JP2006127368A - Graphics rendering device with gamma correction function based on depth value - Google Patents

Abstract

To provide a graphics drawing apparatus capable of adjusting three-dimensional brightness more easily at low cost. The conventional illuminance calculation has a problem that the calculation cost is high, and hardware resources are increased for high-speed processing. Further, setting parameters such as the position of the light source and the material is complicated, and it is not easy for the user to easily program the brightness, and there is a problem that the brightness is not programmable.
A lookup table for setting a plurality of gamma correction curves is provided, and a plurality of color values to which gamma correction is applied by the lookup table according to the depth value (Z value) of each pixel. By selecting one of these, the brightness is adjusted according to the Z value.
[Selection] Figure 1

Description

  The present invention relates to a computer graphics technique, and more particularly to a technique for generating a real image with a sense of depth in a three-dimensional graphics drawing by a simpler method than before.

  In general three-dimensional graphics drawing, a plurality of objects composed of a plurality of polygons (polygons) arranged in a virtual three-dimensional space are placed on a two-dimensional plane (screen) perpendicular to the observer's line of sight. This is done by projecting. At this time, the hidden surface removal of the polygon is generally performed so that the context of the polygon is correct as viewed from the observer. The Z buffer method is widely used as a polygon hidden surface processing method. In the Z buffer method, a Z buffer that stores depth values (Z values) corresponding to the number of pixels of the display is used in addition to color buffers that store color values corresponding to the number of pixels of the display. The Z buffer stores the Z value of the object that has already been drawn and is closest to the viewpoint, and compares the Z value of this Z buffer with the Z value of the polygon to be drawn in pixel units. By doing so, the hidden surface removal is performed by drawing only the pixels of the polygon having a small Z value.

  The manner in which hidden surface removal of a three-dimensional object is performed using a Z buffer will be described with reference to FIGS. FIG. 2A shows how two polygons 201 and 202 arranged in a three-dimensional space are projected on the screen 200. A polygon 203 indicates a figure in which the polygon 201 is projected on the screen 200, and a polygon 204 indicates a figure in which the polygon 202 is projected on the screen 200. FIG. 2B is a diagram showing drawing by the Z buffer, and the two polygons 203 and 204 displayed on the screen 200 of FIG. 2A are the Z buffer Z value and the polygon Z value. These results show the result of rendering while removing the hidden surface for each pixel. FIG. 3 is an example in which hidden surface removal of an object made up of more polygons is performed using a Z buffer. FIG. 3A shows a state in which a scene in which spherical objects 302 and 303 made of a large number of fine polygons are floating in a room object 301 having a depth is drawn. FIG. 3B shows a drawing result projected on the screen 300. As described above, a more complicated three-dimensional image can be generated by using the Z buffer.

  However, it is difficult to obtain a real feeling like a live-action by simply removing the hidden surface of the polygon using the Z buffer. This is because the three-dimensional spread of the brightness of the objects arranged in the three-dimensional space is not considered. That is, the color values attached to the vertices and pixels of the polygon are unnatural to the eye, and thus lack the reality compared to the actual image. Therefore, in applications where a more realistic graphics image is required, a virtual light source is arranged in a three-dimensional space as a means of giving a three-dimensional brightness spread, and is applied to each vertex or pixel of a polygon. It is common practice to calculate the illuminance value. For example, FIG. 4 shows how the virtual point light source 400 is arranged to increase the brightness of the object and enhance the three-dimensional visual effect in the same situation as FIG. .

  According to the literature (OpenGL Specification 1.3), the formula for calculating the illuminance value is represented by the following formula.

  The above illuminance calculation is performed for each RGB color. In the above equation, ME, MA, MD, and MS are parameters relating to the material of the polygon, and indicate the emission color, ambient color, diffuse color, and specular color, respectively. CA is the ambient color of the entire environment (scene). The symbol Σ indicates that addition is performed for a plurality of light sources (i). LA (i), LD (i), and LS (i) are the ambient intensity, diffuse intensity, and specular intensity of the light source (i), respectively. Att (i) and Spot (i) are the attenuation and spotlight terms for light source (i), and Dc (i) and Sc (i) are the diffuse and specular terms for light source (i). Each of these terms is obtained from the interrelationship between the light source (i) and the vertexes or pixels of the polygon. That is, it is a value obtained by calculating the inner product of the vector or calculating the exponent using the position vector of the light source (i) in the three-dimensional space and the position vector or normal vector of the vertex or pixel of the polygon. In this way, in a three-dimensional space, by arranging a light source relative to an object and obtaining an illuminance value for each vertex or pixel of a polygon, the intensity of the three-dimensional brightness can be added, making it more realistic. A near real 3D graphics image can be generated.

