Photon Mapping Combines backward/ reverse ray tracing with - - PDF document
Photon Mapping Photon Mapping Combines backward/ reverse ray tracing with stochastic ray tracing Used to simulate the interaction of light with a variety transparent substances (caustics) Glass Water Diffuse
Combines “backward”/ “reverse” ray tracing
Used to simulate the interaction of light with
Glass Water Diffuse Inter-reflections between illuminated
Effects of particulate matter
Smoke Water vapor
Henrik Wann Jensen 95/96 Simulates the transport of individual photons
Photons bounce off specular surfaces Photons deposited on diffuse surfaces Photons collected by ray tracing from eye http://www.ypoart.com/
The collected photons will need to be
Spatial subdivision algorithms subdivide
Subdivide your scene volume into hierarchical regions
Octrees BSP Trees K-D Trees
Create a tree structure that indicates for each region:
if the region is empty the object present at that particular region
Used in
Ray Tracing Collision detection (animation) Photon Mapping
Subdivision algorithm
Recursively subdivide volume If subregion is empty or contains a predefined #
Otherwise, further subdivide.
Continue until each subregion is empty or
Use tree to represent subdivision hierarchy
Volume is subdivided into 4 equal
Each iteration results in 8 subdivisions.
[Foley/Van Dam]
Subdivisions represented as a tree [Watt/Watt,244]
Quadtree applet
Images from http://www.flipcode.com/tutorials/tut_octrees.shtml
Binary Space Partitioning Trees (BSP
Like Octrees but divides space into a pair
Subregions need not be equally spaced Planes separating regions can be placed at
Octrees vs BSP Trees [Watt/Watt,246]
BSP Trees – Each non-terminal node represents a single
[Watt/Watt,247 ]
K-d Tree is a BSP where dividing planes
K-d Tree applet
Advantages
Efficient means for finding objects within your
Compact representation
Disadvantages
Preprocessing required If scene changes, must rebuild your tree Not foolproof
Octrees vs BSP trees
BSP Trees are generally more balanced
BSP Traversal more efficient BSP Trees have additional storage
Now back to photon mapping
Pass 1 – Shoot and Store Photons
photons are shot from the light into the
Photons are allowed to interact with
Where photons fall are stored in a special
1000s of photons not billions
Statistical approximation based on density
Shooting photons
Photon Scattering
The photon Placed in k-d tree for efficient access
Pass 2 – Gather illumination
Use ray tracing Direct illumination determined by ray
Indirect illumination determined by
Gather photon within a volume in the required
Driven by chance / randomness Rendering and Monte Carlo
Multiple rays per pixel
See how many reach a light
Multiple rays from light
See how many reach the eye
Photon Mapping
Goes from both sides
Pattern of light focused on a surface after
[Jensen]
Three separate photon maps
Global – photon emitted toward all objects Caustics – specular to diffuse interactions Volume – for volumetric effects
2000ET
But doesn’t this take a long time.
No, grasshopper…not with GPUs… Let’s go to the video.
Fixing ray tracing
Spawning complex structures (cones,
Stochastic sampling (distributed ray
“backwards” ray tracing (Photon Mapping)
Questions