JP2003296750A - Graphic image rendering using radiance autotransfer of low frequency lighting environment - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently represent a low frequency lighting environment without causing aliasing by using a low dimensional spherical surface harmonics (SH) base. <P>SOLUTION: Real time image rendering of a diffusion object and a gloss object in a low frequency lighting environment fetches blurry shadow, interreflection, and caustic. A global propagation simulator represents incident lighting transfer of an optional low frequency to radiance to be emitted as preprocessing, but generates a function including a global effect such as shadowing from an object to the object itself and interreflection over the object surface. At execution, those transfer functions are applied to actual incident lighting. Dynamic local lighting is processed by performing sampling near the object in each frame. The object can be turned as a carbody shell about lighting and the lighting can be turned about the object. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to computer graphic image rendering techniques, and more particularly to lighting and shadowing of modeled objects in rendered images.

[0002]

BACKGROUND OF THE INVENTION A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the reproduction by anyone of this patent disclosure in the U.S. Patent and Trademark Office patent file or record, but otherwise reserves all copyright rights whatsoever.

Lighting from surface light sources, soft shadows, and interreflections are important effects for realistic image synthesis. Unfortunately, common methods for integration over large lighting environments, including Monte Carlo raytracing, radiosity, or multipass rendering summing over multiple point sources, are impractical for real-time rendering.
(There are representative documents regarding Monte Carlo ray tracing (for example, see Non-Patent Document 1, Non-Patent Document 2, and Non-Patent Document 3). There is a representative document regarding the multi-pass rendering technology of (for example, see Non-Patent Document 5, Non-Patent Document 6, and Non-Patent Document 7)

[0004]

[Non-Patent Document 1] Cook, R. Porter, T and Carpenter,
L, Distributed Ray Tracing, SIGGRAPH'84, 137〜146

[0005]

[Non-Patent Document 2] Jensen, H, Global Illumination usi
ng Photon Maps, EurographicsWorkshop on Rendering
1996, 21-30

[0006]

[Non-Patent Document 3] Kajiya, J, The Rendering Equation,
SIGGRAPH'86, 143-150

[0007]

[Non-Patent Document 4] Cohen, M and Wallace, J, Radiosity
and Realistic Image Synthesis, Academic Press Pro
fessional, Cambridge, 1993

[0008]

[Non-Patent Document 5] Haeberli, P and Akeley, K, The Acc
umulation Buffer: Hardware Support for High-Qualit
y Rendering, SIGGRAPH'90, 309 ~ 318

[0009]

[Non-Patent Document 6] Keller, A, Instant Radiosity, SIGG
RAPH'97, 49 ~ 56

[0010]

[Non-Patent Document 7] Segal, M, Korobkin, C, van Widenfe
lt, R, Foran, J and Haeberli, P, Fast Shadows and
Lighting Effects Using Texture Mapping, SIGGRAPH'9
2, 249 to 252

[0011]

[Non-Patent Document 8] Cabral, B, Olano, M and Nemec, P,
Reflection Space Image Based Rendering, SIGGRAPH'9
9, 165-170

[0012]

[Non-Patent Document 9] Greene, N, Environment Mapping and
Other applications of WorldProjections, IEEE CG &
A, 6 (11): 21-29, 1986

[0013]

[Non-Patent Document 10] Heidrich, W, Seidel H, Realisti
c, Hardware-Accelerated Shadingand Lighting, SIGGR
APH'99, 171-178

[0014]

[Non-Patent Document 11] Kautz, J, Vazquez, P, Heidrich,
W and Seidel, H, A Unified Approach to Pre-filtere
d Environment Maps, Eurographics Workshop on Rende
ring 2000, 185-196

[0015]

[Non-Patent Document 12] Ramamoorthi, R and Hanrahan, P,
An Efficient Representation for Irradiance Environ
ment Maps, SIGGRAPH'01, 497 ~ 500

[0016]

[Non-Patent Document 13] Edmonds, A, Angular Momentum in
Quantum Mechanics, Princeton University, Princeto
n, 1960

[0017]

[Non-Patent Document 14] Zare, R, Angular Momentum: Under
standing Spatial Aspects in Chemistry and Physics,
Wiley, New York, 1987

[0018]

[Non-Patent Document 15] Chirikjian, G and Stein, D, Kine
matic Design and Communicationof a Spherical Stepp
er Motor, IEEE Transactions on Mechatronics, 4 (4),
Dec. 1999

[0019]

[Non-Patent Document 16] Linde, Y, Buzo and Gray, R, An a
lgorithm for Vector QuantizerDesign, IEEE Transact
ion on Communication COM-28, 1980, 84 ~ 95

[0020]

[Non-Patent Document 17] Greger, G., Shirley, P, Hubbard,
P, and Greenberg, D, The Irrandiance Volume, IEEE
Computer Graphics and Applications, 6 (11): 21 ~ 2
9,1986

[0021]

[Problems to be Solved by the Invention] There are three difficulties involved in real-time realistic global illumination. That is, a complicated spatially varying bidirectional reflection distribution function (BRDF) of a real material has to be modeled (complexity of BRDF), and an integration (light integration) over a hemisphere in the lighting direction at each point is performed. It is necessary, and bouncing / occlusion effects such as shadows due to inclusions in the light path from the light source to the receiver must be considered (complexity of light propagation). Many studies have focused on solving the problem of light integration by expanding the complexity of BRDF (eg, shiny anisotropic reflection) and expressing the incident lighting as a sum of directions or points. It was Therefore, integrating light results in the same as simply sampling an analytical or tabular BRDF at a few points, but is cumbersome for large light sources. In the study of the second system, radiance is sampled, and the radiance is convoluted with kernels of various sizes in advance. (For example, refer to the following (for example, see Non-Patent Document 8, Non-Patent Document 9, Non-Patent Document 10, Non-Patent Document 11, and Non-Patent Document 12))
Although this method solves the problem of light integration, it neglects the complexity of light propagation such as shadows, since the convolution assumes that the incident reflection brightness is not shielded and does not scatter. And there are subtle ways to increase the complexity of light propagation, especially shadows. The problem is the integration of light, and most of the above methods are not suitable for very large light sources.

[0022]

The real-time image rendering techniques described herein better account for the complexity of light integration and light propagation in real time. This technique targets low-frequency lighting environments (area-supported basis functions) and uses low-order spherical harmonic (SH) bases such that aliasing does not occur. Efficiently represent a diverse environment. One aspect of this technique is to represent how an object scatters its light into the object itself or into adjacent space in a way that separates that scatter from the incident lighting. For example, FIG. 8 compares an unshadowed image of a modeled human head with an image generated as described herein using radiance self-transfer.

Briefly describing the techniques described herein, assume first that there is a convex diffuse object illuminated by an infinitely far environment map. The shaded “reaction” of an object to its environment can be thought of as the transfer function that maps the incoming radiance to the outgoing radiance, which in some cases simply does a cosine weighted integration. More complex integrals incorporate how a concave object creates a shadow on itself, in this case multiplying the integrand by an additional propagation factor that represents the visibility in each direction.

The approach of the technique described herein precomputes the heavy-load propagation simulation required by a complex transfer function such as shadowing for a given object. The resulting transfer function is represented as a dense set of vectors or matrices on the object surface. On the other hand, the incident radiance does not have to be calculated in advance. During subsequent real-time rendering, the graphics hardware can dynamically sample the incident radiance at a limited number of points, which the technique transforms into a spherical harmonic (SH) basis with fast computation. . Analytical models such as skylight models or simple geometric shapes such as circles can also be used.

By representing both the incident radiance and the transfer function in the SH basis, the technique described here allows the integral of the light to be calculated as a simple dot product (diffusion object) or a small transfer between their coefficient vectors. It is a simple linear transformation of a lighting coefficient vector through a matrix (gloss target). Very low coefficients (9-25) required in low frequency lighting environment, 1 for graphics hardware
The result can be calculated in a single pass. Unlike Monte Carlo and multi-pass optical integration methods, run-time computations with this technique can be kept constant no matter how many sources or large sources. Retaining a certain amount of run-time computation relies on large, smooth lighting to limit the number of SH coefficients required.

The techniques described herein represent in the transfer function complex propagating effects such as interreflection and caustic. Since these are all simulated as preprocessing, only the basis coefficient of the transfer function changes, and the calculation at runtime does not change. The approach of the technique described here deals with both surface and volume-based geometry. With a high SH factor, this technique can handle not only diffuse, but glossy (but highly specular) objects, including interreflections. For example, 25 SH factors are sufficient to obtain a useful glossy effect.
In addition to the transfer from a rigid object to itself, called self-transfer, this technique is generalized to the proximity transfer from a rigid object to the space adjacent to it, with a cast soft shadow, glossy reflections, And allows caustics on moving objects.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Referring to FIG. 1, the software architecture of a computer graphic image rendering system 100 uses the radiance self-transmitting image rendering technique described herein to provide inter-reflection and self-shadowing. -Real-time image rendering of modeling objects with shadowing). In general, this software architecture is based on the Global Illumination Simulator 120,
It includes a real-time image rendering engine 140 and a graphic display driver 180. In the radiance self-transfer rendering technique, which is described more fully below, the Global Illumination Simulator 12
0 performs the pre-processing stage of this technique, where the radiance self-transfer data 13 from the geometric object model 110 is
0 is calculated in advance. The geometric object model 110 may be a triangular mesh, wavelet composition, or any other representation of the geometry of the object to be modeled. The real-time image rendering engine 140 then uses the pre-computed radiance self-transfer data to dynamically change the lighting environment 150.
And render the image of the object modeled for the view direction 160 and
0 and view direction 160 can be selectively changed or set with user controls 170. The graphic display driver 180 displays an image on an image output device (eg, monitor, projector, printer, etc.).
Output to.

In some embodiments of the graphic image rendering system, the pre-computation of radiance self-transfer by the simulator 120 and the image rendering by the engine 140 are combined into a single image as described in Section 6 "Computing Environment" below. It can be implemented on a computer. More generally, the simulator 120 can be run on another computer and the resulting data can be transferred to a computer that runs the rendering engine 140 to produce a graphic image.