  In addition to the above illuminance calculation, fog is known as a means for giving a three-dimensional visual effect. According to the above literature, the fog calculation formula is expressed by the following formula.

  The above fog calculation is performed for each RGB color. Ci is the color value of the vertex or pixel of the input polygon, and Cf is the fog color. f is a blending coefficient for mixing these two colors, and is a function of the distance (Z value) from the vertex or pixel position of the polygon to the viewpoint. The general fog blend coefficient is a decreasing function with respect to the Z value, and the ratio of the fog color Cf increases as the distance from the viewpoint increases, and the ratio of the input color Ci increases as the distance from the viewpoint increases. That is, the polygon far from the viewpoint has a blurry color as if fog (fog) is applied. As described above, fog is one of the methods (depth of field) for blurring an object located at a far position, and does not calculate brightness intensity three-dimensionally like illuminance calculation.

As another method, in Patent Document 1, opacity (α value) is stored in advance in an index color / texture look-up table (LUT), and a Z value is set as an index of the LUT, and α A method of subtracting a value and performing alpha blending according to the α value is disclosed. As with fog, this method also blurs polygons far from the viewpoint, and does not determine the intensity of brightness three-dimensionally as in illuminance calculation.
Japanese Unexamined Patent Publication No. 2002-92629 (page 22, FIG. 14)

  In the three-dimensional graphics drawing, the conventional method has the following problems for the purpose of giving the object a three-dimensional brightness spread. In other words, the illuminance calculation based on the arrangement of the light sources has a problem that the calculation cost is very high as shown in (Equation 1), and hardware resources are increased for high-speed processing. Further, the setting of parameters such as the position of the light source and the material is complicated, and it is not easy for the user to easily program the brightness, that is, there is a problem that the brightness is not programmable.

  As for the method of performing alpha blending by obtaining the α value from the LUT using fog or the Z value as an index, the object can be blurred according to the Z value, but the brightness in the three-dimensional space is adjusted. Can not solve the above problems.

  In order to solve the above problems, the present invention takes the following measures. That is, a plurality of gamma correction means for converting the color value of each pixel, and selecting one of the plurality of color values output from the plurality of gamma correction means according to the depth value of each pixel, Brightness adjustment according to the depth value is performed by using the selected color value as the color value of the pixel.

  Alternatively, a plurality of gamma correction means for converting the color value of each pixel, and selecting two of the plurality of color values output from the plurality of gamma correction means according to the depth value of each pixel, The selected two color values are linearly interpolated using the depth value, and the linearly interpolated color value is used as the color value of the pixel, thereby adjusting the brightness according to the depth value.

  Alternatively, a plurality of gamma correction means for converting the color value of each pixel are provided, and one of the plurality of color values output from the plurality of gamma correction means according to the screen coordinate and depth value of each pixel is obtained. The brightness is adjusted according to the screen coordinates and the depth value by selecting and using the selected color value as the color value of the pixel.

  According to the present invention, since the user can freely and easily set a gamma correction curve corresponding to the depth, there is an effect that the brightness in the three-dimensional space is highly programmable. Further, according to the present invention, the calculation cost is lower than that of the conventional illuminance calculation, and it is possible to perform high-speed processing with fewer hardware resources.

