JP2008198105A - Three-dimensional graphics rendering apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a three-dimensional graphics rendering apparatus capable of reducing a circuit scale. <P>SOLUTION: A display screen is divided into 4-pixel of 2×2 areas respectively comprising two pairs of pixels different only in lowest order XY-coordinate bit and same in high order bits. In respective 4-pixel areas (1) to (7) having a pixel included in a polygon, two or more pixels are extracted that may include a dummy pixel not included in the polygon but used for LOD calculation, based on XY coordinates of an upper left pixel and a flag indicative of the validity of the 4 pixels concerned, and a single LOD for the 4-pixel area is calculated. Thereby, the LOD can be calculated by a three-dimensional graphics rendering apparatus simply configured, and with a reduced circuit scale. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a three-dimensional graphic drawing apparatus having a texture mapping processing function for pasting a texture image on a polygon.

  When performing a texture mapping process in which a texture image is pasted on a polygon in a 3D graphics rendering process, if the polygon is smaller than the texture image, aliasing may occur due to insufficient sampling, and the image quality may deteriorate.

  Mipmap processing is well known as a method for preventing the deterioration of image quality due to aliasing (for example, Non-Patent Documents 1 and 2). In mipmap processing, multiple texture images reduced in stages by filtering are prepared in advance, and when drawing polygons, texture images of an appropriate size are selected and pasted to prevent aliasing. Reduce.

  In this case, a value indicating “how large a texture image is to be selected” is referred to as LOD (Level Of Detail), and this LOD can be obtained by the following equations (1) and (2).

  In the above formulas (1) and (2), D represents the reduction ratio of the texture image on the screen, and can be approximated by the following formula (3).

  In this equation (3), the reduction ratio D is the largest value of the amount of change in the texture coordinates (u, v) when the coordinates (x, y) have changed by a certain amount in the x and y directions on the screen. Yes.

  By the way, conventionally, various methods for reducing the circuit scale for LOD calculation have been proposed (for example, Patent Documents 1 and 2). That is, Patent Document 1 proposes a method of expanding the formula and reducing the number of multiplications and divisions.

  In Patent Document 2, in a device that processes a plurality of pixels in parallel, there is a method for reducing the circuit scale for LOD calculation by sharing the reduction ratio obtained at the representative point for all the pixels that are simultaneously processed. Proposed.

JP 2003-157445 A Japanese Patent Laid-Open No. 2001-118056 "3D Computer Graphics Third Edition" by Alan Watt, Addison Wesley, 2000, P256-260 `` Real-Time Rendering '' by Tomas Moller and Eric Haines, A K Peters, Ltd., 1999, P111-114

  However, the LOD calculation method described in Patent Document 1 requires a large number of operations for each pixel and has a problem that the circuit scale is large. In addition, since the LOD calculation method described in Patent Document 2 is based on the premise of pixel parallel processing, there is a problem that it cannot be applied to, for example, an apparatus that performs pixel-by-pixel processing with emphasis on circuit scale and power consumption.

  The present invention has been made in view of the above, and an object thereof is to obtain a three-dimensional graphic drawing apparatus capable of reducing the circuit scale.

  In order to achieve the above-described object, the present invention provides a 3D graphic rendering apparatus having a texture mapping processing function for pasting a texture image on a polygon, and the XY coordinates differ only in the least significant bit and are the same in the upper bit. On the display screen divided into 2 4 pixel areas, for each of the 4 pixel areas where the pixels included in the polygon exist, an XY coordinate of the upper left pixel and a flag indicating the validity of the 2 × 2 4 pixels are generated. Whether or not a pixel to be output to the subsequent stage is included in the polygon based on the XY coordinate generation unit that performs and the XY coordinate of the upper left pixel in the one four-pixel region input from the XY coordinate generation unit and the flag If it is not included in the polygon but is used for LOD calculation, As a pixel, dummy pixel generating means for outputting XY coordinates of two or more pixels that may include this dummy pixel one pixel at a time, and based on the XY coordinates of each pixel input from the dummy pixel generating means, Texture coordinate calculation means for calculating texture coordinates subjected to perspective correction, the texture coordinates input from the texture coordinate calculation means, and storing the LOD value from the texture coordinates of adjacent pixels in one of the four pixel regions LOD calculation means for calculating in a shared form is provided.

