Reduced Aggregate Scattering Operators

Transcription

1 Reduced Aggregate Scattering Operators for Path Tracing s Adrian Blumer Jan Novák Ralf Habel Derek Nowrouzezahrai Wojciech Jarosz

2 Overview Reduced Aggregate Scattering Operators for Path Tracing 2

3 Overview Reduced Aggregate Scattering Operators for Path Tracing 3

4 Overview Reduced Aggregate Scattering Operators for Path Tracing 4

5 Overview Reduced Aggregate Scattering Operators for Path Tracing 5

6 Overview Reduced Aggregate Scattering Operators for Path Tracing 6

7 Motivation image sources: clouds: tree: snowmen: sheep: 7

8 Selected Related Work Clustered principal components for precomputed radiance transfer Sloan et al., 2003 Modular radiance transfer Loos et al., 2011 A Practical Model for Subsurface Light Transport Henrik et al., 2001 Much more: importance sampling, caching of light transport, vegetation rendering, subsurface scattering, 9

9 Precomputation Path tracer integration

10 Precomputation Path tracer integration

11 indirect illumination Method Light transport as a matrix operation direct illumination indirect transport matrix direct = 12

12 indirect illumination Method Light transport as a matrix operation direct illumination indirect basis transform transport basis transform direct = 13

13 Method Learning reduced bases configuration 1 # of vertices Modular radiance transfer Loos et al.,

14 Method Learning reduced bases configuration 1 configuration 2 # of vertices Modular radiance transfer Loos et al.,

15 Method Learning reduced bases configuration 1 configuration 2... # of vertices Modular radiance transfer Loos et al.,

16 Method Learning reduced bases configuration 1 configuration 2... # of vertices Modular radiance transfer Loos et al.,

17 Method Learning reduced bases configuration 1 configuration 2... # of vertices Modular radiance transfer Loos et al.,

18 Method Learning reduced bases # of vertices # of configuarions Modular radiance transfer Loos et al.,

19 Method Learning reduced bases # of vertices Compress using SVD + # of configurations truncate Modular radiance transfer Loos et al.,

20 Method Learning reduced bases # of vertices # of basis functions compact basis Modular radiance transfer Loos et al.,

21 Method Learning reduced bases # of vertices # of basis functions compact basis Modular radiance transfer Loos et al.,

22 Method Learning reduced bases # of vertices # of basis functions compact basis Modular radiance transfer Loos et al.,

23 Method Learning reduced bases # of vertices # of basis functions compact basis Modular radiance transfer Loos et al.,

24 Method Learning reduced bases # of vertices # of basis functions clustered compact basis 25 Clustered principal components for precomputed radiance transfer Sloan et al. 2003

25 indirect illumination Method Learning reduced bases direct illumination indirect indirect basis transform transport direct basis basis transform transform direct = 26

26 Precomputation Path tracer integration

27 Path tracer integration Monte Carlo evaluation 29

28 Path tracer integration batching and caching direct illumination cache 30

29 Path tracer integration batching and caching direct illumination cache 31

30 Path tracer integration batching and caching direct illumination cache 32

31 Path tracer integration batching and caching indirect illumination cache direct illumination cache between batches, apply transport 33

32 Path tracer integration batching and caching indirect illumination cache direct illumination cache 34

33 Path tracer integration Importance caching and sampling 35

34 Path tracer integration Importance caching and sampling Importance cache 36

35 Path tracer integration Importance caching and sampling Importance cache 37

36 Path tracer integration Importance caching and sampling Importance cache sampling weights between batches, apply transport 38

37 Path tracer integration Importance caching and sampling sampling weights 39

38 Path tracer integration Importance caching and sampling sampling weights 40

39 Results

40 Quality Path tracing Path tracing with RASO Side lighting (in the training set) *indirect illumination only 42

41 Quality Path tracing Path tracing with RASO Back-lighting (not in the training set) *indirect illumination only 43

42 Quality Path tracing Path tracing with RASO Ambient sky (not in the training set) *indirect illumination only 44

43 Quality Path tracing Path tracing with RASO Front-lighting with occluder (not in the training set) *indirect illumination only 45

44 Performance Equal-time Path tracing 46

45 Performance Equal-time Path tracing with RASOs (ours) 47

46 Temporal stability TODO: video? 48

47 Homogeneous volume PT PT with RASO PT with RASO (17 min) (3.1 min) (6 s) Diffusion dipole (6 s) multiple-scattering only, discretization: voxel grid, dipole parameters hand-tweaked for visually similar result 49

48 Conclusion Advantages - fast convergence - perceptually unobtrusive error Limitations - precomputation & storage - bias - temporal stability 50

49 Future Work - All-frequency Transport - Application to Clouds - Non-negative matrix factorization 51

50 Thanks for your time! Questions?