  In addition, as an effect of gamma correction according to the Z value in terms of image quality, for example, if a plurality of gamma correction curves are set so that the ratio of the output gradation to the input gradation becomes smaller as the Z value becomes larger, it is farther away. It is possible to produce a visual effect that the pixel at the position becomes darker overall for all gradations. In addition, if a gamma correction curve is set so that the degree of nonlinearity increases as the Z value decreases, the closer the pixel is, the brighter the gradation near the halftone and the darker the gradation becomes darker. Therefore, it is possible to produce a visual effect that the contrast is increased. As a result, a gamma correction curve, that is, a difference in brightness and contrast is generated between the pixel having a large Z value and the pixel having a small Z value, and a sense of depth is further obtained, so that a more realistic three-dimensional graphics image can be obtained. There is.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
FIG. 1 shows a configuration according to Embodiment 1 of the present invention. In FIG. 1, the input color Cin is input to the gamma correction unit 100. The gamma correction unit 100 includes four look-up tables LUT0 to LUT3, and different gamma correction curves are set in advance. Each look-up table receives an input color Cin as an index, and outputs a color value corresponding to the index, that is, a gamma-corrected color value. Here, if the input color value Cin is, for example, 8 bits, the lookup table index is 8 bits, and an 8-bit gamma-corrected color value is output according to 256 different index values. Note that it is assumed that the input color Cin of this embodiment is not an RGB color value but a luminance value.

  Examples of gamma correction curves set in these lookup tables LUT0 to LUT3 are shown in FIG. In FIG. 5, a gamma correction curve is set such that the ratio of the output gradation to the input gradation becomes smaller and the degree of nonlinearity becomes smaller as LUT0 → LUT1 → LUT2 → LUT3 changes.

  The selection signal generation unit 101 is a block that extracts specific 2 bits from the Z value of a pixel that is currently drawn. The extracted 2-bit Z value is input to the multiplexer 102 as the gamma correction curve selection signal Zsel. As the gamma correction curve selection signal Zsel, it is simplest to extract 2 bits in order from the most significant bit of the Z value, but it may be configured to extract any continuous 2 bits. . For example, since the Z value may be biased to a specific range depending on the scene, not the most significant 2 bits of the Z value but the lower 2 bits can be adopted as the gamma correction curve selection signal Zsel. In such a case, the selection signal generation unit 101 may be configured to be able to set in advance which bits of the Z value are to be extracted.

  The multiplexer 102 is a block that selects any one of the four gamma corrected colors that are the output of the gamma correction unit 100 in accordance with the gamma correction curve selection signal Zsel. When Zsel is “00”, the output of the lookup table LUT0 is selected, when Zsel is “01”, the output of the lookup table LUT1 is selected, and when Zsel is “10”, the lookup table is selected. When the output of LUT2 is selected and Zsel is “11”, the output of lookup table LUT3 is selected. In this manner, the final color value Cout after gamma correction corresponding to the Z value is obtained by the multiplexer 102.

  FIG. 6 shows the correspondence between the three-dimensional space and the gamma correction curve in the present embodiment. In FIG. 6, it is assumed that a scene in which spherical objects 602 and 603 made of a large number of fine polygons are floating in a room object 601 having a depth is drawn. At this time, the gamma correction curves (A) to (D) of FIG. 5 are applied to the pixels drawn in the four regions (A) to (D) divided by the Z value. That is, a different gamma correction curve is applied according to the Z value of the pixels constituting the object, and as a result, a three-dimensional graphics image as shown in FIG. 7 is generated.

  As described above, an arbitrary gamma correction is applied to the color value of the pixel that is currently drawn based on the Z value of the pixel. Accordingly, for example, if a gamma correction curve is set as shown in FIG. 5, a gamma correction curve having a smaller ratio of output gradation to input gradation as the Z value increases is selected. This pixel becomes darker for all gradations as a whole. In addition, since a gamma correction curve with a higher degree of non-linearity is selected as the Z value decreases, the closer the pixel is, the brighter the gradation near the halftone and the darker the gradation becomes darker. Therefore, the contrast is increased. As a result, a pixel having a large Z value and a pixel having a small Z value have a difference in gamma correction curves, that is, brightness and contrast, and a sense of depth is further obtained, so that a more realistic three-dimensional graphics image can be obtained.

  In the present embodiment, the gamma correction unit 100 includes four lookup tables, but the number of lookup tables is not limited to this. In this case, since the number of bits of the gamma correction curve selection signal Zsel changes according to the number of lookup tables, it is necessary to make the selection signal generation unit 101 and the multiplexer 102 correspond to the number of bits of Zsel. The gamma correction curve is not limited to the example shown in FIG. For example, contrary to FIG. 5, it is possible to increase the brightness or increase the contrast as the Z value increases. In the present embodiment, it is assumed that the input color Cin is a luminance value, not an RGB color value, but may be an RGB color value. In that case, in order to perform gamma correction for each RGB, the number of lookup tables and multiplexers is three times as many.