  According to the present invention, the display screen is divided into 2 × 2 4 pixel regions in which only the least significant bits of the XY coordinates are different and the upper bits are the same, and in each of the 4 pixel regions where the pixels included in the polygon exist, From the XY coordinates of the upper left pixel and the flag indicating the validity of the four pixels, two or more pixels that are not included in the polygon but may include dummy pixels used for the calculation of the LOD are extracted, and 1 for the four pixel region is extracted. Since one LOD is calculated, the LOD can be calculated with a simple configuration, and a three-dimensional graphic drawing apparatus with a small circuit scale can be realized.

  According to the present invention, there is an effect that a three-dimensional graphic drawing apparatus having a small circuit scale can be obtained.

  Exemplary embodiments of a three-dimensional graphic drawing apparatus according to the present invention are explained in detail below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a three-dimensional graphic drawing apparatus according to an embodiment of the present invention. A three-dimensional graphic drawing apparatus 1 shown in FIG. 1 includes a setup processing unit 2, an XY coordinate generation unit (XY coordinate generation unit) 3, a dummy pixel generation unit (dummy pixel generation unit) 4, and a texture coordinate calculation unit (texture). A coordinate calculation unit) 5, an LOD calculation unit (LOD calculation unit) 6, and a texture processing unit 7 are provided. As an external device, a texture memory 8 that holds a texture image and a frame buffer 9 that holds an image of a processing result are shown.

  The setup processing unit 2 performs preprocessing such as calculation of various parameters from input polygon data.

  The XY coordinate generation unit 3 obtains the XY coordinates of the pixels included in the polygon one pixel at a time based on the parameters obtained by the setup processing unit 2, and indicates the XY coordinates and valid pixels for every 2 × 2 pixels. Is output.

  The dummy pixel generation unit 4 outputs the XY coordinates of the pixels included in the polygon pixel by pixel based on the XY coordinates and the flag input from the XY coordinate generation unit 3. At that time, even if the pixel necessary for the calculation of LOD is not included in the polygon, its XY coordinates are output as a dummy pixel.

  The texture coordinate calculation unit 5 calculates the texture coordinates subjected to the perspective correction of the pixel from the XY coordinates of the pixel input from the dummy pixel generation unit 4 and the parameters obtained by the setup processing unit 2.

  The LOD calculation unit 6 accumulates the texture coordinates input from the texture coordinate calculation unit 5, calculates the LOD value from the texture coordinates of adjacent pixels, and outputs the texture coordinate and the LOD value in association with each other.

  The texture processing unit 7 calculates an address on the texture memory 8 based on the texture coordinates and the LOD value input from the LOD calculation unit 6, reads the texture data from the texture memory 8 using the address, and performs predetermined processing. Then, the image of the processing result is written into the frame buffer 9.

  Next, with reference to FIGS. 1-9, operation | movement of the three-dimensional graphic drawing apparatus 1 comprised as mentioned above is demonstrated. FIG. 2 is a diagram for explaining the operation of the XY coordinate generation unit 3. 3 and 4 are diagrams for explaining the operation of the dummy pixel generation unit 4. FIG. 5 is a block diagram illustrating a configuration example of the LOD calculation unit 6. 6 to 9 are diagrams for explaining the operation of the LOD calculation unit 6.

  When polygon data is input to the 3D graphic drawing apparatus 1 from the outside and an activation instruction is input, the setup processing unit 2 starts preprocessing. Since the preprocessing performed by the setup processing unit 2 is general, detailed description is omitted, but the following two processes are performed.

  That is, the setup processing unit 2 calculates internal / external determination parameters used to determine whether a pixel is inside or outside a polygon from the XY coordinates of each vertex of a triangle indicated by polygon data input from the outside. Then, it is output to the XY coordinate generation unit 3. Further, the setup processing unit 2 calculates a coefficient for linearly complementing the texture coordinates (s, t, q) of each vertex of the polygon, and outputs it to the texture coordinate calculation unit 5.

  Next, the operation of the XY coordinate generation unit 3 will be described with reference to FIG. In FIG. 2, the relationship between the XY coordinate plane (display screen) divided for each 2 × 2 4-pixel area in which only the least significant bits of the XY coordinates are different and the upper bits are the same and the polygon 10 to which the texture image is pasted are shown. It is shown. FIG. 2 shows a case where a polygon 10 straddles seven (1) to (7) of a 2 × 2 4-pixel region (hereinafter referred to as “unit 4-pixel region”).

  The XY coordinate generation unit 3 uses the internal / external determination parameters received from the setup processing unit 2 to perform internal / external determination of each pixel on the XY coordinate plane, and the center of the pixel is a polygon. 10, it is determined whether or not the pixel is an effective pixel that needs to be drawn. However, instead of determining each pixel by moving it one by one in a certain direction along one scan line, FIG. As shown, determination is made by moving the adjacent two scan line regions in a zigzag manner. Specifically, as shown in FIG. 2, the unit 4 pixel region is moved in the order of upper left, lower left, lower right, and upper right to determine the inside / outside.

