SL-22 / Iconic Sequences
The Plan
Two people talk through a plan. The camera and the cut do every single joke.
The brief
Build a rapid-fire planning montage between two flatmates: hard whip-pans snapping between them, snap zooms punching in on each beat, match cuts landing on a repeated gesture, and every transition hitting on a percussive sound. Then do the whole thing a second time, shot for shot, with small deliberate variations — the repeat is the joke.
The source sequence
Study the technique, not the copy. Rebuild the blocking, the lens, the movement and the rhythm — not the performances.
Watch the sequence
Embedded via youtube-nocookie.com — YouTube sets no tracking cookies until you press play.
Why it's hard
This is a rhythm problem stacked on a consistency problem. Models smooth whip-pans into leisurely pans and drop snap zooms entirely, so the comic timing dies. Worse, the two faces must survive a dozen fast cuts and then be reproduced shot for shot in the repeat — identity drift that a slow scene would hide is glaring here.
Skill Tests
06 / 28Every challenge is scored against the same vocabulary, so you can compare them. Lit cells are what this one tests.
Success criteria
- Whip-pans carry real directional motion blur and resolve sharp — they are moves, not dissolves.
- Both faces are recognisably the same people across every cut, including the repeat pass.
- The second pass matches the first shot for shot, so the deliberate variations actually read as jokes.
Starter prompt
Built with the CEMI MediaMax system for Google Veo 3.1. Prompts stay in English — that is the language these models take instructions in. Treat it as a starting point and start varying it.
Cramped flatshare kitchen, mid-morning, flat grey light from one north window. Two flatmates in their thirties — one in a red tie, one in a stretched grey t-shirt — rattle a plan across a chipped Formica table. Build it as six metronomic beats. Hard whip-pan between them in a fifth of a second, the room smearing to streaks, snapping sharp on the next speaker; then a snap zoom punching medium to close on each new item, micro-overshooting and settling like a hand on a zoom ring. Match-cut three repeated gestures — a pointed finger, a lifted pint glass, a shrug — every transition landing on a percussive hit. Register of Edgar Wright (dir.) and David M. Dunlap (DP): 21mm held close, mild barrel distortion, hard window-practical contrast, crushed blacks. Dirty foreground: an out-of-focus cereal bowl crusted with dried milk, a dead lighter at the lens edge. Hold camera positions, lens and beat count identical so the montage regenerates shot-for-shot. No on-screen text.
Style reference: Edgar Wright (dir.) / David M. Dunlap (DP), rhythmic whip-pan and snap-zoom montage grammar
The joke is carried by the cut, not the performers, so the prompt fixes a beat count and gives every transition a job — whip-pan, snap zoom, gesture match cut — each landing on a sound hit rather than dissolving. Wright and Dunlap are named because this grammar is theirs and "kinetic" alone gets you nothing; the human operator shows up in the zoom's micro-overshoot, and the crusted cereal bowl keeps a dirty foreground in a scene the model would otherwise render spotless. Notice that the repeatability is a camera problem: lock positions, lens and beat count first, then vary only one thing on the second pass — a swapped gesture, a held beat, a zoom that arrives late.
The model turns the whip-pans into slow drifting pans and the snap zooms into gentle push-ins, so the montage arrives smooth and beatless; if the smear never appears, generate one two-beat pair at a time with the durations spelled out ("0.2 seconds, full motion blur") and stitch the passes in the edit.
Ran a challenge? Show us what you got.
Share your attempt — the triumphs and the glorious failures. The most instructive entries get featured on the board.