Radiance Self-Transmission Techniques Overview: As a pre-processing, the Global Illumination Simulator 120 performs a lighting simulation on the model 110, which allows the modeled object to shadow itself and scatter light. Take in. The results are recorded in the radiance self-transfer data 130 as a dense set of vectors (for diffusion) or matrices (for gloss) on the model. When rendering the image,
The rendering engine 140 uses the lighting environment 15
Project the incident radiance corresponding to 0 into a spherical harmonic (SH) basis (this is described below). The model's transfer vector or matrix field is then applied to the lighting's coefficient vector. If the object is diffusive, the dot product of the transfer vector and the lighting coefficient at multiple points on the object is determined to accurately generate the self-scattering shading.
If the object is shiny, then a transfer matrix is applied to the lighting coefficients to obtain the coefficients of the spherical function representing the self-scattered incident radiance at those points. This function is convolved with the bidirectional reflectance distribution function (BRDF) of the object and then evaluated with the view dependent reflection direction to produce the final shading of the object in the rendered image.

1. Spherical harmonic
General definition of: Spherical harmonics defines an orthonormal basis for the sphere S, similar to the Fourier transform for a one-dimensional circle.
Parameterization s = (x, y, z) = (sin θ cos φ, sin θ s
in φ, cos θ), the basis function is

[0031]

[Equation 1]

Is defined as P _l ^m is a companion Legendre polynomial and K _l ^m is a normalization constant.

[0033]

[Equation 2]

The above definition forms a complex basis, a real-valued basis is a simple transformation.

[0035]

[Equation 3]

Is given by

A low value of l (called the band index) represents a low frequency basis function on the sphere. The basis function for band l will be a polynomial of order l in x, y, and z. The evaluation can be performed by a simple recurrence formula (for example, see Non-Patent Document 13 and Non-Patent Document 14).

Projection and reconstruction: Since the SH basis is orthonormal, the integral (area-supported lighting basis function (area-suppo
rted basis function))

[0039]

[Equation 4]

Via, the scalar function f defined on S can be projected onto its coefficients.

The coefficients are the reconstruction functions of order n.

[0042]

[Equation 5]

Which provides a greater degree of approximation to f as the number of bands n increases. The low frequency signal is
It can be accurately represented by only a few SH bands. Higher frequency signals are band limited (ie smooth without aliasing) by low order projections.

Projecting to the nth order requires n ² coefficients. Often, for vectors of projection coefficients and basis functions indexed one by one,

[0045]

[Equation 6]

It is convenient to rewrite equation (2) via, where i = 1 (l + 1) + m + 1. This formulation makes it clear that the evaluation of the reconstruction function at s represents a simple dot product of the n ² -component coefficient vector f _i and the evaluated basis function y _l (s) vector.

Basic Property: One of the properties of SH projection is its rotation invariance. That is, if g (s) = f (Q (s)) when Q is an arbitrary rotation on S,

[0048]

[Equation 7]

It becomes

This is similar to the transition-invariant property of the one-dimensional Fourier transform. In fact, this property collects, projects, and again samples from f into a rotated set of sample points.

[0051]

[Equation 8]

It means that SH projection does not cause aliasing artifacts when rotating.

Since the SH basis is orthonormal, S
Given any two functions a and b for, their projections are

[0054]

[Equation 9]

The useful property of satisfying is obtained. That is, the integral of the product of the band-limited functions becomes the dot product of the projection coefficients of those functions.

Convolution: The convolution of the function f with the circularly symmetric kernel function h (z) is expressed as h ^* f. H must be circularly symmetric (and thus can be defined as a simple function of z rather than s) so that equation (3) above defines the result in S rather than in the higher order rotation group SO. ). The projection of this convolution is

[0057]

[Equation 10]

Satisfies. In other words, the coefficients of the projected convolution are simply the scaled product of the individually projected functions. This property is h (z) = max (z, 0)
It provides a high-speed method for convolving a hemispherical cosine nucleus defined as to an environment map to obtain an irradiance map (see Non-Patent Document 12, for example). In this case, the projection coefficient of h can be obtained using an analytical formula. The nature of the convolution can also be used to obtain a narrower pre-filtered environment map. Note that h is circularly symmetric about z, so its projection coefficient is nonzero only when m = 0.

Product projection: The projection of the product of the pair of spherical functions c (s) = a (s) b (s) when a is known and b is not known is a matrix

[0060]

[Equation 11]

[0061]

[Equation 12]

It can be thought of as a linear transformation of the projection coefficient b _j through, the summation being indicated in the duplicated j index.

[0063]

[Equation 13]

Note that is a symmetric matrix.

Rotation: Q,

[0066]

[Equation 14]

The reconstruction function rotated by can be projected using a linear transformation of the projection coefficients f _i of f.
Due to its rotation-invariant nature, this linear transform treats the coefficients in each band individually. The most efficient implementation is achieved using the zyz-Euler angle decomposition of the rotation Q using fairly complex recurrence formulas (eg, [13], [1]).
4, Non-Patent Document 15). When dealing with low-order functions only,
Symbolic integration can be used to implement the explicit rotation equations described in these references.

2. Radiance Self-Transmission Radiance self-transmission summarizes how an object O shadows itself and scatters light.
The radiance self-transfer is calculated using the SH basis as L
_It is represented by first parameterizing the incident lighting at the point pεO denoted _p (s). Therefore, the incident lighting is represented as a vector (L _p ) i of n ² coefficients. In practice, this lighting can be dynamically and sparsely sampled near the surface, perhaps with only a single point. This assumes small variations in lighting on O that are not due to the presence of O itself (see Section 4.1 "Spatial Sampling of the Incident Radiance Field"). The radiance self-transfer may be pre-computed as a transfer vector or matrix and stored densely on O.

The transfer vector (M _p ) _i can be used for the diffusive surface and represents a linear transformation on the lighting vector, producing a scalar exit radiance denoted L ′ _p via the dot product.

[0070]

[Equation 15]

That is, each component of (M _p ) _i represents a linear effect of the lighting basis function (L _p ) _i on shading at p.

The transfer matrix (M _p ) _y can be used for glossy surfaces and represents a linear transformation on the lighting vector, where the transformation of the entire spherical function of the transmitted radiance L ′ _p (s) is not a scalar. The projection coefficient is obtained.
Ie

[0073]

[Equation 16]

Of incident and transmitted radiance
The difference is L ’_p(S) is shadowy due to the presence of O
Including the effects of scattering and scattering, while L_p(S) is O
Represents the incident lighting assuming it has been removed from the scene
It is a point. (M_p)_yIs the transmitted radiance
L ’_pIncident radiance (L) for the i-th coefficient of (s)
_p) Show the linear effect of the jth lighting coefficient of j
You In the next section, due to self-scattering on O,
What is the transfer vector of the diffusive surface and the transfer matrix of the glossy surface?
Describe how to ask.

2.1 Diffuse Transmission First, assume that O is diffusive. The simplest transfer function at the point pεO is a scalar function

[0076]

[Equation 17]

Represents the unshadowed diffuse transfer, defined as, which results in an exit radiance that does not change with the view angle with respect to the diffuse surface. Where ρ
_p is the albedo of the object at _p , L _p is the incident radiance at p assuming O has been removed from the scene, N _p is the normal of the object at p, and H _Np (s) = max (N _p [k, 0) is N _p
Is a hemispherical nucleus with a cosine weight centered on. L
By separately projecting _p and H _Np by SH,
Then T _DU is the dot product of those coefficient vectors. The factors obtained by this are called the light function L _p and the transfer function M _p here. In this first simple case,

[0078]

[Equation 18]

It becomes

Since N _p is known, the transfer function

[0081]

[Formula 19]

The SH projection of can be calculated in advance,
This results in a transfer vector. In fact, if N _p is given, the transfer vector can be obtained by a simple analytic formula, so that no memory is required.

[0083]

[Equation 20]

Since is essentially a low-pass filter, the quadratic projection (9 coefficients) provides good accuracy in any lighting environment (even if not smooth).

Diffuse transfer with shadows to include shadows

[0086]

[Equation 21]

And an additional visibility function V _p (s)
→ {0,1} is equal to 1 when the ray from p in direction s does not intersect O again (ie cannot shadow). Similar to the unshadowed transfer, this integral is the SH projection of L _p and the transfer function

[0088]

[Equation 22]

Can be decomposed into two functions. By again SH projecting L _p and M _p separately, the integral in T _DS becomes the dot product of the coefficient vectors.

In this case, communication is important. The radiance self-transfer is pre-computed using a propagation simulator (discussed in the "Pre-calculation of Self-Transfer" section below) and the resulting transfer vector (M _p ) _i
Can be stored at many points p on O. Unlike the previous case, V _p may locally generate high frequency lighting, eg by shadowing itself in a “pinhole”.

[0091]

[Equation 23]

The quadratic projection of can be inaccurate even in a smooth lighting environment. Using 4th or 5th order projections gives good results with regular meshes in a smooth lighting environment. For example, compare the unshadowed and shadowed images of the diffuse surface object of FIG.

Finally, in order to capture not only shadows but also diffuse interreflections (see, for example, the image with interreflections in FIG. 9), the inter-reflected diffuse transmission is taken into account.

[0094]

[Equation 24]

Can be defined as

[0096]

[Equation 25]

Is the radiance from O itself in direction s. The problem is

[0098]

[Equation 26]

Is determined by the outgoing radiance at a point at an arbitrary distance from p, and because the local lighting fluctuates on O, unless the incident radiance is emitted from a light source at infinity, the incident radiance at p alone Radiance even if given brightness

[0100]

[Equation 27]

Is that I do not understand. If the variation of the lighting on O is small, it is as if O is illuminated by L _p everywhere.

[0102]

[Equation 28]

It is very close to Therefore, T _DI depends linearly on L _p and can be factored into a product of two projected functions, a light dependent function and a geometry dependent function, as in the previous two cases.

Pre-computed interreflections must assume that the incident lighting on O does not change spatially, but for simpler shadowed transmission this assumption is not necessary. The difference is that the shadowed transfer depends only on the incident lighting at p, while the inter-reflected transfer depends on many points q ≠ p on O where L _q ≠ L _p . Therefore, if the incident radiance field is sampled sufficiently finely (discussed in the "Spatial Sampling of the Incident Radiance Field" section below), local lighting variations can be captured and shadowed. The transmission will be accurate.