  Then, the XY coordinate generation unit 3 outputs data once for every unit 4 pixel region. The data output by the XY coordinate generation unit 3 is composed of the XY coordinates of the upper left pixel of the unit 4 pixel area and a 4-bit flag indicating which of the 2 × 2 4 pixels is a valid pixel. . Here, the effective pixel is a pixel that is included in the polygon 10 and is a drawing target.

  The initial value of the 4-bit flag that the XY coordinate generation unit 3 has is “0000”, but the center of the pixel is within the polygon 10 in the following procedure (a) to (d). Is determined to be included, and if it is determined to be included, the corresponding bit of the 4-bit flag is set to “1” to clearly indicate that it is a valid pixel. Each of the 4 bits arranged in a horizontal row is referred to as the first, second, third, and fourth bits from the rightmost bit to the leftmost bit.

  (A) When the least significant bit of the XY coordinates is (0, 0), the pixel is located at the upper left of the unit 4 pixel area, so the fourth bit (most significant bit) of the flag is operated. To do. That is, the fourth bit (most significant bit) of the flag is set to “1” when it is included in the polygon 10, and is set to “0” when it is included.

  (B) When the least significant bit of the XY coordinates is (1, 0), the pixel is located in the upper right corner of the unit 4 pixel area, so the third bit of the flag is manipulated. That is, the third bit of the flag is set to “1” if it is included in the polygon 10, and is set to “0” if it is included.

  (C) If the least significant bits of the XY coordinates are (0, 1), the pixel is located at the lower left of the unit 4-pixel area, and therefore the second bit of the flag is manipulated. That is, the second bit of the flag is set to “1” when included in the polygon 10, and set to “0” when included in the polygon 10.

  (D) When the least significant bit of the XY coordinates is (1, 1), the pixel is located in the lower right corner of the unit 4 pixel area, so the first bit (least significant bit) of the flag is set. Set to “1” to operate. That is, the first bit (least significant bit) of the flag is set to “1” when included in the polygon 10, and set to “0” when included in the polygon 10.

  The XY coordinate generation unit 3 repeats the same processing while moving each pixel in a zigzag manner in each of the unit 4 pixel regions. When moving a pixel, when a bit higher than the least significant bit of the X coordinate changes (that is, when moving to the adjacent unit 4 pixel region) and when the processing of 2 scan lines is completed, If the flag value is not “0000”, the XY coordinates of the upper left pixel and the flag value in the current unit 4-pixel area are output, and the flag value is cleared to “0000”. In this way, the XY coordinate generation unit 3 outputs the pixel data (the XY coordinates of the upper left pixel and the 4-bit flag value) to the dummy pixel generation unit 4 for every unit pixel area for all the pixels included in the polygon 10. Output to.

  In the example shown in FIG. 2, since it is determined that the pixel to be rendered is included in seven of the unit 4-pixel area (1) to (7), the upper left pixel of the unit 4-pixel area (1) Output data consisting of the XY coordinates of the pixel 4 and the flag value “1101” for the unit 4 pixel region (1), the XY coordinates of the upper left pixel of the unit 4 pixel region (2) and the flag value “for the unit 4 pixel region (2)” Output data consisting of 1111 ", XY coordinates of unit 4 pixel area pixel (3) and output data consisting of flag value" 1111 "for the unit 4 pixel area (3), upper left pixel of unit 4 pixel area (4) Output data consisting of the XY coordinates of the pixel and the flag value “0101” for the unit 4-pixel area (4), the XY coordinates of the upper-left pixel of the unit 4-pixel area (5) And the output data consisting of the flag value “1111” for the unit 4-pixel region (5), the XY coordinates of the upper left pixel of the unit 4-pixel region (6), and the flag value “1000” for the unit 4-pixel region (6) , Output data consisting of the XY coordinates of the upper left pixel of the unit 4 pixel region (7) and the flag value “1010” for the unit 4 pixel region (7) are output to the dummy pixel generation unit 4 in this order. The

  Next, the operation of the dummy pixel generation unit 4 will be described with reference to FIGS. When the dummy pixel generation unit 4 receives the pixel data of the unit 4 pixel region input from the XY coordinate generation unit 3, the dummy pixel generation unit 4 generates the XY coordinates of the upper right pixel, the lower left pixel, and the lower right pixel based on the XY coordinates of the upper left pixel. Then, the XY coordinates of each pixel in the unit 4-pixel region are output one pixel at a time in the order of upper left, upper right, lower left, and lower right.

