The Frame Points The Eye

The masterclass opens by refusing its own premise. There is no right or wrong in filmmaking, only more interesting and less interesting. Every technique that follows exists to direct the eye, and every broken rule in the guide is a deliberate redirect of it. That refusal is the reason the material is worth keeping. Composition is the mechanism that decides where the eye lands before it decides what the shot looks like, and a mechanism can be written down. A taste cannot. So the guide reduces the discipline to 9 techniques. Focal points. The rule of thirds. The golden triangle. Leading lines. Positive and negative space. Balance and symmetry. Framing within a frame. Depth. Colour and contrast. 9 names, each of which goes into a prompt as a name rather than as a mood. The test case is 8 shots from 1 film, worked shot by shot, with the technique chosen for that shot instead of applied across the whole piece. 8 rows in 1 table, and the same discipline underneath all 8. The guide itself runs to 118 lines and 1,470 words, distilled from a masterclass transcript dated 27 August 2026. It stays short because 9 names do not need more room than that. 1 question sits under the 9. Where do I want them looking, and what should that make them feel? Anything that does not serve the answer is dropped, including a rule pulled from the 9.

The first technique is the cheapest and the most often skipped. Give the frame 1 dominant point and kill everything else. In a wide shot the method is subtraction. Empty the frame until the eye has nowhere else to go. A wide frame with 1 subject beats the same frame with 5, and the 4 elements removed are the reason it beats it. A busy frame can still work, and the guide is exact about when. The subject has to dominate through 1 of 4 levers: size, position, light or focus. Take all 4 away and the frame dissolves into texture. The rule of thirds is 2 horizontal lines and 2 vertical lines, cutting the frame into 9 cells at 33 and 66 per cent on each axis. The 4 intersections are the power points, and 33 by 33, 33 by 66, 66 by 33 and 66 by 66 are the 4 places the eye goes without being told. Subjects in nearly every Hollywood film sit on 1 of those 4. For an interview or a dialogue beat the eyes go on the upper horizontal line, which is 1 line and 1 decision repeated on every setup. The golden triangle is the third technique and it is a different shape. 1 diagonal runs corner to corner, and 2 lines from the other 2 corners meet that diagonal, producing 2 balanced focal points inside 1 frame rather than 1. Centre framing is the exception the guide keeps on purpose. It reads as authority, and it works when somebody speaks straight to camera, which is why the teaching register and the to-camera host register both sit dead centre in the frame.

Leading lines do the most narrative work for the least instruction. Roads, bridges, stairways, hallways, walls and buildings all point somewhere, and every 1 of them can be turned to point at the subject. The storytelling version is exact rather than poetic. Leading lines show where a character is going, because the lines literally point there. The guide also names the break. Broken, crossing and disordered lines read as confusion and disorder, which means the same technique carries the opposite meaning the moment it is inverted. Balance and symmetry come next, and they work below the level of naming. A balanced frame has no side heavier than the other. A viewer usually cannot say why an unbalanced frame is wrong while still feeling that it is, and the fix is 1 move: add an element to the light side. Symmetry is both sides mirroring, which reads as order. Deliberate imbalance is the break, and it makes an audience uneasy on purpose. Both are decisions, and neither is an accident of framing. For wide shots and horizon shots the guide moves the subject off centre and onto 1 of the 2 horizontal lines, at 33 or 66 per cent down the frame, because a dead centre subject flattens the interest and leaves the 4 intersections around it wasted.

