JPH05298455A - Texture mapping device - Google Patents

Abstract

PURPOSE:To realize displaying with smooth texture by mixing texture data where weighing is added and RGB luminance owing to optical reflection, executing plotting and cumulatively adding a picture being a result. CONSTITUTION:Luminance is calculated (13b) by an optical reflection expression through the use of light source information, a reflection coefficient and a normal vector. Then, a polygon is dissolved to spans and information at the end point of a span to be plotted, that is, the coordinate of the both end points, a texture coordinate and RGB luminance owing to optical reflection are obtained from the value of a vertex. The increase portions of RGB luminance, the Z coordinate and the texture coordinate for DDA are obtained (13d). A picture element for indicating whether a texture picture element to be the object of mapping is the element of left down, left up, right down or right up is designated (13e). Moreover, weighing by a decimal part of the coordinate is added to data which is fetched by the texture coordinate and picture element designition, RGB luminance owing to optical reflection is mixed and the element on the span is plotted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a graphic display, and more particularly to a texture mapping device for displaying a realistic display with a smooth material appearance at high speed with a relatively small scale hardware.

[0002]

2. Description of the Related Art Before discussing a conventional texture map device, first, a normal three-dimensional graphic display process will be described. In the 3D graphic display, the displayed 3D graphic is
It is usually approximated to a polyhedron, and is made by sequentially displaying the polygons forming the polyhedron. That is, for the three-dimensional graphic display, a polygonal display should be considered. Figure 9
The fill processing, which is the main processing of the polygon display processing, is shown in FIG. Information on the vertices V1 to V4 of the polygon, that is, the coordinates of the vertices (Xi, Yi, Zi), RGB brightness (ri, gi, b).
i) and a texture map, based on the texture coordinates (si, ti) (both i = 1 to 4), a horizontal line segment within the polygon (hereinafter referred to as a span) from the lower end of the polygon Are sequentially written into the frame buffer (buffer for storing image information of the display screen). That is, it is realized by a method in which color data, that is, RGB luminance, is obtained for each pixel on the span, and this is sequentially written in the corresponding portion of the frame buffer.

The RGB brightness at the apex is calculated from the light source, the direction of the surface and the reflection coefficient of the surface by the light reflection model shown in FIG. That is, the light source light intensity Ip from the light source, the ambient light intensity Ia that is light incident in various directions from the periphery, the unit normal vector N, the unit vector L to the light source, the unit vector V to the viewpoint direction, A vector H obtained by normalizing the sum of L and V, an ambient light reflection coefficient K of the display object
From the values of a, diffuse reflection coefficient Kd, and specular reflection coefficient Ks, FIG.
It can be calculated by the formula shown in (1).

This equation is for one component of RGB,
This formula may be calculated for each RGB.

The first item (1) shows the components of ambient light reflection, the second item the diffuse reflection, and the third item the specular reflection component.
Ambient light reflection and diffuse reflection are reflections that are affected by the color of an object, and specular reflection is for expressing the gloss, or highlight, of an object, and is not significantly affected by the color of the object.
The color of the light source is output as is. That is, while the ambient light reflection coefficient and the diffuse reflection coefficient have different values between RGB depending on the color of the display object, the values of the specular reflection coefficient do not change much between RGB.

Originally, the RGB luminance should be calculated for each pixel in (1) shown in FIG. 10, but there is a problem that the calculation takes time. In order to shorten the time, it is usual to calculate only the values of the vertices of the polygon and obtain the brightness of each pixel by interpolation. Specifically, the RGB luminances at both ends of each span may be obtained by interpolation from the values of the vertices, and the RGB luminances of pixels on the span may be obtained by interpolation from the values at both ends.

Based on the above description of the general three-dimensional graphic display processing, a conventional texture mapping device will be described by taking Japanese Patent Laid-Open Publication No. 3-138779 as an example. As described in the publication, the RGB luminance obtained by the luminance calculation by the light source and the texture data are subjected to an operation such as addition or multiplication and mixed, and the mixture is displayed. That is,
As shown in FIG. 11, by the DDA group, X and Y coordinates of pixels to be written in the frame buffer, RGB luminance,
And texture addresses s, t are determined. The address range of the texture addresses s and t is limited within the valid range by the replacement unit, and the texture data is read from the texture buffer by these outputs. The read texture data is mixed with RGB brightness in the calculation unit, written into the frame buffer and displayed, thereby realizing texture mapping.