  At this time, the upper left, upper right, and lower left pixels of the unit 4-pixel region are used for LOD calculation, and therefore are output even when it is determined that they are not included in the polygon from the 4-bit flag value. The pixel in this case is a “dummy pixel” in this embodiment. On the other hand, the lower right pixel of the unit 4-pixel region is not used for LOD calculation, and therefore is not output when it is determined that it is not included in the polygon from the 4-bit flag value.

  Therefore, when the dummy pixel generation unit 4 outputs the XY coordinates of each pixel, it simultaneously displays a 1-bit flag indicating whether the pixel is a drawing target, that is, whether it is a dummy pixel. Output. As described above, the order of outputting the XY coordinates and the 1-bit flag is the order of upper left, upper right, lower left, and lower right in the unit 4-pixel region.

  FIG. 3A shows an operation when the input data is composed of the XY coordinates and the flag value “1101” in the unit 4-pixel area (1) shown in FIG. The XY coordinates of the upper left pixel in the unit 4-pixel region (1) shown in FIG. 2 are assumed to be (x1, y1). The flag value “1101” indicates that among the four pixels, the upper left, upper right, and lower right three pixels are valid pixels.

  When the dummy pixel generation unit 4 receives data in the unit 4 pixel region (1) shown in FIG. 2 from the XY coordinate generation unit 3, first, the dummy pixel generation unit 4 sends the XY coordinates (x1, y1) of the upper left pixel to the texture coordinate calculation unit 5. Output. At this time, since the fourth bit (most significant bit) of the flag is “1”, the above-described 1-bit flag is set to “1” and is output to the texture coordinate calculation unit 5.

  Next, the dummy pixel generation unit 4 outputs the XY coordinates (x1 + 1, y1) of the upper right pixel to the texture coordinate calculation unit 5. At this time, since the third bit of the flag is “1”, the 1-bit flag described above is set to “1” and is output to the texture coordinate calculation unit 5.

  Next, the dummy pixel generation unit 4 outputs the XY coordinates (x1, y1 + 1) of the lower left pixel to the texture coordinate calculation unit 5. At this time, since the second bit of the flag is “0”, the 1-bit flag described above is set to “0” and is output to the texture coordinate calculation unit 5.

  Finally, since the first bit (least significant bit) of the flag is “1”, the dummy pixel generation unit 4 outputs the XY coordinates (x1 + 1, y1 + 1) of the lower right pixel to the texture coordinate calculation unit 5. . At that time, the 1-bit flag described above is set to “1” and output to the texture coordinate calculation unit 5.

  FIG. 3B shows an operation when the input data is composed of the XY coordinates and the flag value “1000” in the unit 4-pixel area (6) shown in FIG. The XY coordinates of the upper left pixel in the unit 4-pixel area (6) shown in FIG. 2 are (x6, y6). The flag value “1000” indicates that only one pixel at the upper left of the four pixels is valid. Also in this case, the same operation as described in FIG.

  When the dummy pixel generation unit 4 receives data in the unit 4 pixel region (6) shown in FIG. 2 from the XY coordinate generation unit 3, first, the dummy pixel generation unit 4 sets the XY coordinates (x6, y6) of the upper left pixel to the texture coordinate calculation unit 5. Output. At this time, since the fourth bit (most significant bit) of the flag is “1”, the 1-bit flag described above is set to “1” and is output to the texture coordinate calculation unit 5.

  Next, the dummy pixel generation unit 4 outputs the XY coordinates (x6 + 1, y1) of the upper right pixel to the texture coordinate calculation unit 5. At this time, since the third bit of the flag is “0”, the 1-bit flag is set to “0” and output to the texture coordinate calculation unit 5.

  Next, the dummy pixel generation unit 4 outputs the XY coordinates (x6, y6 + 1) of the lower left pixel to the texture coordinate calculation unit 5. At this time, since the second bit of the flag is “0”, the 1-bit flag described above is set to “0” and is output to the texture coordinate calculation unit 5.

  Finally, since the first bit (least significant bit) of the flag is “0”, the dummy pixel generation unit 4 uses both the XY coordinates (x6 + 1, y6 + 1) of the lower right pixel and the above 1-bit flag as texture coordinates. It is not output to the calculation unit 5.