(Embodiment 2)
FIG. 8 shows a configuration according to the second embodiment of the present invention. 8A is the same as that in FIG. 1 with respect to the gamma correction unit 800, but there are two types of outputs from the multiplexer 802 and the input to the linear interpolation unit 803, and the generation of the selection signal. The difference is that the unit 801 outputs the mixing rate Zα to the linear interpolation unit 803 in addition to the gamma correction curve selection signal Zsel.

  This embodiment is characterized in that the color boundary is made inconspicuous by filtering outputs from two different look-up tables by a linear interpolation unit 803. There is such a need when there is a large difference in the set values of two adjacent gamma correction curves among the four gamma correction curves in FIG. ) Or (between B-C) or (between C-D), there is a possibility that a color step will appear in the pixels near the boundary.

  The multiplexer 802 selects two of the total four outputs of the lookup tables LUT0 to LUT3 according to the gamma correction curve selection signal Zsel from the selection signal generation unit 801. When Zsel is “00”, the outputs of the lookup tables LUT0 and LUT1 are selected, when Zsel is “01”, the outputs of the lookup tables LUT1 and LUT2 are selected, and when Zsel is “10” When the outputs of the lookup tables LUT2 and LUT3 are selected and Zsel is “11”, only the output of the lookup table LUT3 is selected.

  As the mixing ratio Zα to the linear interpolation unit 803, 8 bits below the gamma correction curve selection signal Zsel of the Z value can be used as shown in FIG. Using the mixing ratio Zα and the outputs Cm0 and Cm1 of the multiplexer 802, filtering can be performed as follows.

  In the present embodiment, the gamma correction unit 800 is composed of four lookup tables, but the number of lookup tables is not limited to this. In this case, since the number of bits of the gamma correction curve selection signal Zsel changes according to the number of lookup tables, the selection signal generation unit 801 and the multiplexer 802 need to correspond to the number of bits of Zsel. In the present embodiment, the mixing rate Zα is 8 bits, but may be other than this. In the present embodiment, it is assumed that the input color Cin is a luminance value, not an RGB color value, but may be an RGB color value. In that case, in order to perform gamma correction for each RGB, the number of lookup tables, multiplexers, and linear interpolation units is required to be three times.

(Embodiment 3)
In the first and second embodiments described above, a gamma correction curve is assigned by dividing the three-dimensional space in the Z-axis direction as shown in FIG. 6, but in this embodiment, as shown in FIG. The gamma correction curve is assigned by dividing the region not only in the axis but also in the X-axis and Y-axis directions. Alternatively, as shown in FIG. 10, it is possible to divide the region in the direction of an arbitrary vector P defined in the three-dimensional space. In such a case, it is necessary to generate a gamma correction curve selection signal from the X, Y, and Z values of the pixel.

  The configuration of the present embodiment is shown in FIG. FIG. 11 differs from FIG. 1 in that the input to the selection signal generation unit 1101 is the X, Y, and Z values of the pixel. The selection signal generation unit 1101 needs to output a 2-bit gamma correction curve selection signal sel. However, in the area division as shown in FIG. In the area division as shown in FIG. 9B, it is only necessary to take one higher bit from each of the Z value and the Y value.

  In the area division as shown in FIG. 10, the next area determination function F is obtained with the three-dimensional space coordinates of the vector starting point P as (Xp, Yp, Zp), and the upper 2 bits are output as the gamma correction curve selection signal sel. That's fine.

  However, in (Equation 4), sqrt represents a square root operation. Since the calculation of (Equation 4) is somewhat complicated and the calculation cost is high, the configuration of the selection signal generation unit can be simplified by taking the starting point of the vector P as the origin or omitting the square root calculation. it can. In the present embodiment, the three-dimensional space is divided into four regions and different gamma correction curves are assigned, but it may be further divided into more regions.