Space is the technique people notice last and feel first. Positive space is the area given to the subject. Negative space is the empty area around them. The guide maps 3 readings onto the same frame. A subject larger in frame reads as more important. An extreme wide with heavy negative space reads as isolation and smallness. An extreme close up reads as power, and it pulls an audience into the emotion of the shot. Headroom is the negative space above the head, and the guide keeps it minimal. Too much headroom makes a subject feel small and uncomfortable, which is 1 misplaced line doing real damage to an otherwise correct frame. Lead room is the negative space in front of the subject, in the direction of the gaze or the movement, and it usually wants more of it than sits behind. Too little lead room reads as trapped and lost. For a moving subject the number is roughly 2 thirds of the frame in the direction of travel, which is 66 per cent of the width held ahead of the movement. Not half. 66 per cent. 1 shortcut produces the headroom and the lead room at the same time: put the eyes on an intersection of the rule of thirds. 1 placement, 2 problems closed, and no arithmetic in the moment.

Composition decides where the eye lands before it decides what the shot looks like, and a mechanism can be written down where a taste cannot.

Framing within a frame does 2 jobs at once. Put the subject inside a shape that already exists in the scene. A doorway, a window, an arch, a circle, a square. It gives the viewer a smaller zone of focus, which declares who and what matters. It also carries an emotional dial. A small frame reads as claustrophobia and discomfort. A frame sitting between the audience and the character turns the audience into a distant spectator, which is separation rather than intimacy. Depth is the answer to a 2 dimensional medium pretending to be 3 dimensional, and the labour is small. 3 layers: foreground, midground and background, with the subject usually carrying the midground. Shallow depth of field blurs the background, and the eye goes to whatever is in focus, which removes the distraction and the argument in the same move. Even a small out of focus foreground object adds 1 more layer. The rule that gets skipped most often in a small room is this: pull the subject off the wall. Somebody shot flat against a wall has 0 depth. Move them as far from the background as the room allows, or shoot into a corner and get 2 surfaces instead of 1. Rack focus is the last depth tool and the most underrated. A viewer notices a thing at the moment the camera focuses on it, so the focus pull can say what is important without a line of narration. Colour and contrast close the 9. A subject in a distinctive colour pops out of the frame. A dark subject against a monochromatic ground separates cleanly, and complementary colours separate the subject from the background with 2 hues instead of 1. Contrast is the blunt version. The eye goes to the brightest part of the image, so a lit subject in darkness locks the focus on them, and a silhouette reverses the same trick with the light placed behind.

There is 1 caution sitting on top of all 9, and it is the reason a well written prompt still fails. Framing words are the fastest route into a glossy render. The guide's own warning is that this vocabulary is where the video model gets most cinematic glossy. The fix is 1 habit rather than 1 setting. Pair every framing instruction with the documentary anchor: documentary film, 35mm, natural skin texture. Without that anchor the frame comes back as a digital painting instead of a shot. The prompt formula holds 5 slots. The subject and identity block first, then action and motion, then camera and framing, then lighting and atmosphere, and the anchor last. 5 slots, and the last 1 is never optional, which is 1 rule that protects the other 4. The reason to write it down is the same reason the studio writes its other numbers down. The cut grammar is a change every 0.5 to 2 seconds. Captions sit at 36 to 60 point. Vertical frames land at 720 by 1,280 and music runs at 110 to 128 beats per minute. A framing technique deserves the same treatment: a named instruction in a prompt rather than an instinct revisited on every shot. 30 years of building, in 12 countries, under 2 rules: work with what the constraint allows, and refuse the version that looks better on paper than it can look in the room. Built companies across twelve countries. Restructured an industrial group worth EUR 75 million. Deployed 210 energy systems. Took a company through a Frankfurt IPO. Built 200 wood gasification machines at Entrade. Signed a EUR 20 million personal guarantee on a company that folded on Christmas Day 2008. In the building for the 2003 blackout. 393 pages in paperback, published 8 April 2026. The probability that a frame built from 1 named technique and 1 anchor reads as intentional is high. The probability that a frame assembled from adjectives reads as a digital painting is higher than most teams expect, and it costs 2 render cycles to find out. So the method fits on 1 line: name the technique, serve the question, anchor the render. 9 techniques, 4 levers, 3 layers, 2 thirds, 1 anchor.