  FIG. 4 is a diagram for explaining the operation results for all of the unit 4 pixel regions (1) to (7) shown in FIG. The dummy pixel generation unit 4 performs the pixel A included in the polygon 10 in this order for each of the upper left, upper right, lower left, and lower right pixels in the unit 4 pixel region by the method described with reference to FIGS. Outputs the 1-bit flag in which the XY coordinate of the pixel and “1” are set to the texture coordinate calculation unit 5, and for the dummy pixel B that is not used for drawing but is necessary for the LOD calculation, the XY coordinate of the pixel and “ The output of the 1-bit flag set to “0” to the texture coordinate calculation unit 5 is performed for all the unit 4-pixel regions (1) to (7) related to the polygon 10.

  In FIG. 4, the unit 4 pixel regions (1) to (7) shown in FIG. 2 are not labeled, but in the unit 4 pixel region (1), the upper left, upper right, and lower right 3 pixels are the polygon 10. The pixel A is included, and the lower left pixel is the dummy pixel B. In the unit 4-pixel areas (2), (3), and (5), all four pixels are pixels A included in the polygon 10. In the unit 4-pixel region (4), the upper right and lower right two pixels are the pixels A included in the polygon 10, and the upper left and lower left two pixels are the dummy pixels B. In the unit 4-pixel area (6), the upper left one pixel is a pixel A included in the polygon 10, and the upper right and lower left two pixels are dummy pixels B. In the unit 4-pixel region (7), the upper left and lower left two pixels are pixels A included in the polygon 10, and the upper right one pixel is a dummy pixel B.

Next, the operation of the texture coordinate calculation unit 5 will be described. The texture coordinate calculation unit 5 first calculates the XY coordinates (x, y) of the pixels input from the dummy pixel generation unit 4 and the vertices of the polygons input from the setup processing unit 2 in the following equation (4). Apply the coefficients a _s , b _s , c _s , a _t , b _t , c _t , a _q , b _q , c _q to linearly complement the texture coordinates (s, t, q), and the pixel The value of the texture coordinates (s, t, q) at is obtained.

s = a _s × x + b _s × y + c _{s
t = a t × x + b} t × y + c t ... (4)
q = a _q × x + b _q × y + c _q

And the texture coordinate calculation part 5 calculates by the following formula | equation (5), and calculates | requires the value of the texture coordinate (u, v) which performed perspective correction. In Equation (5), USIZE and VSIZE are the sizes of the texture image in the horizontal and vertical directions, respectively.
u = USIZE × s / q, v = VSIZE × t / q (5)

  The texture coordinate calculation unit 5 uses the texture coordinates (u, v) of each input pixel thus obtained as the XY coordinates (x, y) of the corresponding pixel received from the dummy pixel generation unit 4 and the 1-bit flag. It outputs to the LOD calculation part 6 with a value.

  Next, the configuration and operation of the LOD calculation unit 6 will be described. FIG. 5 is a block diagram illustrating a configuration example of the LOD calculation unit 6. The LOD calculator 6 shown in FIG. 5 includes buffers 11 and 12, selectors 13 and 14, a subtraction calculator 15, an absolute value calculator (ABS) 16, a logarithm calculator (log2) 17, and a maximum value selector. 18 and a buffer 19 are provided.

  The buffer 11 temporarily stores the data (texture coordinates (u, v), XY coordinates (x, y), and 1-bit flag value) of each pixel in the unit 4-pixel area input from the texture coordinate calculation unit 5.

  The buffer 12 temporarily holds the data (texture coordinates (u, v), XY coordinates (x, y) and 1-bit flag value) of each pixel output from the buffer 11, and the 1-bit flag value is “1”. In the case of, the texture coordinates (u, v) and the XY coordinates (x, y) are controlled to be sent to the outside (texture processing unit 7).

  The selector 13 selects either the u coordinate or the v coordinate input to the buffer 11 and supplies the selected value to the subtraction calculator 15. The selector 14 gives either the u coordinate or the v coordinate held by the buffer 11 or the u coordinate or the v coordinate held by the buffer 12 to the selective subtraction calculator 15.

  The subtraction calculator 15 calculates a difference between the selection data of the selectors 13 and 14. The absolute value calculator 16 calculates the absolute value of the difference calculated by the subtraction calculator 15. The logarithmic calculator 17 calculates the LOD value by performing the logarithmic calculation of the above equation (1) with 2 as the base for the absolute value obtained by the absolute value calculator 16. The maximum value selector 18 and the buffer 19 perform processing for selecting and holding the larger LOD value among the LOD values obtained by the logarithmic calculator 17.