[0105]

[Equation 29]

In the presence of, the transfer function of interreflection

[0107]

[Equation 30]

It becomes difficult to express explicitly. The section "Precalculation of Self-Transmission" below describes how to calculate its projection coefficients numerically.

2.2 Gloss transfer Self transfer of a glossy object can be defined in a similar way, but generalizes the kernel function to depend on the reflection direction R rather than the (fixed) normal N (view dependent). Do). Similar to the H nucleus above, the radiance self-transfer can model the glossy reflection as the nucleus G (s, R, r), and the scalar r is the "gloss" or extent of the specular response. Define.

Three similar gloss transfer functions, with and without shadowing, and with shadowing, are given below.

[0111]

[Equation 31]

Can be defined as
L is an unknown quantity when calculating _pAnd as a function of R
And outputs a scalar radiance in the direction R. Transmission is no longer
Since it is not just a function of s, a simple vector of SH coefficients
Cannot be replaced with.

Instead of parameterizing the scalar transfer with R and r, a more useful decomposition is the incident radiance L
_p (s) is transmitted over the entire spherical surface of the transmitted radiance represented by L' _p (s). Assuming that the gloss nucleus G is circularly symmetric with respect to R, G'in L' _p (s)
convolve _r (z) = G (s, (0,0,1), r),
It can be evaluated with R to get the final result.

We now represent the transfer to L' _p as a matrix rather than a vector. For example, a glossy shadowed transmission

[0115]

[Equation 32]

Can be defined as, which is its SH
The projection is a symmetric matrix via Eq. (7)

[0117]

[Expression 33]

Is a linear operator for L _p that can be expressed as Even with very smooth lighting, more SH bands are added as the glossiness of O increases.

[0119]

[Equation 34]

For such conditions, a non-square matrix (eg 25 × 9) useful for mapping low frequency lighting to higher frequency transfer radiance is useful under such conditions.

Important in limiting the pre-calculated radiance self-transfer is that the material properties of O (such as albedo and gloss) that affect _{interreflection} in T _DI and T _GI are "burned" in the pre-processed transfer. "Baked in", a point that cannot be changed at run time. In contrast, simpler shadowed transmissions without interreflections allow run-time modification and / or spatial variation of material properties on O. An error occurs when an obstacle or light source enters the O convex hull. O can only move as a rigid body, and cannot deform or move a component as a whole. Also, recall the premise that accurate variation in lighting on O is required for accurate interreflection.

Finally, the diffuse transfer defined above produces a radiance after leaving the surface since the convolution of the cosine-weighted normal hemisphere already produces the radiance, whereas the gloss transfer does not. Produces a radiance incident on a local B
Note that the RDF is convolved to get the final exit radiance. In addition, BR which is fixed to glossy O
It is also possible to burn in the DF and eliminate the G convolution at runtime, but with limited flexibility.

FIG. 9 shows a non-shadowed image, a shadowed image, and an inter-reflected image of a glossy surface object.

3. Pre-Calculation of Self-Transmission Referring now to FIG. 2, the illumination simulator 120 (FIG. 1) uses the SH basis for an infinite sphere (“reference lighting environment” for the simulation) as an emitter, the object O In the global illumination simulation 200 performed in step 1, the self-transfer of the radiance of the modeled object is calculated in advance (also referred to as “pre-calculation of self-transfer”).
This simulation is based on the nth-order SH projection of the unknown spherical surface of the incident light L, that is, n ² unknown coefficients L _i
Parameterize by. The simulation result is
Although the SH basis functions y _i (s) can be used as radiators and can be calculated individually for each L _i , it is more efficient to calculate them all at once. The reference lighting environment (sphere L at infinity) is replaced at runtime with the actual incident radiance L _p around O.

The simulation 200 starts with a path ("shadow path" 202) that simulates a direct shadow from the path issuing L and directly reaching the sample point pεO. Subsequent passes (“inter-reflective passes” 204) add inter-reflective and represent paths (Lp, LDp, LDDp, etc.) from L that bounced off O multiple times before reaching p. In each pass, energy is collected at every sample point on the p surface. For large radiators (ie, low frequency SH basis), energy aggregation is more efficient than shooting updates.

In order to capture the sphere in the direction at the sample point pεO, the simulation is large (10 k
30k) quasi-random direction set

[0127]

[Equation 35]

Is generated. The simulation also pre-computes the evaluation of all SH basis functions in each direction s _d . The direction s _d is organized into hierarchical bins formed by refining the first icosahedron in 1 → 2 halves into spherical triangles of equal area (on a sphere, 1 → 4 subdivisions of a plane A simulation cannot use equal area triangles as in the case) and uses 6-8 subdivision levels to generate 512-2048 bins. Every bin at each level of the hierarchy contains a list of s _d .

First pass or shadow pass 202
Then, the simulation 200 shows that for each pεO, self-occlusion occurs by the object (se
lf-occlude) Tag in the direction from point p (21
1). The simulation uses hierarchies to cull directions that are outside the hemisphere and project shadow rays into the hemisphere centered about the normal N _p of _p . The simulation tags each direction s _d with occlusion bits, ie 1-V _p (s _d ), and sees if s _d is in the hemisphere and intersects O again (ie O can shadow itself). Show. The occlusion bit is also associated with a hierarchical bin and indicates whether there is an occlusion-occurring s _d within the bin. Directions and bins with self-occlusion are tagged so that they can undergo further inter-reflection paths, bins / samples without any occlusion receive only direct light from the environment.

Then, the simulation 200 is performed at 212.
Integrates the transmitted radiance at point p. For diffusive surfaces, the simulation further computes the transfer vector by SH projection of M _p in equation (10) for each point pεO.
For glossy surfaces, the simulation computes the transfer matrix by SH projection of M _{p in} equation (11). In each case, the results represent the radiance collected in p and parameterized in L. The SH projection to compute the transfer is done by numerical integration on the direction samples s _d , the sum of which is the transfer accumulated using the following rules.

[0131]

[Table 1]

The superscript 0 represents the number of repetitions. The vector M _p or matrix M _{p at} each point p is initialized to 0 before the shadow pass, where the shadow pass sums all s _d at all p. This rule is based on the equation (1) of diffusion transfer integral.
And gloss transfer integral (7).

Referring also to FIG. 7, a subsequent interreflection path 20.
At 4, the bin with the occlusion bit set is traversed in the shadow pass, and at 221 the inter-reflection transfer is integrated. Instead of a shadow ray, the simulation impinges on O a ray that returns the transmission from the exit illumination.
If the ray (p, s _d ) intersects another point qεO (q is closest to p), sample the radiance exiting from q in the direction −s _d . The following update rule is used, in which the subscript b is the number of rebound pass iterations.

[0134]

[Table 2]

Similar to shadow path 202, interreflection path 204 begins by initializing the transfer vector or matrix to 0 before accumulating the transfer over direction s _d . The diffusion rule is obtained from the definition of T _DI and equation (1), and the gloss rule is obtained from the T _GI rule and equations (6) and (7). The middle factor in the definition of gloss transfer represents the radiance emanating from q of the previous bounce path b-1 and returning to p. Since M _q stores the incident radiance, convolve it with the BRDF of O at q

[0136]

[Equation 36]

Obtain the outgoing radiance in the direction of, over k
You have to get the sum. a_kIndex one by one
K-th convolution factor represented by the notation
I want you to remember. "Reflect" operation
The lator simply passes its first vector argument to its second
Reflect the vector argument of. Equation (7) is (M
_p)_yIs formed by the product of two spherical functions,
It becomes a symmetric matrix of glossy transfer with shadow
Note that it means. This is the inter-reflected light
It does not apply to rich communication.

As shown at 230, the interreflection path 204
Is repeated until a termination condition is met, such as the total energy of a given path falling below a critical threshold. Interreflections decrease fairly rapidly with standard materials. The sum of the transmissions from all the bounce paths corresponds to interreflection.

Alternatively, a simple enhancement to this simulation enables a mirror-like surface in O. The simulation does not record the transmission at such surfaces. Instead, light rays striking the mirror-like surface are always reflected and propagate until they reach the non-mirror-like surface. Therefore, the path in successive iterations is (L
[S] ^* p, L [S] ^* D [S] ^* p, L [S] ^* D [S]
^* D [S] ^* p, etc., where D is the diffusive or glossy bounce and S is the mirror bounce. This allows caustic (causti) for diffusive or glossy objects that respond dynamically to changes in lighting.
c) is captured.

4. Run-time Rendering of Radiance Transfer The simulation 200 (FIG. 2) described above provides a model that incorporates the transfer of radiance at a number of points p on the object surface represented as a vector or matrix. Referring now to FIG. 3, the real-time image rendering engine 140 (FIG. 1) uses this model in the runtime procedure 300 to render the object's self-communication in the selected lighting environment and view direction during real-time image rendering of the object. Calculate the radiance. 4 and 5 show the process flow of the run-time procedure 300 for diffuse and glossy surfaces, respectively.

In step 300, the rendering engine 14
0 performs the following operation at the time of execution. At 310, one close to O
Or multiple sample points P_iIncident lighting at
{L _pi} For the SH basis. 320 at that
L_PiRotate to fit the O coordinate frame and blend them
(See below) Incident lighting L on O_pThe fi
Get the field. 330-390, at each point p on O (L
_p)_iThe output radiance is obtained by performing a linear transformation on. This line
The shape conversion is based on the diffusion surface of the action 350 (equation (8)).
(Mp) i dot product or action 360
(M on the glossy surface (formula (9))_p)_ijMatrix vector of and
It is multiplication. Further action 370 and glossy surface
And 380, where the release obtained by multiplication 360
Convolve the BRDF of O at p with the radiance vector,
Evaluate in the view-dependent reflection direction R.

For action 310, rendering engine 140 loads a precomputed environment map, evaluates an analytical lighting model in software, or graphics hardware as a representation of the incident lighting of the selected lighting environment. Can be used to sample the radiance. The rotation of action 320 is outlined in Section 1, “Spherical Harmonics Overview”, which is done once per object, not per point p. This rotation is done because the transmission is stored using a common coordinate system for O. If O is moving as a rigid body, it is more efficient to rotate a few radiance samples in L _pi to match O than to rotate multiple transfer functions of O.

