TW487890B - Apparatus and method for pixel data compression during rendering in computer graphics - Google Patents

Abstract

An apparatus and method for pixel data compression during rendering in computer graphics is proposed. The method is to divide the image frame into a plurality of blocks and compute those blocks covered by a rendering triangle. If a block is not totally covered by the triangle, the method will read and decompress the block for reference. Then, the system will render the blocks covered by the triangle and compress each block. At last, the system stores the compressed blocks into memory. The compression method is first to compute a plurality of initial seed colors for clustering the block of pixels. Then, each pixel within the block will be classified into groups with the corresponding initial seed colors. Those pixels with the same initial seed color are averaged to become a new final seed color. Therefore, the coded data comprise the index table and the final seed colors. Due to the that the compression method is simple, the system can compress the block real-time, therefore the memory bandwidth and storage can be reduced.

Description

487890 V. Description of the invention (1) [Explanation of the invention] [Field of the invention] The present invention relates to a drawing method and device for a computer graphics system, and in particular to a method and device for compressing and storing drawn data in memory in real time. /, [Knowledge technology] The use of green graphics (graphics) or animation (animation) needs to grow. For example, computer-generated images or animation are used in the fields of computer games, educational tools, office automation of film automation, or film making. As the capabilities of desktop computers continue to improve, many applications can be developed and leverage this enhanced functionality, and more interesting and interactive graphics interfaces have been created. However, despite the continuous improvement of desktop computer functions, 3D graphics objects and especially 3D moving objects still require expensive and special purpose hardware to achieve. … Today's real-time rendering system (rendering s s s t em) is costly due to the large amount of special memory and the processing resources needed to achieve visual realism. In order to achieve the real interaction and realistic effect, the system must depict the second object in a short time to generate a new image, so that the user feels that the object is continuously moving. ^ Out of the second image, the system must also respond to use with the shortest delay This means that the response of the input can be shown in the displayed image in real time. In most drawing and depicting systems, the surface of the drawing object is modeled and enlarged by the plane polygon in

487890 V. Description of invention (2). For example, color (c ο 1 〇 r) and texture (t e X tu r e) data are stored at the vertices of the polygon (vert ices). The process of generating each frame of the image data includes many steps. First, the system must work out which objects are potentially visible (ρ 〇 t e n t i a 1 1 y v i s i b 1 e). Second, the system rasterizes or scan converts pixel data into an array of pixel data. Finally, the system transfers the rendered display image to the display device. Rate (rate rate pipeline (ρ rate) is the so-called display device day rate is at least per fan rate. However, the image data in China according to the above #, the buffer can reduce the picture frame 0 f display ipel ine) calculate the rate (c The surface update rate is 60 frames per second. In order to make the expected, the conventional description of real-time drawing cost is from the need to remember the buffer, there are two different methods to generate the new frame buffer (frame omputat). i ona1 rate). In order to avoid the flicker of the day and the calculation rate does not need to be generated, the new image data structure includes at least two sets of mapping systems, which require a large amount of frame memory capacity and bandwidth. The presentation shows that one is the speed of the graphics buffer. The other is that the day-to-day update rate is equal to the day-to-day update and will not affect the display frame buffer. Buffer 'and plan. Therefore, if the system of drawing is [Summary of Invention]. In view of the above-mentioned period, the drawing is given to ± n Jing 'The purpose of the present invention is to propose a method for drawing in computer ^ ^ real-time compression of pixel data to reduce frame buffer Device requirements and equipment. Method for compressing pixel data when drawing I I Μ I f brain map

487890

The data of the I ter block is stored in all the colors marked with a number of pixels and recorded. Example] The corners are not merged and decompressed. The memory of each area is decompressed. And the pixel data of each data is a fixed value. The seed angle is fully contracted for block application. The seed color material is compressed and marked. The data closest to each of the same markers and each standard [covered in the real part, reference; will be compressed; the most method is to calculate the best color for the initial color corresponding to the mark; and after the triangle is compressed, the triangle will be compressed from the memory triangle. According to the value of the block, the seed color gives each image near the mark with the information of each image. The volume that has been read has been pixel-delineated. The reduced area is pixel data. The designated pixels are marked according to all the pixels. The following diagram illustrates the method and device for compressing pixel data when computer graphics are drawn. As shown in the figure, the buffer and the drawing receive the corners of the drawing instruction and are stored in the drawing of the Ming. Compression sheet Each additional display does not have the architecture of a computer graphics system for compression and decompression. The illustrated computer graphics system generally includes a host 20, a frame 30, a display device 40, and a graphics engine 10. The graphics engine 10 calculates which graphics objects are potentially visible from the host 20 graphics commands, and stores the graphics commands as an array of pixel data and stores them in the frame buffer 3. , S, and the display image of the frame buffer 30 is transferred to the display device 40. The rendering engine 10 includes an instruction queue 11, a rendering control unit 12: element 13, and a decompression unit 14. The drawing control unit 12 calculates an array of pixels drawn by the angles corresponding to the area of the frame buffer

