Methods for compressing and decompressing texture tiles and apparatuses using the same

Abstract

The invention introduces a method for compressing texture tiles, which contains at least the following steps: lossless-compressing raw data of a texture tile; determining whether a length of the lossless-compression result of the raw data is greater than a target result; and when the length of the lossless-compression result of the raw data is greater than the target length, performing data-reduction control in layers for generating reduced data by reducing the raw data, and generating a lossless-compression result of the reduced data, thereby enabling the length of the lossless-compression result of the reduced data to be equal to the target length or shorter.

What is claimed is: 
     
         1 . A method for compressing texture tiles, comprising:
 lossless-compressing raw data of a texture tile;   determining whether a length of a lossless-compression result of the raw data is greater than a target result; and   when the length of the lossless-compression result of the raw data is greater than the target length, performing data-reduction control in layers for generating reduced data by reducing the raw data, and generating a lossless-compression result of the reduced data, thereby enabling a length of the lossless-compression result of the reduced data to be equal to the target length or shorter.   
     
     
         2 . The method of  claim 1 , wherein the data-reduction control in layers damages information of the raw data. 
     
     
         3 . The method of  claim 1 , wherein the data-reduction control in layers generates the reduced data by reducing the raw data of the texture tile in the minimum level. 
     
     
         4 . The method of  claim 1 , wherein the step of performing data-reduction control in layers comprises:
 reducing the raw data to a degree to generate first reduced data;   lossless-compressing the first reduced data;   when a length of a lossless-compression result of the first reduced data is longer than the target length and a total number of data reductions does not exceed a predefined value, reducing the raw data to a greater degree to generate second reduced data; and   when the length of the lossless-compression result of the first reduced data is longer than the target length and the total number of data reductions exceeds the predefined value, forcibly reducing the raw data to the target length or shorter, without performing a lossless compression.   
     
     
         5 . The method of  claim 4 , wherein the first reduced data or the second reduced data is generated using Equations below:
     R   i   =R   i   /m   j   (1)
       G   i   =G   i   /m   j   (2)
       B   i   =B   i   /m   j   (3)
   wherein R i  represents an R-value of the i th  pixel of the raw data, G i  represents a G-value of the i th  pixel of the raw data, B i  represents a B-value of the i th  pixel of the raw data, m j  represents a constant for the j th  data reduction and a greater j comes a greater m j .   
     
     
         6 . The method of  claim 5 , wherein a length of the raw data is 2048 bits, the target length is 1024 bits, the predefined value is 2, m 1  is set to 2, m 2  is set to 3 and m 3  is set to 8. 
     
     
         7 . The method of  claim 5 , wherein a length of the raw data is 2048 bits, the target length is 1024 bits, the predefined value is 3, m 1  is set to 2, m 2  is set to 3, m 3  is set to 8 and m 4  is set to 16. 
     
     
         8 . The method of  claim 1 , comprising:
 outputting a lossy-level flag corresponding to the reduced data, which indicates that the raw data is forcibly reduced to the target length or shorter, or the raw data has undergone data reduction in a particular layer.   
     
     
         9 . An apparatus for compressing texture tiles, comprising:
 a memory controller;   a compressor;   a subtractor; and   a data-reduction controller, coupled to the memory controller, the compressor and the subtractor, directing the memory controller to output raw data of a texture tile to the compressor, thereby enabling the compressor to generate a lossless-compression result of the raw data; determining whether a length of the lossless-compression result of the raw data is greater than a target result; and when the length of the lossless-compression result of the raw data is greater than the target length, performing data-reduction control in layers for directing the subtractor to generate reduced data by reducing the raw data, and directing the memory controller to output the reduced data to the compressor, thereby enabling the compressor to generate a lossless-compression result of the reduced data, wherein a length of the lossless-compression result of the reduced data equals the target length or shorter.   
     
     
         10 . The apparatus of  claim 9 , wherein the data-reduction control in layers damages information of the raw data. 
     