For a diffuse surface, a simple implementation of action 350 is to store the transfer vector for each vertex and do the dot product in the vertex shader. The transfer vector is stored in the texture map instead of storing it for each vertex,
It can also be evaluated using a pixel shader. These coefficients are signed quantities that are not always in the range [-1,1], so DirectX 8.1 pixel shaders (V1.4) or those that provide a wider range of [-8,8] Corresponding to OpenGL (extension is by ATI) can be used. In one implementation, the pixel shader requires only eight instructions to perform the dot product and stores the projection coefficient of L _{p in} a constant register.

Simulations of color bleeding in a colored environment or O are done in three passes, each pass with a separate dot product for the r, g, and b channels. Otherwise, one pass is sufficient.

For gloss self-propagation, the matrix transformation of equation (9) can be done in software, because the transfer matrix may be too large to operate in either the current vertex shader or the pixel shader. . The result is (L ' _p ) _i , the SH coefficient of the transmitted radiance at point p on O. And in the pixel shader, G ^*
The convolution 370 can be performed using the simple cosine power of (Phong lobe) kernel and the result evaluated in the reflection direction R. This result is

[0146]

[Equation 37]

Can be written as Current graphics hardware can evaluate SH projections up to the fifth order.

4.1 Spatial Sampling of the Incident Radiance Field A simple and useful approach to dynamically sample the incident radiance is to sample the incident radiance at the center point of O. To address local lighting variations on O, more sophisticated techniques sample the incident lighting at multiple points. Given a desired number of points as inputs, an ICP (Iterative Nearest Neighbor) algorithm can be used to obtain a suitable set of sample points as pre-processing (see, eg, Non-Patent Document 16). As a result, the points P that are uniformly distributed on O are close to O at which the incident lighting can be sampled at execution
A representative set of _i is generated. The rendering engine 140 blends the contributions from each sphere _{Lpi of the} resulting sampled radiance with each point p on O blending.
It is also possible to pre-calculate the coefficients at to obtain the incident radiance field on O, denoted L _p above.

FIG. 11 shows a sampling of the incident radiance at a single point, an ICP point, and multiple samples,
5 represents an image resulting from rendering using the runtime procedure 300 of FIG.

4.2 SH Radiance Sampling by Graphics Hardware Graphics hardware is useful for capturing radiance samples {L _Pi } in dynamic scenes. To do this, we render 6 images from each P _i that correspond to the 6 faces of the sphere parameterization of the cubemap. O
Itself must be removed from the rendering.
The cubemap image is then projected onto their SH coefficient using the integral of equation (1).

From the viewpoint of efficiency, textures are calculated in advance for the basis functions weighted by the difference solid angles evaluated for the parameterization of the cube map of s.

[0152]

[Equation 38]

.. The resulting integral is then the sample of the captured L _p (s) and the texture

[0154]

[Formula 39]

It is a simple dot product of

Ideally, this calculation is done in graphics hardware. Alternatively, because of the problem of accuracy and the fact that the inner product cannot be performed by hardware, the sampled radiance image may be read back and projected by software. In this case, it is useful to reduce the resolution of the image to be read back as much as possible.

The low-order SH projections can be calculated using a very low resolution cubemap, provided that they are properly bandlimited. For example, a 6th order band-limited sphere signal can be projected using a 64x4 image, by assuming a power unit signal (ie, a signal whose integral square on the sphere is 1). If the error is normalized, the squared error of the average case is about 0.3.
%, The worst case squared error is about 1%. (More precisely, the mean-case error is the integrated squared difference between the reference signal and the reconstructed signal averaged over all power-unit signals. The worst-case error is the same integral error. However, for a 6 × 8 × 8 map, this error is 0.003% on average and 0.
It drops to 02%. The disadvantage is that typical signals are not band limited to a sphere. According to another analysis, assuming continuous bilinear reconstruction for sampled 2D images, projection to 6th order using 6x8x8 images has a mean-squared error of 0.7%, The worst case error is 2%, which is 0.2% and 0.5% for a 6 × 16 × 16 image, respectively.

6 × 1 from the hardware in the exemplary implementation
Extract a 6x16 image. Aliasing of 2D images is still a problem, as the above analysis uses bilinear reconstruction from point samples as a reference, as is always the case with point-sampling rendering. To reduce aliasing, the cubemap image can be supersampled (eg, doubled) in each dimension, and box filtered decimation can be done in hardware prior to readout and projection. Similarly, as a preprocessing, the basis function texture can also be supersampled and decimated. For radiance samples including readback and SH projection, this exemplary implementation is AT
PIII-933PC with I Radeon 8500
It takes about 1.16 ms.

4.3 Volume model self-propagation The same framework as for the surface is used for volume self-propagation data. The resulting precomputed model allows for runtime changes in lighting, resulting in accurate shadowing and interreflection in any low frequency lighting environment. FIG. 12 shows a self-transfer in a volume with a variation (discussed below) of the run-time procedure 300 for a volume self-transfer in order to capture how the cloud model shadows itself and scatters light. Represents an image of a cloud model using.

Similar to surface transmission, in the preprocessing step S
Simulate lighting on a volume using the H basis function as a radiator. In the case of shadowed transmission without inter-reflections (ie direct shadowing), the simulator collects energy from the radiator to all voxels p of the volume that are attenuated by its path through the volume. The numerical integration required in the direction s _d is

[0161]

[Formula 40]

Where Λ (p → q) is the integrated attenuation of the volume along the path from p to q, and D is the distance the ray (p, s _d ) leaves the volume. Is. To include interreflections, the simulation traverses every voxel p, with a random direction s
_Scatter the transmission forward along _d . Rules of communication

[0163]

[Formula 41]

Place all voxels q along s _d until the volume is exited using. The more passes a volume has, the more indirect bounces are created.

Rendering is done in the conventional manner, ie using alpha blending which takes transparency into account 3D
This is done by drawing slices through the volume in a back-to-front order. Each slice is a 2D image containing samples of the transfer vector. The pixel shader computes the dot product of the lighting coefficients and the transfer vector needed to shade each slice.

5. Proximity transfer of radiance Proximity transfer pre-computes the effect of object O on the environment in its vicinity for parameterized low frequency lighting. FIG. 13 blurs the rough terrain close to which the hang glider is captured, using a near-field transformation (described below) of the simulation 200 (FIG. 2) and a runtime rendering procedure 300 (FIG. 3). The image which casts the shadow is shown. The simulation of propagation is the same as the simulation of self-communication in Section 3, "Precalculation of Self-communication,"
Not at the top, but at a point in 3D space surrounding O.
At run time, any object R can be placed in this volume to capture shadows, reflections, and caustics cast on it by O without knowing it in advance. For example, a moving vehicle O can cast a shadow on the ground R. The thrown shadow and lighting also react to changes in lighting. For example,
Moving the light moves the hazy shadow on R. This generalizes the irradiance volume by allowing for dynamic lighting, taking into account glossy transmissions (see, for example, Non-Patent Document 17).

Since R is unknown in the step of pre-computation,
The neighborhood volume of O stores the transfer matrix rather than the vector. This also applies to diffuse objects, since the normal of R is not known in advance.

In one implementation, the simulator precomputes the transfer matrix M _p at each point in a simple 3D grid around O. At run time, the rendering engine performs the matrix transformation of equation (9) in software at each point in the volume and uploads the result to the graphics hardware. The result is the transmitted radiance (L ′ _p ).
Is a volume texture containing the coefficients of, which is applied to R.

The pixel shader then uses this transmitted radiance to illuminate the object. The diffuse object is convolved with the cosine weighted hemisphere H ^* of the radiance using Eq. (6) and the resulting SH projection is evaluated with the R normal vector. The glossy object is expressed by the formula (1
Perform 2).

For objects for which self-transmission has been calculated in advance, O
And R do not share a common coordinate system.
For this reason, a dense set of transmission samples of one of the two objects must be dynamically rotated to align with the other. Its complexity is O (n ⁴ ), which is the same as the matrix transformation of equation (9), but the rotation of higher-order projections is considerably more expensive than the operation.
Hardware improvements will soon alleviate this problem.

Compared to self-propagation, near-propagation introduces some additional approximation error. It is difficult to combine shadows or lights cast on the same object from multiple proximity transfer objects. Local lighting variations that are not due to the presence of O or R are also problematic. That is, the lighting must be fairly constant over the entire neighborhood of O to provide accurate results. In detail, when an object other than O and R gets into the vicinity of O, an error such as lack of a desired shadow occurs. Also, the vicinity of O must be large enough to contain any shadows or light that O can cast on R. Nevertheless, proximity transfer captures effects that cannot be obtained in real time with traditional methods.

6. Computing Environment FIG. 6 is a schematic example of a suitable computing environment 600 in which the illustrative embodiments may be implemented. As the invention may be implemented in various general purpose or special purpose computing environments, computing environment 600 does not imply any limitation with respect to the scope of use or functionality of the invention.

Referring to FIG. 6, computing environment 600 includes at least one processing unit 610 and memory 620. In FIG. 6, this most basic configuration 6
30 is included in the dotted line. The processing unit 610 executes computer-executable instructions and may be a real or virtual processor. In a multi-processing system, a plurality of processing units execute computer-executed instructions to increase processing power. The memory 620 may be volatile memory (eg, registers, cache, RAM), non-volatile memory (eg, ROM, EEPROM, flash memory, etc.), or some combination of the two. The memory 620 contains software 68 that implements graphic image rendering using radiance self-transfer.
Store 0.

The computing environment can include other features as well. For example, computing environment 600 includes storage 640, one or more input devices 650, one or more output devices 660, and one or more communication connections 670. An interconnect mechanism (not shown), such as a bus, controller, or network, connects the components of computing environment 600 to each other. Operating system software (not shown) typically provides an operating environment for other software executing in computing environment 600 and coordinates the activities of components of computing environment 600. As mentioned earlier, the computing environment is DirectX and Open.
GL function library, ATI Radeon and NVi
It is desirable to include graphics processing hardware and software such as a dia GeForce video card.

Storage 640, which may be removable or non-removable, may be used to store magnetic disks, magnetic tape or cassettes, CD-ROMs, CD-RWs, DVDs, or information and is accessed within computing environment 600. Including any other medium capable of Storage 640 stores instructions for software 680 that implements the image rendering system (FIG. 1).

