Multi-mode texture compression algorithm

• US 6,959,110 B1
• Filed: 08/13/2001
• Issued: 10/25/2005
• Est. Priority Date: 08/17/2000
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Tarolli, Gary M., Danskin, John M., Sundaresan, Murali
• Primary Examiner(s): Patel, Kanjibhai
• Assistant Examiner(s): Bayat, Ali
• Application Number:
  US09/929,427
• Time in Patent Office: 1,534 Days
• Field of Search: 382/232, 382/233, 382/304, 382/162, 382/166, 725/109, 345/555, 345/582, 345/619, 345/505, 345/585, 345/419, 345/473, 345/502, 345/418, 345/426, 348/207.1
• US Class Current: 382/166
• CPC Class Codes: H04N 19/12 Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
• CPC Class Codes: H04N 19/136 Incoming video signal characteristics or properties
• CPC Class Codes: H04N 19/186 the unit being a colour or a chrominance component
• CPC Class Codes: H04N 19/426 using memory downsizing methods
• CPC Class Codes: H04N 19/428 Recompression, e.g. by spatial or temporal decimation

Abstract

A multi-mode texture compression algorithm is provided for effective compression and decompression texture data during graphics processing. Initially, a request is sent to memory for compressed texture data. Such compressed texture data is then received from the memory in response to the request. At least one of a plurality of compression algorithms associated with the compressed texture data is subsequently identified. Thereafter, the compressed texture data is decompressed in accordance with the identified compression algorithm.

Claims

1. A multi-mode texture decompression method for use during graphics processing, comprising;

(a) sending a request for compressed texture data to memory;

(b) receiving the compressed texture data from the memory;

(c) identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(d) decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein, prior to sending the request, the texture data is compressed utilizing each of the plurality of compression algorithms, the most favorable compressed texture data is selected utilizing a comparison operation, and the most favorable compressed texture data is stored in the memory;

wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

2. The method as recited in claim 1, and further comprising storing a mode identifier with the compressed texture data.

3. The method as recited in claim 2, wherein the at least one of the plurality of compression algorithms associated with the compressed texture data is identified utilizing the mode identifier.

4. The method as recited in claim 3, wherein the mode identifier includes a mode bit.

5. The method as recited in claim 1, wherein at least one of the compression algorithms represents a 4×

4 block of texels of the texture data utilizing two bits per texel only if the texels are opaque, each 4×

4 block of texels including two 16-bit colors stored in an RGB 565 format and two additional colors created by interpolating between the two 16-bit colors stored in the RGB 565 format to form a 4-entry lookup table, where a 2-bit index is adapted for being used to determine which 16-bit color from the lookup table is used for each texel of the 4×

4 block of texels, and transparent texels are represented by making one of the four 16-bit colors transparent.

6. The method as recited in claim 1, wherein at least one of the compression algorithms represents a 4×

8 block of texels utilizing two bits per texel only if the texels are opaque, each 4×

8 block of texels including four 15-bit colors in an RGB 555 format to form a 4-entry lookup table, a 2-bit index is adapted for being used to determine which of the four 15-bit colors is assigned to each texel.

7. The method as recited in claim 1, wherein the most favorable compressed texture data is selected by comparing errors associated with each of the compression algorithms.

8. The method as recited in claim 7, wherein the errors are calculated after decompression of the texture data that occurs after the compression thereof.

9. A multi-mode texture decompression method for use during graphics processing, comprising:

(a) sending a request for compressed texture data to memory;

(b) receiving the compressed texture data from the memory;

(c) identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(d) decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein at least one of the compression algorithms represents a 4×

8 block of texels utilizing three bits per texel, each 4×

8 block of texels including two 15-bit colors stored in an RGB 555 format and five additional colors created by interpolating between the two 15-bit colors stored in the RGB 555 format to form an 8-entry lookup table, where an eighth 15-bit color is defined to be a transparent color, and a 3-bit index is used to determine which 15-bit color from the lookup table is used for each texel in the 4×

8 block of texels.

