System and methods for real-time rendering of deformable geometry with global illumination

• US 7,408,550 B2
• Filed: 07/23/2007
• Issued: 08/05/2008
• Est. Priority Date: 07/24/2006
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Bunnell, Michael T.
• Primary Examiner(s): Chauhan; Ulka
• Assistant Examiner(s): PAPPAS, PETER
• Application Number:
  US11/880,658
• Publication Number:
  US 20080018647A1
• Time in Patent Office: 379 Days
• Field of Search: None
• US Class Current: 345/426
• CPC Class Codes: G06T 13/00 Animation
• CPC Class Codes: G06T 15/506 Illumination models

Abstract

Frames of image data representative of a scene containing deformable geometry subject to global illumination are computed in real-time. Frame animation data is computed for a series of frames representative, in animated sequence, of a scene containing deformable geometry. The frame animation data includes vertex attributes of a control mesh descriptive of the three dimensional surface of graphical elements occurring within the scene. Based on the frame animation data, respective illumination values are computed for the individual polygons of the control mesh to determine the global illumination of the scene. The computation is performed iteratively with respect to each frame set of frame animation data until a qualified convergence of global illumination is achieved. Within each iteration, the respective illumination values are determined based on the frame-to-frame coherent polygon illumination values determined in a prior iteration. Qualified convergence is achieved within the frame-to-frame interval defined by a real-time frame rate.

Claims

1. A method, implemented within a computer system including a graphics rendering subsystem, providing for the sequential production of frames of image data representative of a scene containing deformable geometry for display on a display device, said method comprising the steps of:

a) computing frame animation data for a series of frames representative, in animated sequence, of a scene containing deformable geometry, said frame animation data including vertex attributes of a control mesh descriptive of a three-dimensional surface of graphical elements occurring within said scene, said control mesh containing a plurality of polygons; and

b) computing, based on said frame animation data, respective illumination values for said plurality of polygons to determine the global illumination of said scene, wherein said step of computing is performed iteratively with respect to each said frame corresponding frame animation data until a qualified convergence of global illumination is achieved, wherein within each iteration said respective illumination values are determined based on the frame-to-frame coherent polygon illumination values computed in a prior iteration, wherein said qualified convergence is achieved within the frame-to-frame interval defined by a real-time frame rate, and wherein each iteration includes a combined computation of global illumination values as

2. The method of claim 1 wherein said qualified convergence is achieved for each of a continuous plurality of frames occurring within the ongoing sequential production of said frames of image data for display on said display device.

3. The method of claim 2 wherein, for a predetermined polygon of said plurality of polygons, a corresponding one said respective illumination values is determined as the sum of positive illumination contributions of said plurality of polygons directly lighting said predetermined polygon and negative illumination contributions of said plurality of polygons directly shading said predetermined polygon, wherein said positive and negative illumination contributions are said respective frame-to-frame coherent polygon illumination values computed in a prior iteration.

4. A computer system providing for the real-time production of an animated scene containing deformable geometry, said computer system comprising:

a) a primary processing unit, coupled to a main memory, operative to iteratively process a control mesh, representing a deformable geometry, as sequential image frames at a real-time frame rate for display on a display device, wherein each said iteration establishes a polygon data set representative of said control mesh accessible from said main memory; and

b) a graphics rendering subsystem, coupled to said primary processing unit and said main memory, operative to access and process said polygon data set to determine global illumination values for each polygon of said polygon data set within the frame period of said real-time frame rate, said graphics rendering subsystem computing, with respect to a predetermined frame polygon data set, initial global illumination values for each polygon of said predetermined frame polygon data set based upon final global illumination values computed for each polygon in a prior frame polygon data set, wherein polygons of said predetermined and prior frame polygon data sets are related by frame-to-frame coherence of said control mesh, said graphics rendering subsystem being further operative to iteratively compute, with respect to said predetermined frame polygon data set, global illumination values for each polygon of said predetermined frame polygon data with each iteration including a single pass computation of global illumination values as

5. The computer system of claim 4 wherein L_R has a plurality of directional illumination components and wherein each said directional illumination component has red, green, and blue color components.

6. The computer system of claim 5 wherein L_R has four directional illumination components and wherein each said directional illumination component has red, green, and blue color components.

7. The computer system of claim 6 wherein said graphics rendering subsystem includes a quadrant visibility map, representing relative occlusion of pixel relative quadrants predetermined from adjacent high-resolution surface features, wherein said graphics rendering subsystem is operative to render said predetermined frame polygon data set as a frame of pixel data, and wherein said graphics rendering subsystem computes for a predetermined pixel in said frame of pixel data a corresponding pixel shading illumination value as

I=L1*s1+L2*s2+L3*s3+L4*s4

where s1, s2, s3, and s4 are the quadrant corresponding values of an entry in said quadrant visibility map corresponding to said predetermined pixel and where L1, L2, L3and L4 are the quadrant corresponding directional illumination components of L_R.

8. A method, implemented within a computer system including a graphics rendering subsystem, of determining in real-time the global illumination of a deformable geometry, said method comprising the steps of:

a) animating a deformable geometry as a sequence of frames;

b) determining, for a predetermined frame, global illumination values for each of a plurality of polygons defining a control mesh representative of said deformable geometry, wherein, for a predetermined frame of said sequence of frames, said global illumination values are computed from the frame-to-frame coherent global illumination values of said plurality of polygons determined for a prior frame of said sequence of frames as the sum of positive contributions of radiated and reflected light and negative contributions of occluded light directly incident on each said polygon in accordance with

9. The method of claim 8 wherein said single pass computation further provides for a directionally dependent summation of said positive and negative contributions, said global illumination value for said predetermined polygon including a plurality of directional illumination values, wherein each directional illumination value includes a set of color component values representing a defined color model.

10. The method of claim 9 wherein said plurality of directional illumination values includes quadrant components L1, L2, L3, and L4 of said global illumination value, and where in said quadrant components are computed, repeated for each included color component value of said set of color component values, as

11. The method of claim 10 further comprising the step of rendering said predetermined polygon of said predetermined frame as a plurality of pixels, wherein said step of rendering includes determining a pixel shading intensity for each pixel of sold plurality of pixels, and wherein said pixel shading intensity is computed as

I=L1*s1+L2*s2+L3*s3+L4*s4

where s1, s2, s3, and s4 are surface visibility values representing the directionally relative high-resolution texture dependent occlusion of the respective pixels of said plurality of pixels.