     
         11 . The apparatus of  claim 9 , wherein the data-reduction control in layers generates the reduced data by reducing the raw data of the texture tile in the minimum level. 
     
     
         12 . The apparatus of  claim 9 , wherein, in the data-reduction control in layers, the data-reduction controller directs the subtractor to reduce the raw data to a degree to generate first reduced data, and directs the memory controller to output the first reduced data to the compressor, thereby enabling the compressor to generate a lossless-compression result of the first reduced data; when a length of the lossless-compression result of the first reduced data is longer than the target length and a total number of data reductions does not exceed a predefined value, directs the subtractor to reduce the raw data to a greater degree to generate second reduced data; and, when the length of the lossless-compression result of the first reduced data is longer than the target length and the total number of data reductions exceeds the predefined value, directs the subtractor to forcibly reduce the raw data to the target length or shorter, without performing a lossless compression. 
     
     
         13 . The apparatus of  claim 12 , wherein the first reduced data or the second reduced data is generated using Equations below:
     R   i   =R   i   /m   j ;       G   i   =G   i   /m   j ; and       B   i   =B   i   /m   j ,   wherein R i  represents an R-value of the i th  pixel of the raw data, G i  represents a G-value of the i th  pixel of the raw data, B i  represents a B-value of the i th  pixel of the raw data, m j  represents a constant for the j th  data reduction and a greater j comes a greater m j .   
     
     
         14 . The apparatus of  claim 13 , wherein a length of the raw data is 2048 bits, the target length is 1024 bits, the predefined value is 2, m 1  is set to 2, m 2  is set to 3 and m 3  is set to 8. 
     
     
         15 . The apparatus of  claim 13 , wherein a length of the raw data is 2048 bits, the target length is 1024 bits, the predefined value is 3, m 1  is set to 2, m 2  is set to 3, m 3  is set to 8 and m 4  is set to 16. 
     
     
         16 . The apparatus of  claim 9 , wherein the data-reduction controller outputs a lossy-level flag corresponding to the reduced data, which indicates that the raw data is forcibly reduced to the target length or shorter, or the raw data has undergone data reduction in a particular layer. 
     
     
         17 . A method for decompressing texture tiles, comprising:
 receiving a compression result and a lossy-level flag;   decompressing the compression result to generate decompressed data;   determining whether the compression result has undergone data reduction according to the lossy-level flag;   when the compression result has undergone the data reduction, compensating the decompressed data to generate a first restored result; and   outputting the first restored result to a texture operation of a 3D (three-dimensional) pipeline.   
     
     
         18 . The apparatus of  claim 17 , wherein the first restored result is generated using Equations below:
     R   i   =R   i   *m   j   +c   j ;       G   i   =G   i   *m   j   +c   j ; and       B   i   =B   i   *m   j   +c   j ,   wherein R i  represents an R-value of the i th  pixel of the decompressed data, G i  represents a G-value of the i th  pixel of the decompressed data, B i  represents a B-value of the i th  pixel of the decompressed data, m j  represents a restoring constant for the j th  reduction, a greater j comes a greater m j , c j  represents a compensation constant for the j th  reduction, and a greater j comes a greater c j  and c j  is less than c max .   
     
     
         19 . The apparatus of  claim 17 , comprising:
 determining whether raw data of a texture tile is forcibly reduced to a target length or shorter;   when the raw data of the texture tile is forcibly reduced to the target length or shorter, compensating the compression result to generate a second restored result; and   when the raw data of the texture tile is not forcibly reduced to the target length or shorter, decompressing the compression result.   
     
     
         20 . The apparatus of  claim 19 , wherein the second restored result is generated using Equations below:
     R   i   =R   i   *m   max   +c   max ;       G   i   =G   i   *m   max   +c   max ; and       B   i   =B   i   *m   max   +c   max ,   wherein R i  represents an R-value of the i th  pixel of the compression result, G i  represents a G-value of the i th  pixel of the compression result, B i  represents a B-value of the i th  pixel of the compression result, m max  represents a restoring constant and c max  represents a compensation constant.