Why
Cutting rate is the closest thing film has to tempo. An audience reads it pre-consciously, the way they read a musical accelerando, and it tells them how to feel before any content arrives.
The mapping is not arbitrary. Short shots deny processing time, which produces disorientation, urgency, and — if sustained too long — numbness. Long shots grant processing time, which produces dread, intimacy, or boredom depending on what is in the frame. Neither is better. Using one setting for the whole piece is what is wrong.
The working table
| Emotional register | Typical duration | Why |
|---|---|---|
| Panic, chaos, impact | 0.3–1.0s | Below the threshold of full comprehension. The viewer feels before they understand. |
| Urgency, pursuit | 1.0–2.0s | Legible but rushed. Enough to read the frame, not enough to relax. |
| Tension, conflict | 2.0–4.0s | Standard dramatic coverage. The default, and therefore the least interesting choice. |
| Contemplation, grief | 4.0–8.0s | The viewer runs out of things to look at and starts feeling instead. |
| Dread, unease | 8.0–20.0s | Duration itself becomes the threat. Something should have happened by now. |
| Awe, scale | 6.0–15.0s | Wide shots need time to be read. A two-second landscape is a waste of a landscape. |
| Comic beat | 1.5–3.0s + 1 beat | The hold after the joke is the joke. |
These are starting points, not laws. The two real rules are underneath the table:
Rule 1 — contrast beats absolute value. A 4-second shot after a run of 1-second shots feels enormous. The same shot after a run of 12-second shots feels rushed. Duration is relative.
Rule 2 — never repeat the same duration twice in a row. Two consecutive shots of identical length produce a metronome, and the audience stops feeling the rhythm because it has become predictable. Vary by at least 30%.
How
- Set the scene's emotional register from its exit charge, then pick the band from the table.
- Assign durations shot by shot within the band, varying each from its neighbour.
- Compute the scene's average shot length and compare it against the format's target. A drama scene running at a 1.4-second ASL is an action scene wearing a drama script.
- Place the longest shot of the scene at the emotional peak, not at the establishing shot. The instinct to spend time on the wide is almost always wrong.
- Before a sustained short-cut sequence, spend one long shot establishing geography. Fast cutting is only comprehensible against a known space.
- Check the cost: shortening shots shortens runtime, which in a fixed-length format means either more shots or a lost beat. Decide which before cutting.
Notes for AI film
This card carries unusual weight here, because generation models charge by the second and have hard duration ceilings — currently 5 to 10 seconds per clip on most video targets. That constraint quietly pushes every AI film toward a uniform 4-to-6-second ASL, which is precisely the flat rhythm this card exists to prevent.
Fight it deliberately. Long holds must be planned as multiple stitched generations with a locked seed and a static camera, and short cuts must be planned as deliberate sub-second slices of longer generations. Neither happens by accident, and neither survives being left to the export stage.