The present invention relates to the field of processing graphic images, and, in particular, to graphics rasterization.
In the field of computer graphics, the graphics rendering pipeline is the core of real time graphics. The function of the pipeline is to generate, or render, two dimensional images, three dimensional objects, light sources, lighting models, textures and more. The locations and shapes of the objects in the scene are determined by the geometry, placement of the camera in the environment and the characteristics of that environment. The appearance of the objects is affected by material properties, light sources, textures and lighting models.
The process of rasterizing in computer graphics defines a particular scene in terms of primitives, which are typically triangles. For a particular scene, the area to be displayed by a computer system is termed a screen region. The screen region usually is an area that is less than the total scene. Therefore, some of the primitives lie outside of the screen region, while others lie either partially in the screen region or completely within the screen region. During rasterization the primitives are processed such that pixels contained within the primitives are given values in terms of their color, textures, transparency, etc.
Initially during rasterization, the primitives are filled with a solid color or pattern. The function of filling primitives can be broken into two parts. First a decision must be made as to which pixels to “fill” (assign values) within a primitive, and also as to what values to assign to these pixels. If the pixels are to be assigned a solid color, for example, then primitives which lie entirely within the screen region have each of their pixels assigned the color value. However, if a primitive lies only partially within the screen area, to assign all the pixels a particular value and then to discard the portion of the primitive lying outside of the screen area requires large amounts of computer processing time and is inefficient.
The process of clipping is the process of determining the portion of a primitive that is within a clipping region, such as the screen region. By clipping the rectangle to the screen region drawing time is saved during the rasterization. Scissoring is the process of computing all points of the primitive, and then drawing only those lines within in the rasterized clip region. Although a number of algorithms are set forth in the prior art with regards to scissoring and clipping, increasing complexity of graphic images has required more efficient methods of clipping or scissoring primitives with regards to the screen region.
Therefore, there is a need in the prior art to provide a more efficient method and apparatus for culling pixels of a primitive that are not in a scissors region. A drawback of the prior art is an implementation of scissoring primitives which fills all the primitives with pixels and then for every pixel has hardware check x,y pixel coordinates against the scissor planes and cull each pixel individually if it is outside the scissor plane. This prior art method is very inefficient in terms of performance as it requires hardware to completely fill all primitives, even for pixels that are outside the scissor plane. Another prior art method is to perform clipping at intersections of primitive edges with the scissor plane. New primitives are then created by subdividing the original primitives in an interactive manner until there are no primitives to intersect the scissor plane. This prior art method of clipping is slow and costly in terms of hardware implementations.
The features of the present invention which are believed to be novel are set forth with particularity in the appended claims. The invention, together with further objects and advantages, may best be understood by reference to the following description taken in conjunction with the accompanying drawings, and the several figures of which like reference numerals identify like elements.
The present invention is used in a computer system, such as a computer system that displays graphic images. The method of the present invention is, in particular, for rasterizing primitives. A first step of the inventive method is to determine if a primitive is totally outside a predetermined screen region. The primitive is discarded if the primitive is totally outside the screen region. If the primitive is not totally outside the screen region, at least a portion of the primitive is identified that lies within the screen region. Then only those pixels in the portion of the primitive that is inside the screen region are filled. These steps are then executed for each primitive of the plurality of primitives that forms the scene of which the screen region is the portion that the computer system displays.
In general terms, a start point is defined at a vertex of a triangular primitive. It is then determined if the start point is outside the screen region. The edge of the primitive is then edge walked from the start point to a boundary of the screen region. A portion of the primitive inside the screen region is span walked and each pixel in the portion of the primitive that is inside the screen region is filled. In a typical graphic system, the primitive is a triangle and the start point is a vertex of the triangle. Edge walkers and span walkers are known in the art, for example.
The present invention does not fill primitives or any portion of primitives that are outside the screen region nor does it implement any hardware or require any hardware to calculate an intersection of primitives with the screen region. Instead, the present invention incorporates the screen region into the edge walking routine, that is, the edge walking module will only start filling the primitive when it is inside the screen region (image region). According to the present invention, if the edge walker starts at a vertex of the primitive outside the screen region, it then travels to the nearest point of primitive intersection with the screen region. The present invention will also terminate filling of a primitive when it transitions from inside of the screen region to outside the screen region.
Depicted, for example, in
The screen region 300 is defined by a coordinate system, wherein the bottom left corner of the screen region 300 is denoted as XLEFT, YBOTTOM, and has an upper right hand corner denoted by XRIGHT, YTOP. In this coordinate system, the x direction runs from left to right across the screen region 300, and the y direction runs from bottom to top across the screen region 300. Each of the primitives can be defined relative to the coordinate system and in the present invention, each primitive, such as primitive 302, is defined by a vertex XSTART, YSTART, and a value XEND which is the furthest x direction extent and a value YEND which is the furthest y direction extent from the starting value vertex XSTART, YSTART.