The input device 650 is a keyboard, a mouse,
It may be a touch input device such as a pen, trackball, a voice input device, a scanning device, or other device that provides input to the computing environment 600. For audio, input device 650 may be a sound card or similar device that accepts audio input in analog or digital form. Output device 660 may be a display, printer, speaker, or other device that provides output from computing environment 600.

Communication connection 670 enables communication with another computing entity over a communication medium. Communication media carries information such as computer-executable instructions, compressed audio or video information, or other data in a modulated data signal. A modulated data signal is a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal.
By way of example, and not limitation, communication media include hard-wired or wireless technologies implemented with electrical, optical, RF, infrared, acoustic, or other carrier waves.

The present invention can be described in the general context of computer-readable media. Computer-readable media are any available media that can be accessed within a computing environment. By way of example, and not limitation, computer readable media in computing environment 600 include memory 620, storage 640, communication media,
And any combination of the above media.

The invention can be described in the general context of computer-executable instructions, which are contained in program modules and are executed in an actual or virtual target processor in a computing environment. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or split between program modules as desired in various embodiments. Computer-executable instructions for program modules may execute in local or distributed computing environments.

For the sake of presentation, this detailed description refers to “determine,” “get,” to describe computer behavior within a computing environment.
"Adjust" and "appl" (appl
y) ”is used. These terms are a high level abstraction of actions performed by a computer and should not be confused with actions performed by humans.
The actual computer operation that corresponds to these words will vary from implementation to implementation.

Although the principles of the invention have been illustrated and illustrated with reference to exemplary embodiments, it will be appreciated that the exemplary embodiments can be modified in arrangement and detail without departing from principles such as those described above. It is to be understood that the programs, processes, or methods described herein are not related or limited to any particular type of computing environment, unless stated otherwise. Various types of general purpose or specialized computing environments can be used with or perform operations in accordance with the teachings described herein. Elements of the exemplary embodiment shown as software may be implemented in hardware and vice versa.

In light of the many possible embodiments to which the principles of the invention may be applied, all embodiments that come within the scope and spirit of the following claims and equivalents thereto are claimed as the invention. To do.

[Brief description of drawings]

FIG. 1 is a block diagram of a computer graphics software architecture that incorporates precomputed radiance self-transfer for real-time image rendering in a dynamic low frequency lighting environment.

2 is a flowchart of precomputation of self-communication in the image rendering system of FIG.

3 is a flow chart of real-time rendering of radiance transfer in the image rendering system of FIG.

FIG. 4 is a flow chart for processing diffusive surface self-transfer at runtime in the real-time rendering of FIG.

5 is a flow diagram for processing run-time glossy surface self-transfer in the real-time rendering of FIG.

6 is a block diagram of a computing environment suitable for the image rendering system of FIG.

7 is a diagram representing simulated inter-reflections in the pre-computation of self-transmission of FIG.

FIG. 8 is a diagram of an image produced through the real-time rendering of FIG. 3, both without shadowing.

9 is a diagram of an image showing the effect of self-propagation of a diffusing surface, generated through the real-time rendering of FIG. 3, both without shadowing.

FIG. 10 is an image showing the effect of self-propagation of a glossy surface, generated through real-time rendering of FIG. 3, both without shadowing.

11 is a diagram of an image produced via the real-time rendering of FIG. 3, sampling the incident radiance with a single point, an iterative nearest neighbor algorithm (ICP) point, and multiple samples, respectively.

12 is a diagram of an image of volume self-propagation generated via the real-time rendering of FIG.

FIG. 13 is a diagram of an image of a neighborhood transfer generated via the real-time rendering of FIG.

[Explanation of symbols]

100 Computer Graphic Image Rendering System 110 Geometric Object Model 130 Radiance Self-Transmission Data 120 Global Illumination Simulator 140 Real Time Image Rendering Engine 150 Lighting Environment 160 View Direction 170 User Control 180 Graphic Display Driver 600 Computing Environment 610 Processor 620 Memory 640 Storage 650 Input device 660 Output device 670 Communication connection 680 Software

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Peter-Pike Jay. Sloan             United States 98008 Washington             Bellevue Northeast 170 Abe             New 3518 (72) Inventor John M. Snyder             United States 98052 Washington             Redmond Northeast 186             Venue 2522 (72) Inventor Jean Kautz             Germany 66111 Saarbrucken             Humorer Strasse 29 F-term (reference) 5B080 AA14 GA08 GA11

Claims (85)

