two-shot : Two characters share the frame. Their distance, their position and their gazes already tell their relationship.
Over-the-shoulder : The camera films one character over the other's shoulder, who stays at the edge of the frame, from behind and often out of focus. The shoulder places both speakers in the same space.
Shot / reverse shot : Two shots answering each other: one films the first character, the other their partner, from the opposite side. It is the basis of covering a dialogue.
Point-of-view shot : The camera takes the place of the character's eyes: the viewer sees what they see. The shot is often framed by a shot of their gaze, before and after.
Reaction shot : A shot of the face of the one listening or discovering something. The reaction gives the action its meaning: in the edit, you often stay on the listener rather than the speaker.
Insert : A very tight shot on a detail: an object, a hand, a screen, a letter. It isolates information the viewer must not miss.
Composing the frame
The rule of thirds : In practice: divide the frame into three equal parts, horizontally and vertically. This draws four lines - the power lines - and four intersections - the power points. Placing the subject, the horizon or the gaze on one of these lines rather than dead-centre makes the composition more dynamic and better balanced, and gives the subject room.
Symmetry : The subject in the center, the set mirrored on each side. The image feels stable and ordered, sometimes oppressive or strange. A strong choice, the opposite of the rule of thirds.
Leading lines : Roads, rails, walls, rows: the lines of the set guide the eye to the subject. They also give depth to the image.
Frame within a frame : A door, a window, an arch: a second frame inside the image isolates the subject, confines it, or turns the viewer into a hidden observer.
Negative space : A small subject in a large void: the empty space weighs as much as the subject. Loneliness, waiting, isolation.
Shallow depth of field : Only the subject is sharp; foreground and background melt into blur. The eye goes straight to the subject. You get it with a wide aperture, a long lens or a large sensor.
Deep depth of field : Everything is sharp, from the foreground to the back of the set. The viewer chooses where to look and several actions coexist in the same shot. You get it with a small aperture and a short lens.
Camera angles
Eye-level : The camera is at eye level, axis horizontal. A neutral viewpoint: the viewer is on equal footing with the subject.
High angle : The camera is above the subject, looking down. A crushing effect: vulnerability, inferiority, loneliness.
Low angle : The camera is below, looking up. The subject gains power and dominance.
Bird's eye : Straight overhead, vertical. The scene becomes a map, a diagram: the drone's gaze, maximum distance and control.
Worm's eye : Straight below, looking up at the sky. The set crushes everything: excess, oppression, vertigo.
Dutch angle : The Dutch angle tilts the camera to skew the horizon: a feeling of unease, tension or psychological imbalance in the viewer.
Subject orientation
Frontal : The subject faces the camera: frontal, direct. If they look into the lens, they address the viewer (looking at camera) and break the fourth wall.
Three-quarter : The subject is turned about 45°. The most common orientation: the face stays readable and gains relief.
Profile : The subject is seen from the side, at 90°. They look elsewhere, in the frame or off-screen: a position of observer, of distance.
Three-quarter back : You mostly see the nape and the shoulder. The viewer looks with the character, almost in their place.
From behind : The face is hidden: mystery, threat, or identification when following the character into a place they discover.
Camera movements
Pan : The camera rotates horizontally on its axis, without moving. It sweeps the space from left to right.
Tilt : The camera rotates vertically on its axis, bottom to top. The viewpoint stays fixed, only the angle changes.
Dolly in : Dolly in: the camera physically moves toward the subject. The viewpoint changes - unlike a zoom.
Dolly out : Dolly out: the camera physically pulls away from the subject. It opens up the field, isolates the character in space, or closes a scene.
Lateral track : The camera moves sideways while keeping the subject framed. It reveals space and creates depth through parallax.
Pedestal : The camera rises or lowers in translation. It changes height without pivoting in place.
Arc : The camera orbits the subject, keeping it framed. An enveloping, often spectacular move.
Crane : Crane move: the camera combines elevation and travel, from the ground upward. It opens or closes a space.