Page 8 487890 V. Description of Invention (4) Block. The compression unit 1 3 compresses each block individually and stores it in the frame buffer 30. The decompression unit 14 decompresses the block data stored in the frame buffer 30 into an array of pixel data. Since the array of each polygonal day-time pixel data is compressed first and then stored in the frame buffer, the capacity of the frame buffer can be reduced. Figure 2 shows an example of a schematic diagram of a triangle corresponding to a block. When using the drawing method of the present invention, the frame (fr ame) is divided into a plurality of blocks of the same size, and the pixels of the triangle A are drawn according to the drawing. Position to calculate the block where the triangle A is distributed. As shown in the figure, the triangle A is distributed in 15 blocks, which are labeled B 1 to B 1 5 respectively, and only the block B 7 is completely covered by the triangle A. After calculating the blocks distributed by triangle A, each block is processed individually in block units, and the compressed data is stored in memory. Since the block B7 is completely covered by the triangle A, it is not necessary to refer to other block data, so the system can directly compress the block B 7 and save it to the memory. However, for other blocks such as B1, B2, B3, ..., only a part of the pixel data of the triangle A is included. Because the blocks are compressed as a whole, before processing and compressing these blocks, you must read other data of these blocks that are not covered by triangle A. That is, the data of these blocks must be read from the memory and decompressed before further processing. Figure 3 shows an example of a flow chart of the present invention. As shown in the figure, the drawing steps of the polygon (triangle) of the present invention are as follows: Step S3 0 2: Start drawing a triangle; Step S3 0 4: Whether the drawing instruction triangle needs to be drawn, also

487890 V. Description of the invention (5) That is, if the triangle is the block covered by step S 3 0 6 in the next step, step S308 step S31 0 if necessary, go to step S3 1 2 step S 3 1 4 step S 3 1 6 Memory; Step S3 1 8 After the processing is completed, Step S 3 2 0 because the drawing must be slow when decompressing, then the compression method is affected instead. Step S402 Step S4 0 4 Step S 4 0 6 Step S4 0 8 Step S410 Lottery and specify whether it is visible? If it does not need to be drawn, skip to step S3 2 0, NO; calculate the triangle according to the pixel position of the triangle: process the next block data;: do you need to store the block data in the memory? If step S312, otherwise skip to step S314; read the block data from the memory and decompress it; draw and update the block data; compress and save the updated block data to the record: whether each Have all the data of the block been processed? If it does not jump back to step S 3 0 8, otherwise skip to step S 3 2 0;: End the drawing operation of the triangle. The amount of data processing required is quite large, so compression and execution speed are taken into account. If the speed of compression and decompression is too great for the performance of the system. The following describes the beginning of the present invention with reference to FIG. 4; reading pixel array data of a block; calculating an initial value seed color and setting a flag; processing the next pixel; calculating an initial value seed color and an initial value seed color closest to the pixel data The mark is the mark of the pixel data;

Page 10 V. Description of the invention (6) Step S412: Is the image of all blocks like the seed color mark? If you have n’t deducted it all yet ... head ^ ^ 3 to deduct an initial value and skip to step S414 ;,. Sundial J trillion to step S408, otherwise step S414 step S41 6 step S41 8, newly calculate the fixed value seed π fixed value seed color and pixel data of each mark ... Step S40 6 The method of calculating the initial value seed color intentionally selects a block In the portrait series, you can also calculate Dagger = the average value of color. Figure 5 (A) shows that the pixel data of a block is a 4 * 4 matrix. If the initial value is calculated as the average value of the second type ... The color of the zongzi and its mark, the mark and the initial value, and the sub-color centers are 0 (95), 1 (37), 2 (57), and 3 (27) . Next, the ratio data and the four initial value seed colors are assigned, and the target color of the initial value seed color close to the 忒 -pixel-shell material is designated. Finally, according to the legend, the method is to select the same markers in the block, prime, and calculate the color data closest to those pixels. There are many methods to calculate the seed color of the Zeng value. For example, the root-mean-square value or the average pound size = a pixel matrix of 4 * 4 blocks of 32 bits is taken as an example. The size of the uncompressed money is 512 (32 * 16) bit (blt). But after compression, the data size of the 4 * 4 block is ^ 8 (2 ^ 16 + 204) bits, that is, 16 marker data = a fixed value seed color, as shown in Figure 5 (C). Therefore, the stem shrinking method of the present invention can achieve 4 times the I shrinkage. Not only can reduce the diagram of the edge map system 487890 V. Description of the invention (7) Use of the frame buffer, but also reduce the memory bandwidth Figure 6 shows a flowchart of a decompression embodiment: Step S 6 0 2 Step S 6 0 4 step S 6 0 6 step S 6 0 8 step S 6 1 0 color is given to the pixel step S 6 1 2 step S614 step S 6 1 6 has not been fully specified. The steps are described as starting to decompress a block of data; read Take a compressed block data; Separate the marker data from the fixed-value seed color; Process the next pixel; Specify the corresponding fixed-value seed according to the pixel's tag value Are all pixels with the specified fixed-value seed color? If yes, go back to step S608, otherwise go to step S614; output the pixel data of the block; end. Since the above-mentioned decompression method only assigns a corresponding fixed-value seed color to the pixel according to the tag value of the pixel, no other complicated calculation is required, so it can provide real-time decompression. Although the present invention has been described by way of examples, the scope of the present invention is not limited thereby, and those skilled in the art can make various modifications or changes without departing from the gist of the present invention.