10. A multi-mode texture decompression method for use during graphics processing, comprising:

(a) sending a request for compressed texture data to memory;

(b) receiving the compressed texture data from the memory;

(c) identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(d) decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein at least one of the compression algorithms represents a 4×

8 block of texels by two bits per texel, each 4×

8 block of texels including three 20-bit colors stored in a 5555 format, where a first and second one of the 20-bit colors are used for primary colors of a left 4×

4 sub-block of the 4×

8 block of texels, and a second and third one of the colors are used for primary colors of the right 4×

4 sub-block of the 4×

8 block of texels, where two additional 20-bit colors are created in each 4×

4 sub-block of texels by interpolating between the 20-bit colors associated with the corresponding 4×

4 sub-block of texels, where a 2-bit index is adapted for being used to determine which of the four 20-bit colors is assigned to each texel, and a lookup table is used to determine which 20-bit color is applied to each texel.

11. A multi-mode texture decompression computer program product embodied on a computer readable medium for use during graphics processing, comprising;

(a) computer code for sending a request for compressed texture data to memory;

(b) computer code for receiving the compressed texture data from the memory;

(c) computer code for identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(d) computer code for decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein, prior to sending the request, the texture data is compressed utilizing each of the plurality of compression algorithms, the most favorable compressed texture data is selected utilizing a comparison operation, and the most favorable compressed texture data is stored in the memory;

wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

12. The computer program product as recited in claim 11, wherein the most favorable compressed texture data is selected by comparing errors associated with each of the compression algorithms.

13. The computer program product as recited in claim 12, wherein the errors are calculated after decompression of the texture data that occurs after the compression thereof.

16. The system as recited in claim 12, wherein at least one of the compression algorithms represents a 4×

8 block of texels utilizing two bits per texel only if the texels are opaque, each 4×

8 block of texels including four 15-bit colors in an RGB 555 format to form a 4-entry lookup table, a 2-bit index is adapted for being used to determine which of the four 15-bit colors is assigned to each texel.

14. A multi-mode texture decompression system for use during graphics processing, comprising;

(a) a texture fetch module adapted for sending a request for compressed texture data to memory, and receiving the compressed texture data from the memory;

(b) a format detection module adapted for identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(c) a plurality of decompression modules coupled between the texture fetch module and the format detection module, the decompression modules adapted for decompressing the compressed texture data in accordance with the compression algorithm identified by the format detection module;

wherein, prior to sending the request, the texture data is compressed utilizing each of the plurality of compression algorithms, the most favorable compressed texture data is selected utilizing a comparison operation, and the most favorable compressed texture data is stored in the memory;

wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

15. The system as recited in claim 14, wherein at least one of the compression algorithms represents a 4×

4 block of texels of the texture data utilizing two bits per texel only if the texels are opaque, each 4×

4 block of texels including two 16-bit colors stored in an RGB 565 format and two additional colors created by interpolating between the two 16-bit colors stored in the RGB 565 format to form a 4-entry lookup table, where a 2-bit index is adapted for being used to determine which 16-bit color from the lookup table is used for each texel of the 4×

4 block of texels, and transparent texels are represented by making one of the four 16-bit colors transparent.

17. The system as recited in claim 14, wherein the most favorable compressed texture data is selected by comparing errors associated with each of the compression algorithms.

18. The system as recited in claim 17, wherein the errors are calculated after decompression of the texture data that occurs after the compression thereof.

19. A multi-mode texture decompression system for use during graphics processing, comprising:

(a) a texture fetch module adapted for sending a request for compressed texture data to memory, and receiving the compressed texture data from the memory;

(b) a format detection module adapted for identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(c) a plurality of decompression modules coupled between the texture fetch module and the format detection module, the decompression modules adapted for decompressing the compressed texture data in accordance with the compression algorithm identified by the format detection module;

wherein at least one of the compression algorithms represents a 4×

8 block of texels utilizing three bits per texel, each 4×

8 block of texels including two 15-bit colors stored in an RGB 555 format and five additional colors created by interpolating between the two 15-bit colors stored in the RGB 555 format to form an 8-entry lookup table, where an eighth 15-bit color is defined to be a transparent color, and a 3-bit index is used to determine which 15-bit color from the lookup table is used for each texel in the 4×

8 block of texels.

