Sakuga
How a cut gets made

Process

Storyboard to broadcast: the seven real stages a cut passes through, and the frame-hold decision made at every one of them.

CUT 01

The seven stages

In production order
01 Storyboard (e-conte) The director draws every cut as a rough panel with camera direction, rough timing in seconds, and dialogue — the single document the rest of the episode is built from.
02 Layout The storyboard panel is redrawn to final proportions and camera framing — background and character positions locked before any finished drawing begins.
03 Key animation (genga) The key animator draws the extremes of a movement — the poses that define it — plus rough timing notes for how many inbetweens each gap needs.
04 Inbetweens (douga) The inbetween animator draws every frame between each pair of key poses, interpolating position along the path the key animator implied — by far the largest headcount on a production.
05 Finishing (shiage) Digital ink-and-paint: every genga and douga line is cleaned up and filled with the character's official colour model — the same flat colours in every episode, checked against a colour-design reference sheet.
06 Compositing (satsuei) Literally "photography" — a term inherited from the rostrum-camera film era. Characters, backgrounds and effects layers are composited, graded, and given camera moves digitally.
07 Editing Finished cuts are cut together against the sound and dialogue track to the storyboard's intended rhythm, and timing is adjusted where a cut runs long or short.
CUT 02

Frame-hold reference

By convention

Every hold in this table is exact arithmetic against the same 24fps drawing base — not a stylistic label.

On ones (1コマ撮り) 24 ÷ 1 = 24 fps
On twos (2コマ撮り) 24 ÷ 2 = 12 fps
On threes (3コマ撮り) 24 ÷ 3 = 8 fps

On ones: A new drawing every frame — 24 unique drawings a second. Reserved for the highlight cuts (sakuga) a studio wants to sell: a fight climax, a transformation, a hero shot. Expensive in genga and douga both, so it is spent deliberately, not everywhere.

On twos: Each drawing held two frames — 12 unique drawings a second. The default for the large majority of a TV episode: dialogue scenes, walk cycles, ordinary reaction shots. Motion still reads as smooth to most viewers; drawing count is halved.

On threes: Each drawing held three frames — 8 unique drawings a second. Used for slower, calmer or lower-priority cuts, often mixed with held poses and camera moves on the background layer instead of the character, to keep a scene's cost down without going fully static.