[Claims] 1. A computer-implemented method of rendering a computer graphic image of a geometrically modeled object, simulating self-shadowing and interreflection for a plurality of locations on the modeled object. And generating radiance transfer data representing a radiative response of the object including a global transfer effect on a linear combination of area-supported lighting basis functions at the plurality of positions, and a certain lighting based on the radiance transfer data. Evaluating the radiance transfer of the modeled object at the location for a view direction that is the environment; and producing an image of the modeled object shaded according to the evaluation of the radiance transfer. And a step of: 2. The method of claim 1, wherein the step of simulating includes simulating the modeled object illuminated in a reference low frequency lighting environment. 3. The method of claim 1, wherein the radiance transfer data represents a transfer of source radiance of the modeled object at the location to output radiance. 4. The method of claim 1, wherein the radiance transfer data represents a transfer of a source radiance of the modeled object at the location to a transferred incident radiance. 5. Method according to claim 1, characterized in that the step of simulating the lighting is performed as a pre-processing and the steps of evaluating and generating are performed in real time. 6. The step of evaluating the radiance transfer comprises: calculating illuminant lighting data for at least one sampling point in a lighting environment; and pre-determining a plurality of locations on the surface of the modeled object. A linear transformation is performed on the calculated light source lighting data according to the calculated radiance transfer data to obtain output radiance data from the position for a certain view, and the radiance transfer data is based on a reference lighting environment. And representing a radiance response including global propagation effects of the modeled object to source light at the location. 7. The radiance transfer data includes a radiance transfer vector representing a radiance transfer vector at a location on a diffused surface of the modeled object, and the step of evaluating the radiance transfer is performed by the modeling. Calculating a source radiance vector of at least one sampling point in the lighting environment for the diffuse surface of the object, and computing a dot product of the source radiance vector and the radiance transfer vector. The method of claim 1, wherein: 8. The radiance transfer data includes data representing a radiance response of a location on a glossy surface of the modeled object, the step of evaluating the radiance transfer comprising: Calculating a source radiance vector of at least one sampling point in the lighting environment for the glossy surface, the source radiance vector, the data representing a radiance response of the position, and a bidirectional reflectance distribution of the position. Determining the exit radiance from the position for a view direction as a function of a function. 9. The radiance transfer data includes data representing a radiance response of a location on a glossy surface of the modeled object, the step of evaluating the radiance transfer comprising: Calculating a source radiance vector of at least one sampling point in the lighting environment for the glossy surface; The method according to claim 1, further comprising: determining brightness, and determining an output radiance from the position in a certain view direction based on the transmitted radiance and a bidirectional reflection distribution function of the position. The method described. 10. The radiance transfer data includes a radiance transfer matrix representing a radiance transfer matrix at a location on a glossy surface of the modeled object, and the step of evaluating the radiance transfer is performed by the modeling. Calculating a source radiance vector of at least one sampling point in the lighting environment for the glossy surface of the object, and performing a matrix-vector multiplication of the source radiance vector and the radiance transfer matrix; Multiplying the product of matrix-vector multiplications by convolving the bidirectional reflectance distribution function of the position on the glossy surface of the modeled object and evaluating the result of the convolution for a view direction. The method according to Item 1. 11. A computer-implemented method of real-time rendering a computer graphic image of a surface of a geometrically modeled object, the method comprising: calculating illuminant lighting data for at least one sampling point in a lighting environment. And an output radiance from the position for a view direction that has been linearly transformed to the calculated source lighting data according to pre-calculated radiance transfer data for a plurality of positions on the surface of the modeled object. Of data, the radiance transfer data including a radiance response including a global propagation effect of the modeled object to a source light at the position in a reference lighting environment. A step of representing the morphism luminance reaction was shaded according to the outgoing radiance from said position, a method which comprises the steps of generating an image of the modeled object in the lighting environment for the view direction. 12. Performing the linear transformation in the case of a diffused surface of the modeled object comprises computing a dot product of a source radiance vector and a pre-computed radiance transfer vector, wherein: The source radiance vector represents the source light at a sampling point,
12. The method of claim 11, wherein the pre-computed radiance transfer vector represents a radiance response at a location on a diffused surface of the modeled object. 13. Performing the linear transformation in the case of a glossy surface of the modeled object includes calculating a source radiance vector of at least one sampling point in a lighting environment; Determining the exit radiance from the position for a view direction as a function of data representing the radiance response of the position and a bidirectional reflectance distribution function of the position. The method described. 14. Performing the linear transformation in the case of the glossy surface of the modeled object includes calculating a source radiance vector of at least one sampling point in a lighting environment; Determining the transfer radiance of the position based on the data representing the radiance response of the position, and based on the bidirectional reflectance distribution function of the transfer radiance and the position from the position for a view direction. Determining the exit radiance of the. 15. Performing the linear transformation in the case of a glossy surface of the modeled object includes: a source radiance vector representing the source light at a sampling point and a position on the glossy surface of the modeled object. Matrix-vector multiplication with a pre-computed radiance transfer matrix representing the radiance response of the bidirectional reflection distribution of the positions on the glossy surface of the modeled object. The method of claim 11, comprising convolving a function and evaluating the result of the convolution for a view direction. 16. A computer implemented method for real time rendering of a computer graphic image of a diffuse surface of a geometrically modeled object, the method comprising projecting radiance in a lighting environment onto a spherical harmonic basis. Generating a lighting coefficient vector for at least one position of the modeled object, a dot product of the lighting coefficient vector and a radiance transfer vector previously calculated for at least one position on the diffusing surface. To generate an exit radiance at the at least one location on the diffuse surface of the modeled object for the lighting environment and view direction,
Said pre-calculated radiance transfer vector representing a radiance response including global propagation effects of said diffusing surface at said at least one location with respect to source light in a reference low frequency lighting environment; Calculating shading of the diffuse surface of the modeled object at at least one location, and applying the shading to generate an image of the diffuse surface of the modeled object. Method. 17. A computer-implemented method of real-time rendering a computer graphic image of a glossy surface of a geometrically modeled object, the method comprising projecting radiance in a lighting environment onto a spherical harmonic basis. Generating a lighting coefficient vector for at least one position of the modeled object, the lighting coefficient vector and the glossy surface for source light at the at least one position in a reference low frequency lighting environment. Of pre-computed radiance transfer matrix data representing a radiance response including global propagation effects of the at least one location and the at least one position on the glossy surface for the view direction as a function of the bidirectional reflectance distribution function. One Determining exit radiance from a location, calculating shading of the glossy surface of the modeled object at the at least one location, and an image of the glossy surface of the shaded modeled object. And generating. 18. The step of determining exit radiance for a view direction comprises matrix-vector multiplication of the lighting coefficient vector and a radiance transfer matrix previously calculated for at least one position on the glossy surface. , The pre-computed radiance transfer matrix represents a radiance response including global propagation effects of the glossy surface to source light at the at least one location under a reference low frequency lighting environment; Convolving the matrix-vector multiplication product with a bidirectional reflectance distribution function of the position of the modeled object on the glossy surface; and evaluating a result of the convolution for a view direction. The method according to claim 17, wherein: 19. The step of determining exit radiance for a view direction is based on matrix vector multiplication of the lighting coefficient vector and a radiance transfer matrix previously calculated for at least one location on the glossy surface. Calculating the exit radiance, the pre-computed radiance transfer matrix being a global transfer of the glossy surface to source light in the reference low frequency lighting environment at the at least one location. 18. Including a representation of a radiance response including an effect.
The method described in. 20. A computer-implemented method of generating radiance self-transfer data for use in real-time rendering and display of a computer graphic image of a modeled object, represented as a linear combination of region-supporting basis functions. Performing a global illumination simulation on the modeled object for a reference lighting environment, the simulation calculating a radiance transfer of a plurality of sample points on the modeled object, the simulation comprising: Incorporating self-shadowing and inter-reflection of a modeled object, rendering radiative transfer on the modeled object to render the modeled object. For use in generating bets graphic image, a method which comprises the steps of: a global illumination simulation of the plurality of sample points is recorded as a linear transformation of the coefficients of the region supporting basis functions. 21. The method of claim 20, wherein the reference lighting environment is parameterized as a projection of a spherical harmonic basis of a sphere of source light. 22. Performing the global illumination simulation on the modeled object in a plurality of passes, a plurality of on the modeled object including self-shadowing of the modeled object in a first shadowing pass. Simulating direct illumination of the sample points of the direct illumination of the sample points within a hemisphere centered on a normal direction from a sample point to a surface of the modeled object for a certain sample point. Of the sample points at the sample points, tagging as occlusions in directions from the sample points that intersect the modeled object, and direct illumination of the sample points in the directions. Radiation from A step of accumulating luminance transfer to generate radiance transfer data for the sample points, and a subsequent inter-reflection path to simulate inter-reflection illumination of the plurality of sample points on the modeled object. Simulating the inter-reflected illumination by accumulating radiance transfer from the inter-reflected illumination of the sample points in the plurality of directions tagged for occlusion for a sample point. , Generating radiance transfer data for said sample points, iterating an interreflection path until a condition is met, and for a sample point, accumulating at said first shadowing path and subsequent interreflection paths The radiance transfer data that has been sampled is summed to obtain a total radiance transfer of the sample points. 21. The method of claim 20, comprising: 23. The method of claim 20, wherein the condition is that the total inter-reflecting illumination energy of the current inter-reflecting path is below a certain threshold level. 24. Radiance transfer data for sampling points on the diffused surface of the modeled object,
The method of claim 20, further comprising representing as a vector of spherical harmonic basis coefficients. 25. Radiance transfer data of sampling points on the glossy surface of the modeled object,
The method of claim 20, further comprising representing as a matrix of spherical harmonic basis coefficients. 26. A method for real-time rendering of a computer graphic image of a surface of a geometrically modeled object, the low frequency lighting as a reference during a radiance transfer precomputation step performed on a first computer. Performing a lighting simulation of the modeled object in an environment to calculate radiance transfer of a plurality of sample points on the modeled object, the simulation comprising self-shadowing and mutual interaction of the modeled object. Incorporating reflections; recording sampled radiance transfer as a radiance transfer vector of spherical harmonic basis coefficients for sample points on a diffused surface of the modeled object; Recording the calculated radiance transfer as a radiance transfer matrix of spherical harmonic basis coefficients for sample points on a glossy surface of a ringed object, a second computer performing an image rendering step, A step of rendering an image of the modeled object, wherein the rendering step samples the light source light of the modeled object in a lighting environment at the time of rendering to obtain the plurality of samples on the modeled object. Obtaining a source radiance vector of spherical harmonic basis coefficients of a point, and for a sample point on a diffused surface of the modeled object, performing a dot product of the source radiance vector and the radiance transfer vector. View direction And, for sample points on the glossy surface of the modeled object, based on the source radiance vector, the radiance transfer matrix, and the bidirectional reflectance distribution function of the glossy surface of the modeled object. A step of obtaining an exit radiance for a certain view direction, and a step of shading the modeled object at the plurality of sample points based on the exit radiance in the view direction of each of the plurality of sample points. And a step including. 27. Determining the exit radiance for the view direction comprises multiplying the source radiance vector by the radiance transfer matrix, and determining a product of the source radiance vector and the radiance transfer matrix. 27. Evaluating the convolution of the bidirectional reflectance distribution function of the glossy surface of at the sample point for the view direction to obtain an exit radiance in the view direction. . 28. In an image rendering step performed on a second computer, selecting a lighting environment at the time of rendering under user control, and selecting the modeled object for various lighting environments selected by the user control. The method of claim 26, further comprising rendering a further image. 29. In a second computer-implemented image rendering step, changing the view direction under user control, and providing additional images of the modeled object for various view directions controlled by the user. The method of claim 26, further comprising: rendering. 30. A memory for storing a geometrical model of an object and a region-supported lighting basis function at a plurality of positions simulating self-shadowing and inter-reflection of lighting at a plurality of positions on a modeled object. An illumination simulator that produces radiance transfer data representing the radiative response of the object including a global transfer effect for a linear combination of the radiance transfer data, and the radiance transfer data for the view of the modeled object in a lighting environment for a view A real-time image rendering engine that evaluates radiance transfer at a plurality of positions and generates an image of the modeled object shaded according to the radiance transfer evaluation, and an image output device that presents the image. A computer graphic image rendering system comprising. 31. The computer graphic image rendering of claim 30, wherein the radiance transfer data represents a transfer of source radiance of the modeled object at the plurality of positions to output radiance. system. 32. The computer graphic of claim 30, wherein the radiance transfer data represents a transfer of the source brightness of the modeled object at the plurality of locations to a transferred source radiance. Image rendering system. 33. A memory for storing a geometric model of an object, a reference including self-shadowing and interreflection by the object, parametrized by a spherical harmonic basis at a plurality of points on the surface of the object; An illumination simulator for calculating a radiance response to a low-frequency lighting environment, calculating a source radiance at the plurality of points parameterized by the spherical harmonic basis, and transforming the source radiance and the radiance response in combination. Obtaining the exit radiance from the plurality of points to the processed view and generating an image of the object that shades the object at the plurality of points according to the exit radiance of each of the plurality of points to the view A real-time image rendering engine, A computer graphic image rendering system comprising: an image output device that presents an image. 34. A plurality of objects on the surface of the object including self-shadowing and inter-reflection by the object simulated in a reference lighting environment represented as a linear combination of a geometric model of the object and a region-supporting basis function. A memory for storing radiance response data for the points, calculating the source radiance at the plurality of points, and processing the transformation combining the source radiance and the radiance response from the plurality of points for the view Presenting the image and a real-time image rendering engine that obtains an output radiance of the object and generates an image of the object that shades the object at the points according to the output radiance of each of the points for the view. An image output device. -Time graphics rendering system. 35. The radiance response data is parameterized with a spherical harmonic basis.