20. A multi-mode texture decompression system for use during graphics processing, comprising:

(a) a texture fetch module adapted for sending a request for compressed texture data to memory, and receiving the compressed texture data from the memory;

(b) a format detection module adapted for identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(c) a plurality of decompression modules, coupled between the texture fetch module and the format detection module, the decompression modules adapted for decompressing the compressed texture data in accordance with the compression algorithm identified by the format detection module;

wherein at least one of the compression algorithms represents a 4×

8 block of texels by two bits per texel, each 4×

8 block of texels including three 20-bit colors stored in a 5555 format, where a first and second one of the 20-bit colors are used for primary colors of a left 4×

4 sub-block of the 4×

8 block of texels, and a second and third one of the colors are used for primary colors of the right 4×

4 sub-block of the 4×

8 block of texels, where two additional 20-bit colors are created in each 4×

4 sub-block of texels by interpolating between the 20-bit colors associated with the corresponding 4×

4 sub-block of texels, where a 2-bit index is adapted for being used to determine which of the four 20-bit colors is assigned to each texel, and a lookup table is used to determine which 20-bit color is applied to each texel.

21. A multi-mode texture decompression system for use during graphics processing, comprising;

(a) means for sending a request for compressed texture data to memory;

(b) means for receiving the compressed texture data from the memory;

(c) means for identifying at least one of a plurality of compression algorithms associated with the compressed texture data; and

(d) means for decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein, prior to sending the request, the texture data is compressed utilizing each of the plurality of compression algorithms the most favorable compressed texture data is selected utilizing a comparison operation, and the most favorable compressed texture data is stored in the memory;

wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

22. A multi-mode texture compression method for use during graphics processing, comprising;

(a) compressing texture data utilizing each of a plurality of compression algorithms in parallel;

(b) selecting the most favorable compressed texture data using a comparison operation;

(c) storing the most favorable compressed texture data in memory;

(d) storing a mode bit with the most favorable compressed texture data in the memory;

(e) sending a request for the compressed texture data to the memory;

(f) receiving the compressed texture data from the memory;

(g) determining the mode bit associated with the received compressed texture data;

(h) identifying at least one of the plurality of compression algorithms associated with the compressed texture data based on the mode bit; and

(i) decompressing the compressed texture data in accordance with the identified compression algorithm;

wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

23. A multi-mode texture compression method for use during graphics processing, comprising;

(a) compressing texture data utilizing each of a plurality of compression algorithms in parallel;

(b) selecting the most favorable compressed texture data using a comparison operation;

(c) storing the most favorable compressed texture data in memory; and

(d) storing a mode bit with the most favorable compressed texture data in the memory;

(e) wherein the mode bit associated with the received compressed texture data is capable of being used to identify at least one of the plurality of compression algorithms associated with the compressed texture data such that the compressed texture data is capable of being decompressed in accordance with the identified compression algorithm;

(f) wherein the most favorable compressed texture data provides the most accurate replication of an original version of the texture data.

24. A data structure stored in memory for compressing texture data representing a Y×

Z block of texels utilizing three bits per texel, each Y×

Z block of texels including two X-bit colors stored in a predetermined format and five additional colors created by interpolating between the two X-bit colors stored in the predetermined format to form a lookup table, where an eighth X-bit color is defined to be a transparent color, and a W-bit index is used to determine which X-bit color from the lookup table is used for each texel in the Y×

Z block of texels, wherein the compressed texture data is used during graphics processing.

25. A data structure stored in memory for compressing texture data representing a Y×

Z block of texels utilizing two bits per texel only if the texels are opaque, each Y×

Z block of texels including four X-bit colors in a predetermined format to form a lookup table, a W-bit index is adapted for being used to determine which of the four X-bit colors is assigned to each texel, wherein the compressed texture data is used during graphics processing.

26. A data structure stored in memory for compressing texture data representing a Y×

Z block of texels by two bits per texel, each Y×

Z block of texels including three X-bit colors stored in a predetermined format, where a first and second one of the X-bit colors are used for primary colors of a left Y×

Y sub-block of the Y×

Z block of texels, and a second and third one of the colors are used for primary colors of the right Y×

Y sub-block of the Y×

Z block of texels, where two additional X-bit colors are created in each Y×

Y sub-block of texels by interpolating between the X-bit colors associated with the corresponding Y×

Y sub-block of texels, where a W-bit index is adapted for being used to determine which of the four X-bit colors is assigned to each texel, and a lookup table is used to determine which X-bit color is applied to each texel, wherein the compressed texture data is used during graphics processing.