cartoon effect

cartoon effect applied to the studio's source frame — beforecartoon effect applied to the studio's source frame — afterSourceEffect

01/ ORIGIN

Cel animation forced this look before anyone chose it. Each frame was ink outlined on the front of a transparent sheet and painted flat on the back, so a character could only ever be areas of solid colour bounded by a line, and shading meant adding another flat area rather than a gradient. Reusing cels across frames made flatness an economic necessity as well as a stylistic one. The convention outlasted the process by decades.

02/ SWEEP

LEVELS sweep

LEVELS 2: two flat areas only, the most graphic possible read
LEVELS 2LEVELS 2: two flat areas only, the most graphic possible read
LEVELS 5: the default, enough areas to model the form
LEVELS 5LEVELS 5: the default, enough areas to model the form
LEVELS 12: many bands, approaching continuous tone under the lines
LEVELS 12LEVELS 12: many bands, approaching continuous tone under the lines

03/ SPEC

This effect quantises tone into a small number of flat bands and composites detected edges over them as dark outlines, which is posterisation and edge detection working together rather than either alone. LEVELS runs from 2 to 12 with a default of 5 and sets how many flat colour areas the image is reduced to, which is the control that decides between a hard graphic read and something closer to painted animation. EDGE THRESHOLD governs how readily a contour becomes a line, and EDGE WIDTH its weight. In kott it renders in natural colour so the flat areas carry the source hue.

LEVELS2 to 12 · default 5
CONTRAST0.2 to 3 · default 1.1
EDGE THRESHOLD0 to 2 · default 0.5

04/ FAQ

What is a cartoon effect?
A cartoon effect reduces an image to a few areas of flat colour and draws dark outlines along detected edges, reproducing how cel animation was inked and painted.
What LEVELS value looks most like animation?
Four to six. Traditional cel work used very few flat values per character, so low numbers read as animation while higher ones start looking like a photograph with outlines added.
Why do the outlines look broken in places?
Edges are detected from tonal change, so where two areas have similar brightness there is nothing to detect. Lower EDGE THRESHOLD to catch softer boundaries.