Method and system for implementing compressed normal maps

• US 9,081,681 B1
• Filed: 12/19/2003
• Issued: 07/14/2015
• Est. Priority Date: 12/19/2003
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Donovan, Walter E.
• Primary Examiner(s): Hsu, Joni
• Application Number:
  US10/742,434
• Time in Patent Office: 4,225 Days
• Field of Search: 345/584, 345/582, 345/581, 345/418, 345/585, 345/543, 345/544, 345/555, 345/565, 345/564, 345/572, 345/530, 345/552
• US Class Current: 1/1
• CPC Class Codes: G06F 12/08 in hierarchically structured memory systems, e.g. virtual memory systems
• CPC Class Codes: G06T 15/005 General purpose rendering architectures
• CPC Class Codes: G06T 15/04 Texture mapping
• CPC Class Codes: G06T 9/00 Image coding bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00

Abstract

A method for compressing normal maps in a computer system. The method includes accessing a map of input normals. A memory block having a first portion and a second portion is defined. A table of indices is stored in the first portion of the memory block and a table of normals is stored in the second portion of the memory block. The indices of the first portion of the memory block reference the normals of the second portion. The normals in the second portion of the memory block are unit normals of a sphere defined to represent the map of input normals.

Claims

1. A method for compressing normal maps in a computer system, comprising:

accessing a map of input normals, said map of input normals comprising a plurality of normal vectors on a surface of an image;

defining a memory block having a first portion and a second portion;

storing a table of indices in the first portion of the memory block; and

storing a table of normals in the second portion of the memory block, the normals in the second portion of the memory block comprising a subset of normals out of a set of unit normals of a quantized unit sphere,

wherein the indices of the first portion of the memory block reference the normals of the second portion of the memory block,

wherein the quantized unit sphere is quantized into points and the subset of normals is derived from a subset of the points selected to represent the map of input normals,

further wherein the subset of the normals stored in the second portion of the memory block is selected by applying a compression algorithm to the set of unit normals of the quantized unit sphere according to a pre-defined metric for manipulating an average value corresponding to the unit normals.

2. The method of claim 1, wherein the memory block is a 64-byte memory block.

3. The method of claim 2, wherein the first portion is 256 bits and the second portion is 256 bits.

4. The method of claim 1, wherein the first portion of the memory block comprises a first table of 64 4-bit indices.

5. The method of claim 1, wherein the second portion of the memory block comprises a second table of 16 entries, and wherein each entry comprises a 16 bit encoded normal.

6. The method of claim 1, wherein 16 selected normals in the second portion are used to represent 64 input normals.

7. The method of claim 1, wherein the pre-defined metric for manipulating an average value corresponding to the unit normals comprises a metric to minimize the average angular deviation of the unit normals of the set of unit normals of the quantized unit sphere.

8. The method of claim 1, wherein the pre-defined metric for manipulating an average value corresponding to the unit normals comprises a metric to minimize the average error of the dot product of the unit normals of the set of unit normals of the quantized unit sphere.

9. A system for compressing normal maps in a graphics computer system, the system having a graphics processor unit (GPU) coupled to a memory, the memory having computer readable code which when executed by the GPU cause the GPU to implement a method comprising:

accessing a map of input normals, said map of input normals comprising a plurality of normal vectors on a surface of an image;

defining a memory block having a first portion and a second portion;

storing a table of indices in the first portion of the memory block; and

storing a table of normals in the second portion of the memory block, the normals in the second portion of the memory block comprising a subset of normals out of a set of unit normals of a quantized unit sphere,

wherein the indices of the first portion of the memory block reference the normals of the second portion of the memory block,

wherein the quantized unit sphere is quantized into points and the subset of normals is derived from a subset of the points selected to represent the map of input normals,

further wherein the subset of the normals stored in the second portion of the memory block is selected by applying a compression algorithm to the set of unit normals of the quantized unit sphere according to a pre-defined metric for manipulating an average value corresponding to the unit normals.

10. The system of claim 9, wherein the memory block is a 64-byte memory block.

11. The system of claim 10, wherein the first portion is 256 bits and the second portion is 256 bits.

12. The system of claim 9, wherein the first portion of the memory block comprises a first table of 64 4-bit indices.

13. The system of claim 9, wherein the second portion of the memory block comprises a second table of 16 entries, and wherein each entry comprises a 16 bit encoded normal.

14. The system of claim 9, wherein 16 selected normals in the second portion are used to represent 64 input normals.

15. The system of claim 9, wherein the pre-defined metric for manipulating an average value corresponding to the unit normals comprises a metric to minimize the average angular deviation of the unit normals of the set of unit normals of the quantized unit sphere.

16. The system of claim 9, wherein the pre-defined metric for manipulating an average value corresponding to the unit normals comprises a metric to minimize the average error of the dot product of the unit normals of the set of unit normals of the quantized unit sphere.

17. A non-transitory computer readable storage media having computer readable code, which when executed by a processor of a computer system cause the computer system to implement a method for compressing normal maps, comprising:

accessing a map of input normals, said map of input normals comprising a plurality of normal vectors on a surface of an image;

defining a 64-byte memory block having a first 256-bit portion and a second 256-bit portion;

storing a first table of 64 4-bit indices in the first portion of the memory block; and

storing a second table of 16 entries in the second portion of the memory block, wherein each entry comprises a 16 bit encoded normal,

wherein the indices of the first table reference the normals of the second table, the normals in the second table comprising a subset of a greater number of unit normals of a quantized unit sphere,

wherein the quantized unit sphere is quantized into points and the subset of normals is derived from a subset of the points selected to represent the map of input normals,

further wherein the subset of the normals stored in the second portion of the memory block is selected by applying a compression algorithm to the set of unit normals of the quantized unit sphere according to a pre-defined metric for manipulating an average value corresponding to the unit normals.

18. The non-transitory computer readable storage media of claim 17, wherein the first table comprises and 8×

8 table of 4-bit indices.

19. The non-transitory computer readable storage media of claim 17, wherein the second table comprises a 4×

4 table of 16 entries.

20. The non-transitory computer readable storage media of claim 17, wherein 16 selected normals in the second table are used to represent 64 input normals.

21. The non-transitory computer readable storage media of claim 20, wherein the 64 input normals are used to represent a plurality of possible unit normal vectors over a hemisphere.

22. The non-transitory computer readable storage media of claim 17, wherein the pre-defined metric for manipulating an average value correspond to the unit normals comprises a metric to minimize the average angular deviation of the unit normals of the set of unit normals of the quantized unit sphere.

23. The non-transitory computer readable storage media of claim 17, wherein the pre-defined metric for manipulating an average value corresponding to the unit normals comprises a metric to minimize the average error of the dot product of the unit normals of the set of unit normals of the quantized unit sphere.