Shape-Aware ASCII Renderer

6D character matching with contrast enhancement for high-quality ASCII art. Based on Alex Harri's technique.

Proof model

ASCII characters have shape, not just brightness.

The renderer compares 6D character vectors so contours survive the translation into text.

Sample

Character cells are measured.

Vector

Six regions describe the shape.

Match

Contours choose the character.

Render

ASCII preserves edges.

 
 
0×0 characters 0.0ms render time Frame: 0
 
Scene
 
   
 
 
Contrast Enhancement
 
 
 
Cell Size
 
 
 
Quality
 
 
How It Works
 
6D Shape Vectors
 
Each ASCII character is represented as a 6D vector capturing how it fills its cell.
					Six sampling circles in a staggered 2×3 grid measure density at different positions.
 
Nearest Neighbor Matching
 
For each cell in the image, we sample a 6D vector and find the character whose
					pre-computed shape vector is closest in Euclidean distance.
 
Global Contrast
 
Normalizes the sampling vector to [0,1], applies an exponent to crunch darker values,
					then denormalizes. Exaggerates shape differences for better character picks.
 
Directional Contrast
 
External sampling circles detect edges by reaching into neighboring cells.
					When a neighbor is lighter, it pushes the internal value down, sharpening boundaries.