Page 12 487890 Brief Description of Drawings [Brief Description of Drawings] Figure 1 shows the architecture of a computer graphics system with compression and decompression in accordance with the present invention. FIG. 2 shows a schematic flowchart of the present invention. Figure 3 shows a schematic diagram of a pixel depicted in the present invention that does not include the entire block. FIG. 4 shows a compression flowchart of the present invention. Figure 5 shows the data compression of a block. FIG. 6 shows a decompression flowchart of the present invention. [Pattern number] 10 Drawing engine 11 Instruction queue 12 Drawing controller 13 Compression unit 14 Decompression unit 20 Host 30 Memory 40 Display device

Claims (1)

487890 VI. Application for Patent Scope 1. A method for compressing pixel data during computer drawing and drawing, which includes the following steps: executing drawing instructions to generate graphic elements; determining whether the foregoing graphic elements need to be drawn, and ending if no drawing is needed, Otherwise, continue with the following steps; calculate the block where the graphic element is distributed according to the pixel position of the aforementioned graphic element; decompress the 绾 block step, if each of the aforementioned blocks needs to refer to the tribute of the compressed block stored in the aforementioned memory When the data is read, the data of the compressed block is read from the foregoing memory and decompressed; the drawing step is to draw the pixels of the graphic element in the block; the compression step is to compress and store the parent block individually. In step, the data of the compressed block is stored in the memory. 2. The method for compressing pixel data in computer graphics as described in item 1 of the scope of the patent application, wherein the aforementioned compression step includes: calculating an initial seed color, specifying a plurality of marks according to the pixel data of the block, and the mark location Corresponding initial value seed color; Designate the mark, assign the mark of the aforementioned seed color closest to each pixel data to each pixel; Calculate the fixed value seed color, and calculate all the closest to the mark based on the pixel data of the same mark The color of the pixel data is used as the fixed seed color; and the data is sent, with the mark of each pixel and the aforementioned seed of each mark Page 14 487890 VI. Scope of patent application Color is used as compressed information. 3. The method for compressing pixel data during computer graphics as described in item 2 of the scope of patent application, where the aforementioned block is a 4 * 4 pixel matrix. 4. The method of compressing pixel data during computer meeting drawing as described in item 3 of the scope of patent application, wherein the aforementioned block contains four different marks and corresponding four fixed-value seed colors. '5. A drawing engine of a computer drawing system, which performs block compression before image data is stored in memory. The drawing engine includes: a command line. It stores drawing commands; a drawing controller receives the aforementioned A drawing instruction, and calculates a drawing element of the drawing instruction and all blocks covered by the drawing element; a compression unit that compresses the block pixel data output by the aforementioned drawing controller into block data; and a decompression The unit is to decompress a compressed block data into block pixel data; thus, before the block pixel data is stored in the memory, the block data is compressed and stored by the compression unit, and stored by the memory. The block data read from the volume is first decompressed into block pixel data by the aforementioned decompression unit, and then further processed. 6. The drawing guide of the computer drawing system described in item 5 of the scope of patent application Page 15 487890 6. The scope of the patent application, wherein the aforementioned compression unit includes: a seed color generation unit, which designates a plurality of initial value seed colors and their corresponding marks according to the aforementioned block pixel data; a mark designation unit, which designates The pixel color closest to the initial value of the seed color is given to each pixel; the seed color update unit calculates and calculates the data of the pixel color closest to the mark as the fixed value seed color based on the pixel color of the same mark; and the data sending unit Using the mark of each pixel and the aforementioned fixed value seed color of each mark as the compressed data. · 7. The method of compressing pixel data during computer drawing and drawing as described in item 6 of the scope of patent application, where the aforementioned block is a 4 * 4 pixel matrix. 8. The method of compressing pixel data during computer graphics as described in item 7 of the scope of the patent application, wherein the aforementioned block contains four different marks and corresponding four fixed-value seed colors. Page 16