  In short, in the LOD calculation unit 6, each pixel data of the unit 4-pixel region (texture coordinate (u, v), XY coordinate (x, y) and 1-bit flag value of each pixel) input from the texture coordinate calculation unit 5. Are sequentially stored in the buffers 11 and 12, the selectors 13 and 14 and the subtraction calculator 15 obtain the difference between the texture coordinates (u, v) of two adjacent pixels, and the difference is calculated by the absolute value calculator 16 , | Du / dx |, | du / dy |, | dv / dx |, | dv / dy | in the above equation (3), and the logarithmic calculator 17 performs the logarithmic operation of the above equation (1). To calculate the LOD value, the maximum value selector 18 and the buffer 19 select and hold the maximum LOD value, and the 1-bit flag value for the unit 4 pixel area for which the processing is completed is “1”. is there In this case, the pixel data (the texture coordinates (u, v) and XY coordinates (x, y) of the upper left pixel) held by the buffer 12 and the LOD value held by the buffer 19 are associated with each other and output to the texture processing unit 7. Is configured to do.

  The LOD calculation unit 6 calculates the LOD value from the upper left, upper right, and lower left pixel data in the unit 4 pixel region by the procedure shown in FIGS. Note that the LOD calculation unit 6 determines which pixel in the unit 4-pixel area the pixel input from the texture coordinate calculation unit 5 is based on the least significant bit of the XY coordinates. 6 and 7 are diagrams for explaining the operation at the timing when the data of the upper right pixel is input after the data of the upper left pixel in the unit 4 pixel region is held in the buffer 11. 8 and 9 show the operation at the timing when the data of the upper left pixel in the unit 4 pixel region is transferred from the buffer 11 to the buffer 12 and the data of the lower left pixel is input after the data of the upper right pixel is held in the buffer 11. It is a figure explaining.

  In FIG. 6, when data of the upper left pixel of the unit 4 pixel region is input from the texture coordinate calculation unit 5, the data is stored in the buffer 11. Thereafter, the data of the upper right pixel in the unit 4 pixel region is input from the texture coordinate calculation unit 5. At that time, the u coordinate of the upper left pixel held in the buffer 11 and the u coordinate of the upper right pixel input from the texture coordinate calculation unit 5 are respectively selected by the selectors 13 and 14, and the subtraction calculator 15, absolute The LOD is calculated by the value calculator 16 and the logarithmic calculator 17, and the calculation result is stored in the buffer 19 via the maximum value selector 18.

  In FIG. 7, in the next cycle, the v coordinate of the upper left pixel held in the buffer 11 and the v coordinate of the upper right pixel input from the texture coordinate calculation unit 5 are respectively selected by the selectors 13 and 14 and subtracted. The LOD is calculated by the calculator 15, the absolute value calculator 16, and the logarithmic calculator 17, and the maximum value selector 18 selects the larger one compared with the value held in the buffer 19, and this is selected as the buffer 19. To store. At this time, the LOD value held in the buffer 19 is expressed by the following equation (6).

  In FIG. 8, when the above processing is completed, the upper left pixel data held in the buffer 11 is moved to the buffer 12, and the upper right pixel data input from the texture coordinate calculation unit 5 is held in the buffer 11. Thereafter, the data of the lower left pixel of the unit 4 pixel region is input from the texture coordinate calculation unit 5. At that time, the u-coordinate of the upper-left pixel held in the buffer 12 and the u-coordinate of the lower-left pixel input from the texture coordinate calculation unit 5 are respectively selected by the selectors 13 and 14, and the subtraction calculator 15, absolute The LOD is calculated by the value calculator 16 and the logarithmic calculator 17, and the maximum value selector 18 selects the larger one compared with the value held in the buffer 19 and stores it in the buffer 19.

  9, in the next cycle, the v coordinate of the upper left pixel held in the buffer 12 and the v coordinate of the upper right pixel input from the texture coordinate calculation unit 5 are respectively selected by the selectors 13 and 14 and subtracted. The LOD is calculated by the calculator 15, the absolute value calculator 16, and the logarithmic calculator 17, and the maximum value selector 18 selects the larger one compared with the value held in the buffer 19, and this is selected as the buffer 19. To store. At this time, the LOD value held in the buffer 19 is expressed by the following equation (7).

  When the LOD calculation using the upper left, upper right, and lower left pixel data in the unit 4 pixel region is completed by the above procedure, the LOD value finally held in the buffer 19 is changed to the upper left pixel held in the buffer 12. Output to the texture processing unit 7 together with the data. However, if the 1-bit flag indicating whether or not to draw a pixel is “0”, the pixel is a dummy pixel and is discarded without being output.