4. The real-time graphic rendering system according to item 4. 36. The real-time graphic rendering system of claim 34, wherein the incident radiance is parameterized with a spherical harmonic basis. 37. Program code for processing data comprising a geometric model of an object and radiance self-transfer reactions of a plurality of points on said object, thereby rendering an image of the geometrically modeled object. A computer-readable data-bearing medium for carrying the program code means for calculating light source lighting at one or more sample points close to the object in a lighting environment represented as a linear combination of region-supporting basis functions. And computer program readable means for combining the illuminant lighting data and the radiance self-transmission response to obtain exit radiance from the point on the object in a view direction. Carrier medium. 38. The radiance self-transfer response data is
For a point on the diffuse surface of the object, including a radiance self-transfer response vector of a region-supported lighting basis, the program code further includes a source light vector and the radiance for the point on the diffuse surface of the object. 36. Computer-readable data carrier according to claim 35, comprising program code means for performing a dot product with a self-transfer reaction vector to obtain the exit radiance from the point on the object in the view direction. Medium. 39. The computer-readable data-bearing medium of claim 38, wherein the area-supported lighting basis is a spherical harmonic basis. 40. The radiance self-transfer response data is
A point on the glossy surface of the object, including a radiance self-transfer reaction matrix of a spherical harmonic basis, the program code further comprising a source light vector, the radiance for the point on the glossy surface of the object, 38. Program code means for determining the exit radiance from the point on the object in a view direction based on a self-transfer response matrix and a bidirectional reflectance distribution function. Computer-readable data-bearing medium. 41. Obtaining the outgoing radiance comprises: matrix-vector multiplication of a light source light vector and the radiance self-transfer reaction matrix;
41. The method of claim 40, comprising convolving a bidirectional reflectance distribution function evaluated for the view direction to obtain the exit radiance from the point on the object in the view direction. Computer-readable data-bearing medium. 42. Determining the exit radiance further comprises convolving a product of the matrix-vector multiplication with a bidirectional reflectance distribution function evaluated for the view direction to determine the points on the object in the view direction. 42. The computer-readable data-bearing medium of claim 41 including obtaining the exit radiance from the. 43. The radiance self-transfer reaction data includes a radiance self-transfer reaction matrix, and the program code further performs matrix-vector multiplication of the source light vector and the radiance self-transfer reaction matrix. 38. The computer-readable data-bearing medium of claim 37, comprising program code means for determining the exit radiance from a point on the object for a view direction. 44. A computer implemented method for evaluating an incident radiance field of a lighting environment for use in rendering and displaying a computer graphic image of a modeled object in a low frequency lighting environment, the lighting environment. Generating an image of the source radiance in multiple directions from a point in the image, transforming the image as a linear combination of area-supported lighting basis functions of the source radiance field at the point in the lighting environment. Rendering, shading the modeled object at the points using a lighting basis representation when rendering an image of the modeled object in the lighting environment; Method characterized by including the step of visually presenting the image of the modeled object in Lighting environment. 45. The method of claim 44, wherein the lighting basis representation is a spherical harmonic basis representation. 46. A lighting environment within a dynamic scene for use in rendering and displaying a computer graphic image of a modeled object in the dynamic scene using graphics hardware on the computer. Storing a set of pre-computed texture images representing region-supported lighting basis functions, the method comprising: capturing, from the sample points in the dynamic scene, the image from the sample points in the dynamic scene. Rendering a set, wherein the image represents source light from the lighting environment of the dynamic scene, the set of source light images and the pre-computed texture image in the graphics hardware. Source radiance as a function of set Projecting degrees at the sample points to obtain the source radiance at the sample points in the region-supported lighting basis; and, when rendering an image of an object modeled in the lighting environment, the sample points at the sample points. Shading the modeled object at the point using a source radiance, and visually presenting the image of the modeled object in the lighting environment. . 47. The method of claim 46, wherein the pre-computed texture image represents a set of spherical harmonic (SH) basis functions. 48. The method of claim 47, wherein the set of spherical harmonic basis functions is weighted by a differential solid angle and evaluated via a cubemap sphere parameterization of the angle. . 49. The method of claim 46, wherein the rendered image corresponds to a surface of a cubemap spherical parameterization of source light from a low frequency lighting environment of the dynamic scene. 50. A radiance sample lighting environment within a dynamic scene for use in rendering and displaying computer graphic images of modeled objects in the dynamic scene using graphics hardware on the computer. Of a texture image pre-computed for a set of spherical harmonic (SH) basis functions weighted by the difference solid angle and evaluated via a cubemap sphere parameterization of said angle. Storing a set, and rendering the set of images in the graphics hardware from sample points in the dynamic scene, the image of the source light from a low frequency lighting environment of the dynamic scene. Cubemap spherical surface Projecting the light radiance at the sample point onto a spherical harmonic coefficient as a dot product of the set of source light images and the set of SH basis function texture images in the graphics hardware, Obtaining the source radiance at a point in the SH basis, and using the light radiance of the SH basis at the sample point in rendering the image of the object modeled in the lighting environment at the point. Shading a modeled object, and visually presenting the image of the modeled object in the lighting environment. 51. The method of claim 50, further comprising repeating rendering and projection operations for a plurality of spatially different sample points in the dynamic scene. 52. The method of claim 50, further comprising supersampling and box filter decimation of the source light image prior to projecting to SH coefficients, thereby reducing aliasing. 53. The method of claim 50, further comprising supersampling and box filter decimation of the SH basis function texture image prior to projecting onto SH coefficients. 54. A computer system for graphic rendering, prior to a set of basis functions of spherical harmonics (SH) weighted by a difference solid angle and evaluated via a cubemap sphere parameterization of said angle. A memory for storing a set of calculated texture images, and a graphics processing board operative to render the set of images from sample points in a dynamic scene having a lighting environment, the image comprising: Source radiance at the sample point corresponding to the cubemap spherical parameterized surface of the source light from the lighting environment of the scene, and further as a dot product of the set of source light images and the set of SH basis function texture images. Of the sample points by projecting A graphics processing board that operates to obtain the source radiance at the SH basis, and uses the source radiance at the SH basis at the sample point in rendering an image of an object modeled in the lighting environment. And a system for executing a real-time image rendering engine operative to shade the modeled object at the point, and an image output device for presenting the image. 55. Using computer graphics hardware, a radiance sample used to render and display a computer graphic image of a modeled object in a dynamic scene using a lighting environment in the scene. A computer-readable data-bearing medium carrying captured program code, the program code causing the graphics hardware to render a set of images from sample points in the dynamic scene, the images being the dynamic images. Cube map of source light from a low frequency lighting environment of the scene, program code means corresponding to the surface of the spherical parametrization, said graphic hardware having a set of source light images and a cube of said angles weighted by a differential solid angle. Spherical harmonics of the incident radiance at the sample points as a dot product with a set of SH basis function texture images previously calculated for a set of spherical harmonic function (SH) basis functions evaluated via spherical surface parameterization Program code means for projecting onto a function coefficient to obtain the incident radiance at the sample point in the SH basis, and the SH basis at the sample point in rendering an image of an object modeled in the lighting environment. Program code means for shading the modeled object at the point using the incident radiance of, and program code means for visually presenting the image of the modeled object in the lighting environment. A compilation featuring Computer-readable data-bearing medium. 56. A computer-implemented method of generating radiance self-transfer data for use in real-time rendering and display of a computer graphics image of a volume model, the method including self-shadowing of lighting on an illuminated volume, and Simulating inter-reflections to generate radiance transfer data representing radiance responses including global transfer effects for a linear combination of area-supported lighting basis functions for a plurality of volume elements of the volume model; Evaluating the radiance transfer of the plurality of volume elements in a lighting environment for a view based on the brightness transfer data; and an image of the volume model shaded according to the evaluation of the radiance transfer. And generating. 57. A computer-implemented method of generating radiance self-transfer data for use in real-time rendering and display of a computer graphics image of a volume model, the reference lighting represented as a linear combination of area-supported basis functions. A global illumination simulation of a volume for an environment, the simulation calculating radiance transfer for a plurality of volume elements of the volume model, the simulation incorporating self-shadowing and mutual reflection of the volume model. And a plurality of volumes for use in rendering a radiance transfer on the volume model to generate a graphic image of the volume model. Global illumination simulation for the original,
Recording as a linear transformation of the coefficients of the basis function. 58. The method of claim 57, wherein the reference lighting environment is parameterized as a projection of a spherical harmonic basis of a sphere of source light. 59. In the reference lighting environment, comprising: performing the global illumination simulation on the volume model in a plurality of passes; and including an attenuation along a path through the volume model in a first shadowing pass. Simulating direct illumination of multiple volume elements of the volume model, with subsequent inter-reflection paths including attenuation through the volume model of the volume element of the volume model by other volume elements of the volume model. Simulating inter-reflection illumination, repeating the inter-reflection path until the conditions are met, and for a sample point, sum the radiance transfer data accumulated in the first shadowing path and the following inter-reflection paths. do it Claim 5, characterized in that it comprises the steps of generating a total radiance transfer in the sample point
7. The method according to 7. 60. A computer-implemented method of real-time rendering a computer graphic image of a volume model, the method comprising: calculating illuminant lighting data for at least one sampling point in a lighting environment; Performing a linear transformation on the calculated illuminant lighting data according to the radiance transfer data previously calculated for the volume element,
Obtaining data of outgoing radiance from the volume element for a view, the radiance transfer data including global propagation effects of the volume model on a source light at a position in a reference lighting environment. Representing a radiance response, and generating an image of a modeled object for the view in the lighting environment, shaded according to the exit radiance from the location. 61. As a pre-calculation, a step of performing global illumination simulation on the volume for a reference lighting environment parameterized as a projection of a spherical harmonic basis of the spherical surface of the source light, the projection being a coefficient of a spherical harmonic function. And the simulation calculates radiance transfer of a plurality of volume elements of the volume model, the simulation incorporating self-shadowing and inter-reflection of the volume model, the global for the plurality of volume elements. 61. recording the illumination simulation as a linear transformation of the coefficients of the spherical harmonics of the source light to form the radiance transfer data. 62. In the reference lighting environment, comprising: performing the global illumination simulation on the volume model in multiple passes; and including attenuation along a path through the volume model in a first shadowing pass. Simulating direct illumination of multiple volume elements of the volume model, with subsequent inter-reflection paths including attenuation through the volume model of the volume element of the volume model by other volume elements of the volume model. Simulating inter-reflective illumination, repeating the inter-reflective path until a condition is met, and for a sample point the radiance transfer data accumulated in the first shadowing pass and the following inter-reflective pass. total Te method of claim 61, characterized in that it comprises the steps of generating a total radiance transfer in the sample points. 63. A computer graphic image rendering system for real time rendering a computer graphic image of a volume model, the memory storing a volume model and a plurality of volume elements of the volume model illuminated in a reference lighting environment. Simulates self-shadowing and inter-reflection of lighting in
