Simulating Wax Crayons Dave Rudolf David Mould Eric Neufeld Depart. of Computer Science, U. of Saskatchewan
Overview Previous Work Goals Representation of Paper and Wax Interactions between Wax and Paper Rendering Results
Previous Work Adobe Photoshop TM Conté Crayon Filter
Previous Work Corel Painter TM Grainy Crayon Brush
Previous Work Viscous Paint Model (Baxter et al.)
Goals Improve realism Consider individual strokes Suitable paper texture Approximate mechanics of crayons Simulate optical properties of wax Efficient Rendering Not too insanely realistic
Representation of Paper Height-Mapped Texture Texture Synthesis Other sources (scans of Real Paper)
Convolution Mask Somewhat Expensive Can be Pre-computed Van Wilk’s Lunar Texture
Other Textures Make no assumptions about texture
Representation of a Crayon 2D Mask Height Area
Compression of Wax
Assume that wax compresses linearly Hooke’s Law: F = Y L A / L 0 Can’t solve directly Crayon & Paper profiles are not predictable Must binary-search for force ( F )
Compression of Wax
Wax Deposition
Determined by: Friction Force of Crayon Direction of Stroke Whether there is existing wax F f = F n = N cos(N, F c )
Wax Smearing
Depends on texture and crayon force Flow Smearing Independent of crayon heading S f xy = z Directional Smearing Depends on crayon heading S d xy = cos( (x, y), V )
Deposition and Smearing Without Smearing With Smearing
Rendering Need a good colour model Wax scatters and absorbs light Additive/Subtractive models are insufficient Kubelka-Monk Colour Model Approximates of Translucent Pigments Considers transmittance, scattering, and interference of pigment layers.
Approximated Kubelka-Monk Real Crayon Generated Crayon
Results Not Quite Interactive Between 0.3 and 2.0 seconds per frame Updating model takes the most time Not Quite Realistic No wax flaking, carving, transporting
Results
Questions ?