  Then, the upper right pixel data held in the buffer 11 is transferred to the buffer 12 and output to the texture processing unit 7. At this time, the LOD value of the buffer 19 is not changed similarly, and the same LOD value as that of the previous pixel is output. However, when the 1-bit flag indicating whether or not to draw a pixel is “0”, the pixel is similarly a dummy pixel and is discarded without being output.

  Similarly, the lower left and lower right pixel data subsequently input from the texture coordinate calculation unit 5 are also output to the texture processing unit 7 via the buffers 11 and 12. At this time, the LOD value of the buffer 19 is not changed similarly, and the same LOD value as that of the previous pixel is output. However, when the 1-bit flag indicating whether or not to draw a pixel is “0”, the pixel is similarly a dummy pixel and is discarded without being output.

  As described above, each time the upper left pixel in the unit 4 pixel region is input, the LOD calculation unit 6 calculates the LOD for each unit 4 pixel region and repeatedly shares the calculation result with the unit 4 pixel region. . When the 1-bit flag indicating whether or not to draw a pixel is “1”, the LOD value shared by the unit 4 pixel area is output to the texture processing unit 7 together with the XY coordinate and texture coordinate of the pixel. .

  Next, as described above, the texture processing unit 7 calculates an address on the texture memory 8 based on the texture coordinates and the LOD value input from the LOD calculation unit 6, and uses the address to generate a texture from the texture memory 8. Data is read and predetermined processing is performed, and an image of the processing result is written in the frame buffer 9. Since the processing method in the texture processing unit 7 is a general one, the detailed description is omitted here, but the dummy pixels used for the calculation of the LOD generated as described above are input to the texture processing unit 7. Therefore, the texture processing unit 7 can perform a predetermined process as usual without any trouble.

  As described above, according to this embodiment, the display screen is divided into 2 × 2 4 pixel areas in which only the least significant bits of the XY coordinates are different and the upper bits are the same, and pixels included in the polygon exist. In each of the four pixel areas, two or more pixels that are not included in the polygon but may include dummy pixels that are used for the LOD calculation are extracted from the XY coordinates of the upper left pixel and the flag indicating the validity of the four pixels. Since one LOD for the 4-pixel region is calculated, the LOD can be calculated with a simple configuration, and a three-dimensional graphic drawing apparatus with a small circuit scale can be realized.

  Further, when generating a dummy pixel, a flag indicating whether or not the pixel is a dummy pixel is generated and passed to the LOD calculation unit. When the LOD calculation unit indicates that the flag is not a dummy pixel, Since the pixel data and the LOD value are output to the texture processing unit, the texture processing unit can execute a predetermined process without any trouble.

  As described above, the three-dimensional graphic drawing apparatus according to the present invention is useful for reducing the circuit scale.

It is a block diagram which shows the structure of the three-dimensional graphic drawing apparatus by Embodiment 1 of this invention. It is a figure explaining operation | movement of the XY coordinate production | generation part shown in FIG. It is a figure explaining operation | movement of the dummy pixel production | generation part shown in FIG. It is a figure explaining the operation result about all the unit 4 pixel area | regions (1)-(7) shown in FIG. 2 of the dummy pixel production | generation part shown in FIG. It is a block diagram which shows the structural example of the LOD calculation part shown in FIG. It is a figure explaining the operation | movement which the LOD calculation part shown in FIG. 5 calculates a LOD value from each u coordinate of upper left and upper right. It is a figure explaining the operation | movement which the LOD calculation part shown in FIG. 5 calculates a LOD value from each v coordinate of upper left and upper right. It is a figure explaining the operation | movement which the LOD calculation part shown in FIG. 5 calculates a LOD value from each u coordinate of lower left and upper left. It is a figure explaining the operation | movement which the LOD calculation part shown in FIG. 5 calculates a LOD value from each v coordinate of lower left and upper left.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 3D graphic drawing apparatus 2 Setup processing part 3 XY coordinate production | generation part (XY coordinate production | generation means)
4 Dummy pixel generator (dummy pixel generator)
5 Texture coordinate calculator (texture coordinate calculator)
6 LOD calculation part (LOD calculation means)
7 Texture processing section 8 Texture memory 9 Frame buffer 10 Polygon 11, 12, 19 Buffer 13, 14 Selector 15 Subtraction calculator 16 Absolute value calculator (ABS)
17 Logarithm calculator 18 Maximum value selector (1) to (7) Unit 4 pixel area A Pixel included in polygon B Dummy pixel

Claims (6)