An illumination simulator that generates data representing radiance transfer from light source light to output light of the volume element of the volume model; and the volume model in a second lighting environment for a certain view direction based on the radiance transfer data. And a real-time image rendering engine for generating an image of the volume model shaded according to the evaluation of the radiance transfer, and an image output device for presenting the image. System to do. 64. A computer-readable program-bearing medium carrying a program, the method executing on a computer to generate radiance self-transfer data for use in real-time rendering and display of a computer graphic image of a volume model. And the method is a step of performing global illumination simulation on a volume for a reference low frequency lighting environment parameterized as a projection of a spherical harmonic basis of a spherical surface of the source light, the projection being a spherical harmonic function. And the simulation computes the radiance transfer of a plurality of volume elements of the volume model,
The simulation incorporates self-shadowing and mutual reflection of the volume model, and the plurality of volume elements for use in rendering radiance transfer on the volume model to generate a graphic image of the volume model. Recording the global illumination simulation as a linear transformation of the spherical harmonic coefficient of the source light. 65. The method further comprising: performing the global illumination simulation on the volume model in multiple passes; and damping along a path through the volume model in a first shadowing pass. Simulating direct illumination of a plurality of volume elements of the volume model in a low frequency lighting environment, the subsequent inter-reflection path including attenuation through the volume model by the other volume elements of the volume model. Simulating the inter-reflecting illumination of the volume element of the volume model, repeating the inter-reflective path until a condition is satisfied, and for a sample point, accumulating in the first shadowing path and the following inter-reflective path. Before By summing the radiance transfer data, according to claim 64, characterized in that it comprises the steps of generating a total radiance transfer of the sample point
The medium described in. 66. A computer-readable program-bearing medium carrying a program, the program executing on a computer to perform a method of real-time rendering of a computer graphic image of a volume model, the method comprising: a writing environment. Calculating light source lighting data for at least one sampling point with a linear transformation to the calculated light source lighting data according to radiance transfer data previously calculated for a plurality of volume elements of the volume model,
Obtaining data of outgoing radiance from the volume element for a view direction, the radiance transfer data comprising a global transfer effect on a linear combination of area-supported lighting basis functions of the volume element of the volume model. Representing a radiance response comprising: producing a modeled image of the modeled object for the view in the lighting environment, shaded according to the exit radiance from the location. Program carrying medium. 67. The method further comprises, as a pre-computation, performing global illumination simulation on the volume for a reference low frequency lighting environment parameterized as a projection of a spherical harmonic basis of a sphere of source light, the method comprising: The projections have coefficients of spherical harmonics, the simulation calculating radiance transfer of a plurality of volume elements of the volume model, the simulation incorporating self-shadowing and mutual reflection of the volume model; Recording the global illumination simulation for a plurality of volume elements as a linear transformation of coefficients of the spherical harmonics of the source light to form the radiance transfer data. Computer readable program carrying medium according to. 68. The method further comprising: performing the global illumination simulation on the volume model in a plurality of passes; and dampening along a path through the volume model in a first shadowing pass. Simulating the direct illumination of a plurality of volume elements of the volume model in a low frequency lighting environment, the subsequent inter-reflection path comprising attenuation through other volume elements of the volume model by the other volume elements of the volume model. Simulating the inter-reflecting illumination of the volume element of the volume model, repeating the inter-reflective path until a condition is met, Before By summing the radiance transfer data, according to claim 67, characterized in that it comprises the steps of generating a total radiance transfer of the sample point
A computer-readable program-bearing medium according to item 1. 69. A computer-implemented method of generating radiance neighborhood transfer data from modeled objects in a scene for use in real-time rendering and display of computer graphic images of modeled objects in the scene. A step of performing a global illumination simulation on the modeled object for a reference lighting environment represented as a linear combination of basis functions, the simulation comprising a plurality of adjacent spaces in a proximity space around the modeled object. Compute the radiance transfer of the sample points and the simulation incorporates self-shadowing and inter-reflections cast from the modeled object to the plurality of sample points. And the global illumination simulation for the plurality of sample points as a linear transformation of the coefficients of the basis function for use in rendering the radiance transfer on the volume model to generate a graphic image of the volume model. A step of recording. 70. Performing the global illumination simulation on the modeled object in a plurality of passes, comprising shadowing the plurality of sample points from the modeled object in a first shadowing pass. Simulating direct illumination of the plurality of sample points within the proximity space around the modeled object in a reference low frequency lighting environment, with subsequent inter-reflection paths from the modeled object from the modeled object. Simulating the reflected illumination of the plurality of sample points in the proximity space around the object, repeating the inter-reflection path until a condition is met, and for a sample point, the first shadowing path. By summing the radiance transfer data accumulated in the subsequent inter-reflection path, claim 69, characterized in that it comprises the steps of generating a total radiance transfer of the sample point
The method described in. 71. A record of the global illumination simulation of the plurality of sample points as a linear transformation of a spherical harmonic coefficient of the source light is recorded for a plurality of points in a three-dimensional grid centered on the modeled object. 70. The method of claim 69, which is in the form of a transfer matrix of 72. A computer implemented method for real time rendering a computer graphic image of a modeled object in a scene including a modeled object, the source lighting data for at least one sampling point in a lighting environment. And performing a linear transformation on the calculated illuminant lighting data according to pre-computed radiance neighborhood transfer data for a plurality of sample points in a near space around the modeled object for a view. Obtaining transfer radiance data at a plurality of locations on the modeled object, the radiance transfer data being the modeled object in a reference lighting environment. Representing a radiance response including global propagation effects of the modeled object to source light at the plurality of sample points in the near space of an edge, shaded according to the transferred radiance from the plurality of locations, Generating an image of the modeled object in the lighting environment for a view. 73. The operation of performing the linear transformation includes dynamically rotating the coordinate system direction of the plurality of sample points to the coordinate system direction of the modeled object. Item 72. The method according to Item 72. 74. As a pre-computation, performing global illumination simulation on the modeled object for a reference low frequency lighting environment parameterized as a projection of a spherical harmonic basis of a sphere of source light, the projection comprising: Having a coefficient of spherical harmonics, the simulation calculating radiance transfer of the plurality of sample points in the proximity space around the modeled object, the simulation shadowing from the modeled object. Injecting and reflecting, and recording the global illumination simulation for the plurality of sample points as a linear transformation of the coefficients of the spherical harmonics of the source light to generate the radiance transfer data. The method of claim 72, characterized in that it comprises a-up. 75. The global illumination simulation recording of the plurality of sample points as a linear transformation of the spherical harmonic coefficient of the source light is performed for a plurality of points in a three-dimensional grid centered on the modeled object. 75. The method of claim 74, which is in the form of a transfer matrix of 76. A step of performing the global illumination simulation in multiple passes, the first shadowing pass including a shadowing from the modeled object, around the modeled object in the reference lighting environment. Simulating direct illumination of the plurality of sample points in the near space; simulating reflected illumination of the plurality of sample points by the modeled object in a subsequent reflection pass; Repeating the interreflection path up to, and, for a sample point, summing the radiance transfer data accumulated in the first shadowing path and the subsequent interreflection path to produce a total radiance transfer for the sample point. Including steps to Claim 74, wherein the door
The method described in. 77. A computer graphic image rendering system for real time rendering a computer graphic image of a modeled object in a scene containing modeled objects, the memory storing a model of the object and the object. A plurality of sample points from the modeled object simulating self-shadowing and inter-reflection of lighting for a plurality of sample points in a close space around the modeled object illuminated in a lighting environment Illumination simulator for generating data representing radiance transfer from source light to transferred incident light for a point, and a second lighting environment for a view based on the radiance transfer data A real-time image that evaluates the radiance transfer from the modeled object to the modeled object at the plurality of sample points and produces an image of the modeled object shaded according to the evaluation of the radiance transfer. A system comprising a rendering engine and an image output device for presenting the image. 78. A computer for performing a method of generating radiance neighborhood transfer data from a modeled object in a scene for use in real-time rendering and display of a computer graphic image of a modeled object in the scene. A computer-readable program-bearing medium carrying a program to be programmed, the method comprising: performing a global illumination simulation on the modeled object in a reference lighting environment represented as a linear combination of basis functions, the simulation comprising: Computes the radiance transfer of a plurality of sample points in the near space around the modeled object, and the simulation calculates the plurality of subsamples from the modeled object. Incorporating self-shadowing and inter-reflections that are cast on pull points, for the plurality of sample points for use in rendering radiance transfer on the volume model to generate a graphic image of the volume model. The global illumination simulation,
Recording as a linear transformation of the coefficients of the basis function. 79. The method comprises: performing the global illumination simulation on the modeled object in multiple passes; and shadowing the plurality of sample points from the modeled object in a first shadowing pass. Simulating direct illumination of the plurality of sample points in the proximity space around the modeled object in the reference lighting environment, including the following inter-reflection paths from the modeled object: Simulating the reflected illumination of the plurality of sample points in the near space around the modeled object, repeating the inter-reflection path until a condition is met, and for a sample point the first shadowing Scan and by summing the radiance transfer data accumulated in inter-reflection path followed, claim 78, characterized in that it comprises the steps of generating a total radiance transfer of the sample point
A computer-readable program-bearing medium according to item 1. 80. The recording of the global illumination simulation of the plurality of sample points as a linear transformation is in the form of a transfer matrix for a plurality of points in a three-dimensional grid centered on the modeled object. 80. The computer-readable program-bearing medium of claim 79. 81. A computer-readable program-bearing medium carrying a program for programming a computer to perform a method of real-time rendering a computer graphic image of a modeled object in a scene including a modeled object, comprising: The method calculates light source lighting data for at least one sampling point in a lighting environment, and according to pre-calculated radiance neighborhood transfer data for a plurality of sample points in a near space around the modeled object. Performing a linear transformation on the calculated light source lighting data to obtain output radiance data at a plurality of positions on the modeled object for a view direction, the radiance transfer A step of representing a radiance response including global propagation effects of the modeled object to source light at the plurality of sample points in the proximity space around the modeled object under a reference lighting environment. And generating an image of the modeled object in the lighting environment for the view direction, shaded according to the exit radiance from the plurality of positions. 82. The operation of performing the linear transformation includes dynamically rotating the coordinate system direction of the plurality of sample points with the coordinate system direction of the modeled object. Item 81. The medium according to Item 81. 83. The method comprises, as a pre-computation, performing a global illumination simulation on the modeled object in a reference lighting environment parameterized as a projection of a spherical harmonic basis of the sphere of the source light, the method comprising: The projections have coefficients of spherical harmonics, the simulation computes the radiance transfer of the plurality of sample points in the proximity space around the modeled object, the simulation from the modeled object Capturing the shadowing and reflection of the source illumination, and recording the global illumination simulation for the plurality of sample points as a linear transformation of the coefficients of the spherical harmonics of the source light to generate the radiance transfer data. Medium of claim 81, characterized in that it comprises a flop. 84. The recording of the global illumination simulation of the plurality of sample points as a linear transformation of the spherical harmonic coefficient of the source light is performed for a plurality of points in a three-dimensional grid centered on the modeled object. 84. The medium of claim 83 in the form of a transfer matrix of 85. The method of modeling in the reference lighting environment, comprising: performing the global illumination simulation in multiple passes; and shadowing from the modeled object in an initial shadowing pass. Simulating direct illumination of the plurality of sample points in the near space around the object, and simulating reflected illumination of the plurality of sample points by the modeled object in a subsequent reflection pass; Repeating the interreflection path until a condition is met, summing the radiance transfer data accumulated in the first shadowing path and the following interreflection path for a sample point to obtain a total radiance of the sample point The stage that produces the luminance transfer. Claim, characterized in that it comprises a flop 83
The medium described in.