Handheld : Handheld: the camera is held on the shoulder or by hand. Slight shakes bring presence, urgency, a documentary realism.
Zoom : Zoom: the camera doesn't move, the focal length changes. The field narrows, but the viewpoint stays the same.
Dolly zoom : Dolly zoom (Vertigo effect): the camera moves in while the field of view widens. The subject keeps the same size, but the background stretches and warps - vertigo, unease, an inner shift.
Steadicam : A stabilized arm, worn by the operator in a harness: the camera follows the characters, crosses rooms, climbs stairs, with the fluidity of a dolly but without tracks. The handheld gimbal is its light version.
Aerial shot : Filmed from a drone or a helicopter: the camera flies over the set and reveals its scale. Often at the start of a sequence, as an establishing shot.
Long take : A whole scene in a single shot, without a cut: the camera chains its moves (approach, orbit, rise). Continuous tension, real-time immersion, and a very demanding setup for the whole crew.
Whip pan : A pan so fast the image becomes a motion blur. It serves as a transition between two shots, a surprise, or adds energy.
Crash zoom : A brutal zoom in on the subject: a shock, a realization, often with a comedy or B-movie tone.
Body-mounted camera : The camera is rigged to the actor's body and films them from the front (snorricam): the face stays still in the frame, the set moves around them. Drunkenness, panic, loss of control.
Rack focus : The camera does not move: the focus slides from one element to another, here from the foreground to the subject. It moves the eye without cutting.
Motion control : A case apart: the camera sits on a robotic arm whose move is programmed and recorded, frame by frame.
Bullet time : Time freezes while the camera circles: an arc of still cameras fired one after another, each taking a single picture. Played in sequence, they make a camera move around a stopped instant.
Where a travelling shot is performed by hand, motion control replays the move identically, as many times as needed. You can shoot several passes of the same shot - empty set, actor, separate elements - all perfectly superimposable.
The benefits in post-production are enormous: clean compositing (passes line up to the pixel), multiplicity effects (the same actor several times in frame), set extensions, and above all a camera move that is known exactly - 3D tracking becomes trivial, since the camera's coordinates are already known.
On the set of The Matrix (1999), about 120 still cameras were lined up on a rig around the actor, with a film camera at each end. A computer fired them one after another at set intervals, or all at once for completely frozen time. Each camera takes a single picture: played in sequence, they form an impossible camera move around an almost motionless instant.
The shot is then rebuilt in post-production: in-between frames are computed between two cameras to smooth the move, the rig is painted out and the background replaced with CGI built from photos of the set. The idea goes back a long way: Eadweard Muybridge's camera arrays (1878), then artist Tim Macmillan's time-slice in the early 1980s. Supervised by John Gaeta, the effects of The Matrix won the Academy Award for Best Visual Effects.
Lenses: focal length
Focal length (in mm) sets the angle of view: the shorter it is, the wider the field.
In full-frame terms: a wide-angle (≈ 14-35 mm) opens the field wide and accentuates perspective - sets, space. A standard (≈ 40-60 mm, often the 50 mm) renders a view close to the eye. A telephoto (≈ 85 mm and up) tightens the field, compresses planes and isolates a distant subject - portraits, far details.
Sensors & size
The surface that receives light. Its size changes framing, depth of field and rendering.
Full frame (36 × 24 mm, inherited from 35 mm film) is the reference. Any smaller sensor captures only a central portion of the image: it “crops”. This is measured by the crop factor = 43.27 ÷ sensor diagonal - i.e. 1.0 for full frame, ~1.5 for APS-C, 2.0 for Micro 4/3.
At equal focal length, a small sensor gives a tighter frame and a large sensor opens the field. At equal framing, the small sensor needs a shorter focal length and gives greater depth of field; the large sensor isolates the subject better (background blur) and gathers more light. In cinema, Super 35 is the historical standard.
Focal length is a physical property of the lens: it does not change. It's the smaller sensor that records only a central part of the image - hence a tighter frame, expressed by the “full-frame equivalent”.