(Embodiment 4)
FIG. 12, FIG. 13 and FIG. 14 show an embodiment relating to a graphics drawing apparatus having a gamma correction unit based on Z values according to the present invention. FIG. 12 shows the configuration of the highest hierarchy of the graphics drawing apparatus, which is composed of a CPU (Central Processing Unit) 1201, a graphics engine 1202, a memory controller 1203, a display controller 1204, a memory device 1205, and a display device 1206. Yes. Note that these blocks may be composed of separate semiconductors, or some blocks may be integrated as a single system LSI. The memory controller 1203 controls the memory device 1206, and the display controller 1204 controls the display device 1207.

  The graphics engine 1202 is connected to and controlled by the CPU 1201 via the CPU bus 1207. The graphics engine 1202 is connected to the memory controller 1203 via the memory bus 1208 and can access data on the memory device 1205. The memory device 1205 stores color buffer and Z buffer pixel data, a display list, texture data, and the like that are used in three-dimensional graphics rendering by the graphics engine 1202. Note that a memory for storing these data may exist inside the graphics engine 1202 in order to increase the memory access speed.

  Next, the internal configuration of the graphics engine 1202 is shown in FIG. The host interface 1301 is connected to the CPU bus and communicates with the CPU. The graphics controller 1302 inputs polygon drawing data to the graphics pipeline 1304 in accordance with an instruction from the CPU. The polygon drawing data may be included in a command sequence from the CPU, or may be included in a display list on the memory device 1205. In the latter case, the graphics controller 1302 reads the display list via the memory bus interface 1303 and acquires polygon drawing data. The graphics pipeline 1304 is divided into the following three main blocks. The vertex processing unit 1305 is a block that performs processing such as coordinate conversion and illuminance calculation for each vertex of the polygon. The rasterizer 1307 is a block that scans the inside of a polygon in units of pixels and performs hidden surface processing using Z values, texture mapping, various blending processing, and the like for each pixel. The setup unit 1306 is between them, and obtains polygon information such as the polygon shape, the color value inside the polygon, texture coordinates, and the like from the three vertex data that have been subjected to coordinate conversion and illuminance calculation output from the vertex processing unit 1305. Then, the data is converted into a data format that can be input to the rasterizer 1307. FIG. 13 shows the configuration of the graphics engine in the present invention, but a general graphics engine has almost the same configuration, and the configuration characterizing this embodiment is in the lower layer than the rasterizer 1307.

  FIG. 14 shows the internal configuration of the rasterizer. In FIG. 14, a pixel coordinate generation unit 1401 scans the inside of a polygon in units of pixels, and obtains position coordinates (X coordinate and Y coordinate) of pixels included in the polygon. A texture address generation unit 1402, a color generation unit 1403, a fog generation unit 1404, and a Z value generation unit 1405 obtain the texture coordinate value, color value, fog value, and Z value of the pixel from the position coordinates, respectively. The texture color generation unit 1406 reads the texture color on the external memory (memory device 1205 in FIG. 12) via the memory bus interface, and performs processing such as filtering. The texture color output from the texture color generation unit 1406 is blended with the color output from the color generation unit 1403 and the texture environment calculation unit 1407, and then the fog value output from the fog generation unit 1404 and the fog calculation unit 1408 Blended and sent to the gamma correction unit 1409 based on Z value. Needless to say, when texture mapping is not performed, the texture environment calculation unit 1407 outputs the color input from the color generation unit 1403 as it is. The same is true when fog is not performed. 14 includes a fog generation unit 1404 and a fog calculation unit 1408, but the configuration of this embodiment does not impair the characteristics of the present embodiment even if these blocks are not present. In FIG. 14, the block that characterizes this embodiment is a gamma correction unit 1409 based on Z values. The internal configuration of this block is the same as the configuration shown in FIGS. 1 and 8 already described in the first and second embodiments. That is, the gamma correction unit 1409 based on the Z value has a plurality of lookup tables that can store gamma correction curves, and selects one of the lookup table outputs according to the Z value.

  The lookup table may be configured so that it can be shared by the Z-value gamma correction unit 1409 and the texture color generation unit 1406. This is due to the following reasons. Conventionally, a texture data format called “index texture” is generally used for the purpose of reducing the capacity of texture data in an external memory. In this type of texture, a look-up table is required inside the texture color generation unit 1406. However, recent advances in semiconductor technology tend to increase the memory bus bandwidth and increase the usable memory capacity. It is also true that the situation where index textures must always be used due to the demand for higher image quality. Therefore, if the same look-up table can be used for the conventional index texture and the gamma correction using the Z value of the present invention, the circuit can be used efficiently and the increase in circuit area can be suppressed. is there.

