Computer graphics processing and selective visual display system – Computer graphics processing – Three-dimension
Reexamination Certificate
2000-08-02
2003-05-27
Zimmerman, Mark (Department: 2671)
Computer graphics processing and selective visual display system
Computer graphics processing
Three-dimension
Reexamination Certificate
active
06570565
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a 3 dimensional (D) graphic accelerator and a method for processing a graphic acceleration using the same.
2. Description of the Related Art
In general, 3D computer graphics is a core part of constructing an environment of multimedia. In order to assist in a vivid 3D image, however, an exclusive 3D graphic accelerator of high performance is required. Highly efficient 3D graphic accelerators are introduced these days to PCs and game devices, and active studies are under way for 3D graphic accelerators.
FIG. 1
is a block diagram illustrating the processing steps of the 3D graphics. Referring to
FIG. 1
, the processing steps of the 3D graphics comprises a real-time hardware acceleration by a 3D application software in a 3D graphic accelerator through an application program interface (API), and a transfer of the result to the display.
The above 3D graphic accelerator mainly functions as geometry processing and rendering. The geometry processing is a process of converting an object on a 3D coordinate into a viewpoint, and projecting the viewpoint onto a 2D screen coordinate. The rendering is a process of determining a color value of the 2D coordinate and storing the value in a frame buffer. The color value stored in the frame buffer after processing all the 3D data inputted to one frame is transferred to a display, which is called a “display refresh.” In general, a geometry processing section and a rendering section are pipelined to enhance their performance.
The 3D graphic accelerators are roughly divided into two styles of following an order of inputted primitives, i.e., an object-order style and an image-order style.
The primitives here refer to dots, lines and polygons. In most of the applicable fields in general, polygons occupy most of the primitives. A hardware accelerator is constructed to process the polygons at high speed.
The object-order style is a style that the objects are transferred to a display in the order of primitives after undergoing the geometrical process and rendering process. The object-order style is advantageous for high performance because it can pipeline the geometrical process and the rendering process performed per primitive.
For a hidden surface removal, however, the object-order style needs to have a depth buffer (z-buffer) and a color-buffer corresponding to a full screen. For an overlap of the rendering process and the display refresh, two z-buffers and two color buffers are required. This is called a double buffering. Most of the 3D graphic accelerators currently being introduced to public perform the processing in the object-order style.
The image-order style is not a style of processing the image in the order of primitives but of processing the primitives at the corresponding positions of the images in the order of the positions.
Assuming that the start point of the screen is (0, 0) and the end point is (n−1, m−1), for instance, all of the related primitives at the position of (0, 0) are searched for calculation of color values. The same process is repeated in the given order until reaching to the position of (n−1, m−1). This necessitates a buffer that has a pipeline of the geometrical process and the rendering process for entire primitives and all the information on the geometrically processed primitives. Since only a part of the screen rather than a full screen is required for the hidden surface removal, the image-order style is more advantageous for lowering the price than for enhancing the performance. The 3D graphic accelerator of this style had been mainly adopted in the past.
The following is a comparative explanation between the object-order style and the image-order style made with reference to FIG.
2
.
FIG. 2
shows triangles consisting of A, B, C and D, E, F. Assume that, of the two triangles, the one consisting of A, B, C is defined earlier than the one consisting of D, E, F and first inputted to the 3D graphic accelerator. According to the object-order style, the information on colors and depths generated from the geometrical process and the rendering process through a 3D graphic pipeline with respect to the triangle consisting of A, B, C is stored in a frame buffer. As a next step, information on the colors and the depths is calculated in the geometrical process and the rendering process with respect to the triangle consisting of D, E, F. The calculated information is compared with the information on the depths of the triangle A, B, C already stored with respect to the overlapping interval, and the information closer to the viewpoint is selected and stored.
FIG. 2
shows an occasion where an overlapping interval exists with respect to the two triangles and the one consisting of D, E, F is closer to the viewpoint than the one consisting of A, B, C with respect to the overlapping interval.
According to the image-order style, the processing order of the triangles depends on the position of the image currently being processed.
For instance, if the position currently being processed according to the image-order style in
FIG. 2
is the k
th
scan-line, the values in the X axis are processed in a reverse order of the numerical values. Subsequently, all the triangles corresponding to the pixel currently being processed are searched irrespective of an input order of the triangles, and the one closest to the viewpoint is selected for calculation of the ultimate color values. In other words, when the value in the X axis is the k
th
scan-line, the triangle consisting of A, B, C is processed within the interval corresponding to the triangle consisting of A, B, C except within the overlapping interval. Of the two triangles, the position closer to the viewpoint (the triangle consisting of D, E, F in
FIG. 2
) is searched and processed within the overlapping interval. The triangle consisting of D, E, F is processed in the remaining interval corresponding to the triangle consisting of D, E, F.
Meanwhile, the most outstanding image-order style is a scan-line style, as shown in FIG.
3
.
Still referring to
FIG. 3
, in the geometrical process, all the primitives are geometrically processed according to the viewpoint, and are transferred to the position on the screen. The pertaining information is stored in buckets existing per scan-line. This means that the buckets are supposed to have all the information pertaining to the primitives corresponding to the corresponding scan-lines. This is called a “bucket sorting.” Such a bucket sorting is a part to be processed by a geometrical processing section or a separate device. The rendering process is performed upon completion of the geometrical process and storage of the information pertaining to all the primitives in the buckets existing per scan-line. The rendering process is performed in a predetermined order of scan-lines.
Assume that the rendering process is performed from the 0
th
scan-line to the final n−1
th
scan-line, as illustrated in
FIG. 3
, for instance, and the currently being processed scan-line is the k
th
, and the (x, y) position on the screen of the k
th
scan-line begins from (0, k) and ends with (m, k). Then, the primitives corresponding to the position (0, k) is first searched from the buckets and the one closest to the viewpoint is selected for calculation of an ultimate color value, according to the scan-line style. If such a process is performed to the (m, k) position, the rendering process per scan-line is completed, and the information rendered on a scan-line is transferred for a display refresh. The same process is performed with respect to the k
+1
th scan-line, which corresponds to the next scan-line, and to the entire scan-lines.
The following are the characteristic features of the image-order style.
First, a huge memory space is required to retain information in buckets allotted per scan-line in proportion to an increase of the number of primitives.
Second, whereas the object-order style requires a z-buffer for a full screen, the scan-line style
Han Tack Don
Park Woo Chan
Nguyen Kimbinh T.
Sheridan Ross PC
Zimmerman Mark
LandOfFree
3D graphic accelerator and method for processing graphic... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with 3D graphic accelerator and method for processing graphic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and 3D graphic accelerator and method for processing graphic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3067185