Formats & ratios
The ratio is the shape of the image (width:height); the resolution is its pixel count - two different things. Here are the common formats and their resolutions.
4:3 (1,33:1) : 2K : 1440 × 1080 px · 4K : 2880 × 2160 px
16:9 (1,78:1) : 2K : 1920 × 1080 px · 4K : 3840 × 2160 px
Flat (1,85:1) : 2K : 1998 × 1080 px · 4K : 3996 × 2160 px
DCI Full (1,90:1) : 2K : 2048 × 1080 px · 4K : 4096 × 2160 px
Scope (2,39:1) : 2K : 2048 × 858 px · 4K : 4096 × 1716 px
1:1 (carré) : 2K : 1080 × 1080 px · 4K : 2160 × 2160 px
4:5 (portrait) : 2K : 1080 × 1350 px · 4K : 2160 × 2700 px
9:16 (vertical) : 2K : 1080 × 1920 px · 4K : 2160 × 3840 px
Open gate : Shooting open gate means recording the entire sensor (the full “gate”) rather than a crop chosen up front (16:9, 2.39:1…). You get a taller image - often close to 3:2 or 4:3 - holding more information, both vertically and on the sides. The payoff is in post: from a single master you reframe to the target format (16:9, 9:16, 1:1, 2.39:1), stabilize or reposition the frame, with no loss of resolution. Don't confuse it with “full frame”: open gate isn't a sensor size but a capture choice, available on Super 35 or Micro 4/3 too. The trade-off: heavier files and sometimes a limited frame rate (fps).
Continuity rules
180-degree rule : An imaginary line links the two characters: the axis. As long as the camera stays on the same side, each keeps their place on screen (A on the left, B on the right). Crossing the axis swaps their positions and disorients the viewer.
30-degree rule : Between two consecutive shots of the same subject, the camera changes its angle by at least 30°, or clearly changes the shot size. Otherwise the cut looks like a mistake: a jump in the image (jump cut), which some filmmakers use precisely as an effect.
Eyeline match : When a character looks off-screen, the next shot shows what they are looking at. The directions answer each other: if they look right, their partner looks left. This is what makes two shots filmed separately seem to face each other.
L'équipe sur le plateau
Un plan, c'est un travail d'équipe. Les principaux rôles autour de la caméra et de la mise en scène :
Relever pour la post
On ne tourne pas seulement pour le plateau : on tourne aussi pour la post-production.
À chaque plan, un membre de l'équipe - souvent le DIT, le 2e assistant caméra, et sur les plans à effets le data wrangler ou le superviseur VFX - relève les informations techniques de la caméra : modèle, objectif et focale, ouverture, hauteur de la caméra, angle (tilt), distances, parfois des repères de tracking dans le décor.
Il assure aussi les sauvegardes (backups) : copie et vérification des cartes, redondance des fichiers - la matière première du film, qu'on ne peut pas perdre.
Pourquoi tant de soin ? En post, le tracking (ou matchmoving) reconstitue en 3D le mouvement réel de la caméra pour y intégrer des éléments - décor, créature, correction - parfaitement calés. Plus les données de tournage sont précises, plus l'intégration est rapide et juste. C'est aussi pourquoi un mouvement en motion control, déjà connu au pixel près, simplifie radicalement le travail.
Camera & VFX report, to download
Based on the VES-style VFX camera report: one line per take, start and end values, set references, backups.
Glossaire
Les termes clés de cette page, en une phrase.
Focal length equivalence
A focal length doesn't frame the same on every sensor. Enter a focal length and a sensor: here's the full-frame equivalent and the angle of view.
- Crop factor
- 1.00×
- Full-frame equivalent
- 50 mm
- Angle of view (H)
- 39.6°
Focal length is a physical property of the lens: it does not change. It's the smaller sensor that records only a central part of the image - hence a tighter frame, expressed by the “full-frame equivalent”.
3D characters: “Sophia Animated 003” and “Nathan Animated 003” by Renderpeople, CC BY 4.0.
Ten questions on shot sizes, angles, movements and focal lengths, in camera view.