  In the pixel test unit 1410, an alpha test, a stencil test, a Z test (hidden surface processing using Z values), and the like are performed. When the Z test is performed, the Z value output from the Z value generation unit 1405 is compared with the Z value read from the external memory, and only the pixels for which the former is determined to be smaller (in front) than the latter. Passes this test. In the destination blend unit 1411, the gamma corrected color output from the gamma correction unit 1409 based on the Z value and the color value read from the rendering buffer are blended as necessary. Finally, the color value after blending and the Z value generated by the Z value generation unit 1405 are written into the external memory.

  A graphics drawing apparatus having a gamma correction function based on a depth value according to the present invention is a variety of electronic devices equipped with a graphics drawing function, such as a mobile phone, a PDA, a digital TV, a car navigation system, a home game machine, and a personal computer. It can be used in a computer or the like.

The figure which shows the structure in Embodiment 1 of this invention. The figure for demonstrating the hidden surface removal using Z buffer in a prior art The figure for demonstrating the hidden surface removal using Z buffer in a prior art The figure for demonstrating the illumination intensity calculation by the light source in a prior art The figure which shows the example of a setting of the gamma correction curve in Embodiment 1 of this invention The figure which shows the correspondence of the three-dimensional space and gamma correction curve in Embodiment 1 of this invention The figure which shows the three-dimensional graphics image produced | generated in Embodiment 1 of this invention. The figure which shows the structure in Embodiment 2 of this invention. The figure which shows the correspondence of the three-dimensional space and gamma correction curve in Embodiment 3 of this invention. The figure which shows the correspondence of the three-dimensional space and gamma correction curve in Embodiment 3 of this invention. The figure which shows the structure in Embodiment 3 of this invention. The figure which shows the whole structure in Embodiment 4 of this invention. The figure which shows the structure of the graphics engine in Embodiment 4 of this invention. The figure which shows the structure of the rasterizer in Embodiment 4 of this invention.

Explanation of symbols

100, 800, 1100 Gamma correction unit 101, 801, 1101 Selection signal generation unit 102, 802, 1102 Multiplexer 200, 300 Screen 201-204 Polygon 301-303, 601-603 Drawing object 400 Light source 803 Linear interpolation unit 1201 CPU
1202 Graphics Engine 1203 Memory Controller 1204 Display Controller 1205 Memory Device 1206 Display Device 1207 CPU Bus 1208 Memory Bus 1301 Host Interface 1302 Graphics Controller 1303 Memory Bus Interface 1304 Graphics Pipeline 1305 Vertex Processing Unit 1306 Setup Unit 1307 Rasterizer 1401 Pixel Coordinate Generation Unit 1402 texture address generation unit 1403 color generation unit 1404 fog generation unit 1405 Z value generation unit 1406 texture color generation unit 1407 texture environment calculation unit 1408 fog calculation unit 1409 gamma correction unit based on Z value 1410 pixel test unit 1411 destination blend Part

Claims (5)

A plurality of gamma correction means for converting the color value of each pixel; and selecting one of a plurality of color values output from the plurality of gamma correction means according to a depth value of each pixel, A graphics drawing apparatus, wherein brightness adjustment is performed in accordance with the depth value by using the color value obtained as the color value of the pixel. A plurality of gamma correction means for converting the color value of each pixel; and selecting two of the plurality of color values output from the plurality of gamma correction means according to the depth value of each pixel, The two color values are linearly interpolated using the depth value, and the brightness value is adjusted according to the depth value by using the linearly interpolated color value as the color value of the pixel. Graphics drawing device. A plurality of gamma correction means for converting the color value of each pixel are provided, and one of the plurality of color values output from the plurality of gamma correction means is selected according to the screen coordinates and the depth value of each pixel. A graphics drawing apparatus, wherein brightness adjustment is performed in accordance with the screen coordinates and the depth value by using the selected color value as a color value of the pixel. The graphics drawing apparatus according to claim 1, wherein the gamma correction unit is a look-up table. 5. The graphics drawing apparatus according to claim 4, wherein the lookup table is shared for generating an index texture.

Images