Computer graphics processing and selective visual display system – Computer graphics processing – Three-dimension
Reexamination Certificate
1997-09-25
2003-05-20
Padmanabhan, Mano (Department: 2671)
Computer graphics processing and selective visual display system
Computer graphics processing
Three-dimension
C345S589000
Reexamination Certificate
active
06567083
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to computer graphics, and in particular, to shading computer-generated images.
2. Related Art
Shading is an important component that contributes to the realism of a geometric scene to be rendered. The calculations associated with shading can be split into two parts: finding surface reflectance properties at each point being shaded, and computing the incident illumination at the point being shaded. Surface reflectance properties are used by computer graphics application programming interfaces (graphics APIs) to define how a surface dissipates light. For example, one graphics API, OpenGL™ by Silicon Graphics, Inc., supports five surface reflectance properties: diffuse and ambient properties, specular and shininess properties, and an emissive property. See, e.g. R. Fosner,
OpenGL™ Programming for Windows
95
and Windows NT
(Addison-Wesley Pub.: Reading, Mass. 1997), Chapter 9, “Colors, Materials, and Lights,” pp. 181-201.
The shading calculation for an incident light component, however, is fairly primitive in current graphics hardware. As such the overall shading realism is diminished regardless of the sophistication in which surface reflectance properties are represented. In current commercial graphics shaders, incident light sources are only represented by point source lighting models. For example, OpenGL™ only supports directional lights, point light sources with varying distance-based falloff functions, and spot lights, which provide illumination in a cone.
One important type of point light source not yet represented in commercial graphics shaders is a point light source with emission varying by direction, as described by the IES standard luminaire file format. The IES standard luminaire file format is a standard representation of emission from complex emitters, assuming an observer is sufficiently far away that the source can be treated as a point emitter. An IES recommended standard file format for electronic transfer for photometric data, IES LM-63-1990, is available from the Illuminating Engineering Society of North America.
In contrast to point lights, real-world light sources are area sources. Current methods for computing incident light from an area light source are impractical for computer graphics applications and systems. Current techniques are either too inefficient and expensive for hardware or real-time applications, or are only effective for area luminaires with limited shapes or emission distributions. For example, Ashdown has developed a method for mapping the emission distributions from complex area light sources onto a sphere (Ashdown, I.,
J. Illuminating Eng. Soc.
22(1):163-180 (Winter 1993)). Outgoing emission at a number of points on the sphere is computed for each point shaded—the computational requirements of this technique are excessive for current graphics hardware. Stock has developed a similar method, where the directional distribution of a light source is computed at various points over a sphere (Stock, R. D.,
Fourier Formulation of Illumination Optics and Computer
-
Automated Reflector Design,
Ph.D. thesis, CMU (1995)). This distribution is projected into a Fourier space, and the coefficients are interpolated over the source of the sphere to construct intermediate distributions at arbitrary points. However, the techniques of Ashdown and Stock are still impractical for hardware or real-time applications.
Snyder has investigated closed form expressions for incident illumination for various types of area light sources (Snyder, J., “Area light sources for real-time graphics,”
Microsoft Res. Tech. Rep.
MSR-TR-96-11 (March 1996)); this work is based on a paper written by Arvo in SIGGRAPH '95 (Arvo, J., “Applications of irradiance tensors to the simulation of non-Lambertian phenomena,” in
ACM SIGGRAPH '
95
Conf Proc.,
Cook, R., ed., Addison Wesley (August 1995), pp. 335-342). However, Snyder's techniques are still not sufficiently efficient for hardware implementation, and are limited to a small set of types of emitters.
Sun et al. have investigated techniques for tabularizing the calculation of the differential area to area form-factor (Sun, J., et al.,
Computer Graphics Forum
12(4):191-198 (1993)). This computation is equivalent to finding the incident energy from a diffusely emitting light source to a point. This technique reduces computation by tabularizing some aspects of it, but still has significant computational demands and does not easily generalize to more general emitter shapes or non-diffuse emission distributions.
Algorithms based on radiosity have been used to generate images of environments, taking into account both light from area light sources as well as interreflections of light from non-emissive surfaces. For a general introduction to radiosity, see: Francois Sillion and Claude Puech,
Radiosity and Global Illumination,
Morgan Kaufmann Publ: USA, 1994. Radiosity algorithms typically compute a simulation of light transport in a time-consuming pre-processing step, and the result can then be displayed interactively. With radiosity techniques, the lighting solution is computed at a fixed set of points in the environment independent of the viewing parameters used for display. This one set of solution points can never be sufficient to provide an accurate solution for all possible viewing parameters.
Although complete radiosity solutions are impractical for interactive viewing of scenes with area light sources, some of the techniques used in radiosity solutions offer possibilities for computing light from area sources. These techniques are too computationally expensive and too inflexible for general applicability.
One of the techniques commonly used in radiosity algorithms to compute incident light at a point (the form-factor computation) is based on computing a contour integral around the edges of a polygonal light source (Daniel R. Baum and Holly E. Rushmeier and James M. Winget, “Improving Radiosity Solutions Through the Use of Analytically Determined Form-Factors”, SIGGRAPH '89 Proceedings, p. 325-334.). Although this is a potentially exact computation of the incident light, it is currently too expensive to evaluate in an interactive system, and cannot easily be extended to account for lights with non-diffuse emission distributions.
Wallace et al. have proposed a more efficient but less accurate method to compute form-factors, based on a simplifying assumption that the light is disc shaped (John R. Wallace and Kells A. Elmquist and Eric A. Haines, “A Ray Tracing Algorithm for Progressive Radiosity”, SIGGRAPH '89 Proceedings). This is not an exact computation, unless the light is disc shaped and the point where incident light is being computed is parallel to the emitter and directly above or below it. Furthermore, the computation is still not efficient enough for real-time applications and cannot be easily extended to account for lights with varying emission distributions.
Ward has developed algorithms for global illumination and area light source computation based on computing scalar irradiance values at points on surfaces and interpolating them to compute irradiance at nearby points. See, Ward, Gregory J., et al., “A Ray Tracing Solution for Diffuse Interreflection”, in
Computer Graphics
(
SIGGRAPH '
88
Proc.
); Ward, Gregory J. and Heckbert, P., “Irradiance Gradients” in Third Eurographics Workshop on Rendering, May 1992; Ward, Gregory J., “The RADIANCE Lighting Simulation and Rendering System” in
SIGGRAPH '
94
Proc.,
Orlando, Fla., Jul. 24-29, 1994. Although these algorithms produce high-quality lighting simulations and naturally adapt to the rate at which irradiance is changing in a scene, they depend on storing irradiance estimates in a spatial data structure that is too complex for interactive applications, and the irradiance estimates are combined and weighted by a function that is also too complex for interactive applications. This algorithm doesn't store irradiance in a vector form, and
Baum Daniel R.
Hanrahan Patrick M.
Pharr Matthew M.
Microsoft Corporation
Padmanabhan Mano
Woodcock & Washburn LLP
LandOfFree
Method, system, and computer program product for providing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method, system, and computer program product for providing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method, system, and computer program product for providing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3002277