Procedural Shading & Texturing

A part of the slides of my work (before joining Autodesk) that I used at CEDEC some years back.

Transcript

1. Procedural Shading & Texturing Shinji Ogaki
2. Procedural Shading • Procedural shading Procedural shading is a proven rendering technique in which a short user- written procedure, called a shader, determines the shading and color variations across each surface. - Marc Olano • Procedural texture A procedural texture is a computer-generated image created using an algorithm intended to create a realistic surface or volumetric representation of natural elements such as wood, marble, granite, metal, stone, and others, for use in texture mapping. - Wikipedia
3. Advantages • Small memory footprint • Resolution independent
4. Famous Algorithms • An Image Synthesizer – Ken Perlin (Siggraph 1985) • A Cellular Texture Basis Function – Steve Worley (Siggraph 1996)
5. Today’s Topics • Voronoi cell • Fourier transform
6. Voronoi Cell
7. Voronoi Cell • Useful for mesoscale features (micro – meso - macro) • Inclusions, scratches, etc. Modeling Aventurescent Gems with Procedural Textures –Weidlich et al. (SCCG 2008) Multi-Scale Rendering of Scratched Materials using a Structured SV-BRDF Model – Raymond et al. (TOG 2016)
8. Voronoi Cell • Grid based Algorithm • No precomputation • Zero memory cost
9. Shading point Feature point 2D
10. Shading point Closest feature point 2D
11. Single Level
12. Multiple Levels
14. Coloration • Use exemplar • Histogram matching • Poor performance • Spatial information is not taken into account • Fractional Brownian motion in 2D texture space
15. Voronoi Cell – Hash Function • 4D vector input – 4D vector output In (x,y,z,t): x,y,z:grid t:instance ID, etc. Out(x,y,z,t): x,y,z:feature point t:normal, color, density, etc. • Chaotic Hash Algorithms • Secure Hash Algorithms • MD5, SHA-0, SHA-1, SHA2, SHA-3, etc.
16. Voronoi Cell – Hash Function • 4D vector input – 4D vector output In (x,y,z,t): x,y,z:grid t:instance ID, etc. Out(x,y,z,t): x,y,z:feature point t:normal, color, density, etc.
17. w/o instance ID w/z instance ID
18. Voronoi Cell – Hash Function • 4D vector input – 4D vector output In (x,y,z,t): x,y,z:grid t:instance ID, etc. Out(x,y,z,t): x,y,z:feature point t:normal, color, density, etc.
19. Scratches • Remove if δ<density • Stretch
20. Voronoi Cell – Hash Function • 4D vector input – 4D vector output In (x,y,z,t): x,y,z:grid t:instance ID, etc. Out(x,y,z,t): x,y,z:feature point t:normal, color, density, etc.
21. Inclusions • Ray marching • Normal vector • Sample warping δ>density, density’=density/δ • Spherical Fibonacci lattice id=(int)(N x density’)
22. Deformation • Surface: Pref (Reference position) • Volume: Open problem (Tetrahedrarization? Curve interpolation?)
23. Fourier Transform
24. Fourier Transform • Convolution • Ocean waves • Useful tool to design or analyze procedural noise patterns
25. Phase
26. Wave
27. Coil
29. Fourier Transform Convolution
30. Convolution Input Kernel Input (Frequency space) Kernel (Frequency space) FFT FFT Element-wise multiplication IFFT Output
31. Convolution
32. Convolution
33. Convolution
34. Convolution
35. Convolution
36. Convolution
37. Convolution
38. Convolution
39. Convolution
40. Convolution
41. Convolution
42. Convolution Sum
43. Convolution Blurred signal
44. Convolution Input Kernel Input (Frequency space) Kernel (Frequency space) FFT FFT Element-wise multiplication IFFT Output
45. Convolution 2D
46. Convolution 2D • Gabor noise
47. Convolution 2D • Glare effect
48. Fourier Transform Ocean Waves
49. Ocean Waves • Sum of multiple waves
50. Ocean Waves • Sum of multiple waves
51. Ocean Waves • Sum of multiple waves
52. Ocean Waves • Sum of multiple waves
53. Ocean Waves • Sum of multiple waves
54. Ocean Waves • Sum of multiple waves
55. Ocean Waves • Empirical Directional Wave Spectra for Computer Graphics - Christopher J. Horvath (Digipro 2015) IFFT Ocean! Generate waves (Points in frequency space)