Presentations text content in A Hierarchical Volumetric Shadow
Slide1
A Hierarchical Volumetric ShadowAlgorithm for Single Scattering
Ilya Baran, Jiawen Chen, Jonathan Ragan-Kelley,Frédo Durand, Jaakko LehtinenComputer Science and Artificial Intelligence LaboratoryMassachusetts Institute of Technology
Slide2Volumetric scattering with shadows
Photo by Frédo Durand
Slide3Alan Wake
by Remedy Entertainment
Slide4Slide5Slide6Slide7Slide8Related work
Ray marching(brute force)Analytical scattering modelsSky lighting, bloom, etc.Doesn’t account for visibility
Sun et al. [2005]
Slide9Shadow volume methods
Max [1986]Analytical integrationWyman and Ramsey [2008]Ray marching along intervals
Slide10Related work
Engelhardt and Dachsbacher [2010]Detect discontinuities, subsample and interpolatePerformance depends on occluder complexityEpipolar geometry
Detected
discontinuties
Engelhardt
and
Dachsbacher
[2010]
Slide11Overview
Incremental integrationApproximating single scatteringEpipolar rectification
~
Slide12Simplified scenario
Orthographic camera
Light direction perpendicular to view direction
Visibility only
Slide13r
d
Slide14Slide157
5
Brute force complexity:
O(
rd
)
1
Slide167
5
5
1
Brute force complexity:
O(
rd
)
Slide171
1
1
1
1
1
1
Bit mask
Process top down incrementally
Slide187
1
1
1
1
1
1
1
Bit mask
Slide197
1
1
1
1
1
1
1
1
Bit mask
Slide207
1
1
1
1
1
1
1
1
Bit mask
Slide217
1
1
0
0
1
1
1
1
Bit mask
Slide227
1
5
1
0
0
1
1
1
1
Bit mask
Slide237
1
5
Bit mask algorithm complexity:
O(
rd
)
5
2
2
3
1
0
0
1
1
1
1
Bit mask
Slide24Partial sum trees
Binary treeEach node storessum of children
4
2
2
2
0
1
1
1
1
0
0
1
0
1
0
Slide25Tree update
4
2
2
2
0
1
1
1
1
0
0
0
0
1
0
Slide26Tree update
4
2
2
2
0
0
1
1
1
0
0
0
0
1
0
Slide27Tree update
4
2
1
2
0
0
1
1
1
0
0
0
0
1
0
Slide28Tree update
3
2
1
2
0
0
1
1
1
0
0
0
0
1
0
Slide29Tree query
3
2
1
2
0
0
1
1
1
0
0
0
0
1
0
∑ = 3
Slide30Complexity
2D
Brute force:
O(
rd
)
Incremental with bit mask:
O(
rd
)
Incremental with tree:
O( (
r+d
) log d )
3D:
s
independent
slices
Brute force:
O(
srd
)
Incremental with tree:
O(
s (
r+d
) log d )
Slide31Textured lights
0.8
0.1
0.5
0.3
0.2
1
0.8
Light texture
1
0.9
0.8
0.6
0.5
0.4
0.3
Light attenuation
Slide32SVD approximation
~
A
U
SV
T
=
+
+
+
Slide33Epipolar
rectification
Light direction
Epipolar
slices
Eye
Slide34Epipolar coordinates
Point light
Directional light
Slide35Depth map resampling
Camera depth map
Light depth map
Rectify
Rectify
Corresponding
epipolar
slice
Slide36
9.9 sec (17.1x)
169 sec
Equal quality comparison:
Sibenik
Our method
Ray marching
Slide37Our method
Ray marching
Equal quality comparison: Terrain
1.3 sec (41.5x)
54 sec
Slide38Equal quality comparisons: Trees
Our method
Ray marching
2.3 sec (120.4x)
277 sec
Slide39Slide40GPU Performance on
Sibenik
Slide41GPU Performance on Trees
Slide42Limitations and future work
Non-homogeneous media
All points in volume must be visited
SVD will have high rank
GPU performance
Dynamic data structure: limited parallelism
Bandwidth intensive
Requires CUDA, not suitable for consoles
Slide43Summary
Hierarchical volumetric shadow algorithm with complexity guaranteesSignificant speedupson CPUsModerately faster than state-of-the art on GPUs
Slide44GPU Performance on Sibenik
Slide45Slide46Slide47
A Hierarchical Volumetric Shadow
Download Presentation - The PPT/PDF document "A Hierarchical Volumetric Shadow" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.