In a 3D graphic drawing device having a texture mapping processing function for pasting a texture image to a polygon,
XY coordinates of the upper left pixel for each of the 4 pixel areas in which the pixels included in the polygon exist in a display screen divided into 2 × 2 4 pixel area units in which only the least significant bit is different and the upper bits are the same XY coordinate generation means for generating a flag indicating the validity of the 2 × 2 4 pixels,
Based on the XY coordinates of the upper left pixel in one of the four pixel regions input from the XY coordinate generation means and the flag, it is determined whether or not the polygon to be output in the subsequent stage is included in the polygon. A dummy pixel generating means for outputting the XY coordinates of two or more pixels that may include the dummy pixel one by one as dummy pixels when used for the calculation of LOD,
Texture coordinate calculation means for calculating texture coordinates subjected to perspective correction of each pixel based on XY coordinates of each pixel input from the dummy pixel generation means;
LOD calculation means for accumulating the texture coordinates input from the texture coordinate calculation means and calculating the LOD value from the texture coordinates of adjacent pixels in a form shared by one of the four pixel regions. Characteristic 3D graphic drawing device.   2. The dummy pixel generation unit according to claim 1, wherein the upper left, upper right, and lower left pixels in one of the four pixel regions are output as the dummy pixels even when they are not included in the polygon. 3D graphic drawing device. The LOD calculation means includes:
A buffer for two pixels for sequentially storing the texture coordinates input from the texture coordinate calculation means;
Using the input pixels to the two-pixel buffer and the retained output pixels of the two-pixel buffer, the LOD is calculated from the difference in texture coordinates between the upper left, upper right, and lower left pixels in one 4-pixel region. A computing means for calculating a value;
The three-dimensional selection according to claim 1, further comprising: a maximum value selection / holding unit that holds a maximum value among the LOD values calculated by the calculation unit as a LOD value in the four-pixel region. Graphic drawing device. In a 3D graphic drawing device having a texture mapping processing function for pasting a texture image to a polygon,
In the display screen divided into 2 × 2 4 pixel areas whose XY coordinates are different only in the least significant bit and are the same in the upper bits, the upper left pixel of each of the 4 pixel areas in which the pixels included in the polygon exist are displayed. XY coordinate generation means for generating XY coordinates and a flag indicating the validity of the 2 × 2 4 pixels,
Based on the XY coordinates of the upper left pixel in one of the four pixel regions input from the XY coordinate generation means and the flag, it is determined whether or not the polygon to be output in the subsequent stage is included in the polygon. Is not included but is used as a dummy pixel when used in the calculation of LOD, the XY coordinates of two or more pixels that may include this dummy pixel are output pixel by pixel together with a flag indicating whether or not the pixel is a dummy pixel. Dummy pixel generating means for performing
Based on the XY coordinates of the pixels input from the dummy pixel generation means, texture coordinates subjected to perspective correction of each pixel are calculated, and the calculated texture coordinates are input from the XY coordinates of the pixels and the dummy pixel generation means. Texture coordinate calculation means for outputting together with the flag,
The texture coordinates input from the texture coordinate calculation means are accumulated, the LOD value is calculated from the texture coordinates of adjacent pixels in a form shared by one of the four pixel areas, and the texture coordinates calculation means inputs the texture coordinates. LOD calculation means for outputting the calculated LOD value to the subsequent stage together with the accumulated texture coordinates and the XY coordinates of the pixels input from the texture coordinate calculation means when the flag does not indicate that the pixel is a dummy pixel And a three-dimensional graphic drawing apparatus.   The dummy pixel generation means outputs each of the upper left, upper right, and lower left pixels in one of the four pixel areas as the dummy pixel even if it is not included in the polygon, and outputs it together with the flag indicating the dummy pixel. The three-dimensional graphic drawing apparatus according to claim 4. The LOD calculation means includes:
A buffer for two pixels for sequentially storing the texture coordinates input from the texture coordinate calculation means;
Using the input pixels to the two-pixel buffer and the retained output pixels of the two-pixel buffer, the LOD is calculated from the difference in texture coordinates between the upper left, upper right, and lower left pixels in one 4-pixel region. A computing means for calculating a value;
Maximum value selection holding means for holding the maximum value among the LOD values calculated by the calculation means as the LOD value in the 4-pixel region;
When the flag input from the texture coordinate calculation means does not indicate that the pixel is the dummy pixel, the LOD value held by the maximum value selection holding means is the texture coordinates stored in the buffer, and The three-dimensional graphic drawing apparatus according to claim 4, further comprising: output control means for outputting to the subsequent stage together with XY coordinates of the pixels input from the texture coordinate calculation means.

Images