Directional values used by the present invention are depicted in
In step 504 it is determined if the primitive is totally outside of the screen region 300, such as a primitive 302 in
XDIR AND ((X<XLEFT) OR (XEND>XRIGHT))
YDIR AND ((Y<YTOP) OR (YEND>YBOTTOM))
If the primitive is found to be totally outside of the screen region, such as primitive 302 in
The portion of the primitive lying within the screen area 300 is identified as follows. First, the variable y is incremented if the following first value is logically true:
(((YDIR AND (Y>YBOTTOM)) OR ((
Then the variable x is incremented if the following second value is logically true:
(((XDIR AND X>XRIGHT)) OR ((
These steps (508, 510) are then repeated until the first and second values are not true, which identifies a beginning of a portion of the primitive that is inside the screen region. That is, the test is to determine if the variable y is between YTOP and YBOTTOM, and if the variable x is between XLEFT and XRIGHT. If the answer to this question is no, then x and y are incremented. If the answer to the question is yes, then the pixel corresponding to that x,y coordinate lies within the portion of the primitive within the screen region 300, and the pixel is then filled (step 512), that is, assigned the proper color value. The process ends when all pixels within the portion of the primitive inside the screen region 300 have been filled, the test being performed in step 514 in
It is to be noted that in general the method steps and the equations involve comparisons of the current x and y coordinates to the XLEFT, XRIGHT and YTOP and YBOTTOM of the screen area, as well as the X direction and Y direction values at that x,y coordinate. Note that when x and y are incremented, there are eight possible directions of movement as depicted in
The primitive location module 202 in one embodiment is software that implements the step of determining if the primitive is totally outside the screen region. Similarly, the edge walker module 204 and the span walker module 206 are software implementations for performing the above described steps.
When the edge walker 204 in the raster engine 110 begins at the XSTART, Y START values of a primitive, it walks one of the edges of the primitive until it reaches the closest boundary point of intersection of the primitive with the screen region. For example, as shown in
For primitive 306, the edge walker starts at vertex 330 and proceeds until it identifies point 332 as being the first point of intersection between the primitive 306 and the screen region 300. It is to be understood that once the area within the screen region 300 is identified, the span walker and fill module 206 then take over and fill each of the pixels and the portion of the primitive which is within the screen region 300. This is indicated by the area of the primitives which are filled with lines in
Thus, in general, in the primitive location module 202 the primitive values are compared to the screen region values to determine if the primitive is totally outside the screen region. If this is not true, then the edge walker module 204 and the span walker and fill module 206 determine the start point of the intersection of the primitive in the screen region for filling the pixels that are within this portion. The filling of the pixels within the portion within the screen region is finished when one of the following is true:
(XDIR AND (X<XLEFT)),
(
(YDIR AND (Y<YTOP)),
(
Thus, the invention fulfills the need in the prior art for an efficient method of filling the portions of primitives lying within a screen region. The method of the present invention in particular does not fill the portions of the primitives outside of the screen region, thus saving significant computing time and resources. Furthermore, the method of the present invention does not require a hardware calculation of the intersection of the primitives with the screen region as is done in the prior art.
The invention is not limited to particular details of the apparatus and method depicted and the modifications and applications may be contemplated. Certain other changes may be made in the above-described method and apparatus without departing from the true spirit of the scope of the invention herein involved. For example, the method of the present invention can be utilized in other computer systems other than the 3D pipeline embodiment depicted in
Number | Name | Date | Kind |
---|---|---|---|
3639736 | Sutherland | Feb 1972 | A |
4888712 | Barkans et al. | Dec 1989 | A |
5040130 | Chang et al. | Aug 1991 | A |
5051737 | Akeley et al. | Sep 1991 | A |
5079719 | Maillot | Jan 1992 | A |
5357599 | Luken | Oct 1994 | A |
5455897 | Nicholl et al. | Oct 1995 | A |
5488684 | Gharachorloo et al. | Jan 1996 | A |
5777625 | Rossin | Jul 1998 | A |
5856829 | Gray et al. | Jan 1999 | A |
5877773 | Rossin et al. | Mar 1999 | A |
6005586 | Morita et al. | Dec 1999 | A |
6052129 | Fowler et al. | Apr 2000 | A |
6144387 | Liu et al. | Nov 2000 | A |
6169554 | Deering | Jan 2001 | B1 |
6268875 | Duluk et al. | Jul 2001 | B1 |
6831660 | Kipping et al. | Dec 2004 | B1 |
20010013867 | Watanabe et al. | Aug 2001 | A1 |