FIG. 7 shows an example of texture data. Here, a circle indicates a pixel of texture data, that is, a texture pixel, coordinates s and t are given to the texture data, and the center of the texture pixel indicates an integer coordinate value. In this example, the texture data has a cross shape.

At this time, the texture coordinates of each pixel obtained by DDA are not limited to integer values, but have values below the decimal point. In this case, it is necessary to consider not only the closest one texture data but also the influence of the adjacent texture pixel.

As shown in FIG. 12, if the pixels on the span having the Y coordinate of Ys and the texture coordinates corresponding to the pixels 0 to 8 are moved from P1 to P8 on the line, the closest texture is obtained. When only pixels are used, as indicated by the one-dot chain curve in the graph of the change in texture data of pixels on the span, one pixel on the span suddenly changes to another texture data in a stepwise manner, and a smooth texture can be expressed. Disappear.

As a measure against this, it is assumed that the texture data changes with a gradient shown by a broken line, and a method of interpolating between texture pixel centers is considered as a curve of the improved method shown by a solid line. There is. In order to realize this method, the function to take a weighted average by the proximity of texture pixels of neighboring texture pixels is added to the part that accesses the hardware texture data, considering the decimal part of the texture coordinates. However, there is a problem that the hardware becomes complicated and the quantity increases.

[0012]

In the above-mentioned prior art, since the decimal part of the texture coordinate obtained in the texture coordinate calculation is not taken into consideration, it is difficult to display more realistically.
Alternatively, if a function for calculating a weighted average of neighboring texture pixels is added to a portion for controlling access to texture data to realize a smooth texture mapping display, there is a problem that the hardware becomes large in scale. An object of the present invention is to realize a smooth textured display using a relatively small-scale hardware.

[0013]

As a first means, the value of the decimal part of the texture coordinates of each pixel on the span
The weighted texture data and the RGB brightness due to light reflection are mixed and drawn, and this is repeated for adjacent texture pixels, and the resulting images are cumulatively added.

As a second means, minute displacements are given in various directions with respect to the texture coordinates given to the display object,
The conventional method, that is, the method of using the value of the texture pixel closest to the texture coordinate is repeatedly displayed,
By adding a predetermined weight to the resulting image and performing cumulative addition. This aims to capture the influence of not only the nearest texture pixel but also neighboring texture pixels by giving a displacement. Smooth texture mapping is realized by considering neighboring texture pixels.

[0015]

In the case of the first means, the four closest texture pixels are weighted according to the distance from the texture coordinates, mixed with the RGB luminance due to the reflection of light, and the results are cumulatively added. Normally, luminance mixing is linear in addition and multiplication, so the result of cumulative addition after luminance mixing and
A value obtained by adding weights to the texture data and cumulatively added and the result of mixing the RGB luminance due to light reflection match. Therefore, smooth texture mapping can be performed by corresponding to the texture data obtained by interpolation from the values of neighboring texture pixels.

Further, in this method, the value of one texture pixel is fetched for each pixel as in the conventional method, and the RGB brightness obtained by the light reflection method or the like may be mixed. On the other hand, it is only necessary to add the function of weighting the texture pixels by the decimal part of the texture coordinates and the designation of which of the adjacent texture pixels is used.

The second means will be described below. FIG. 14 shows an example of how to give a minute displacement to the texture coordinates. Here, the texture coordinate of a pixel is indicated by P. With the conventional method of using the value of the texture pixel closest to the texture coordinate, the value of the texture pixel (i, j) is used. Here, the texture coordinates are given a small displacement with respect to P. For example, by giving a small displacement, the coordinates of P1 to P8 can be given, and P,
In the case of P5, P6 and P7, the texture pixel (i,
j) is used, for P3 and P4, texture pixel (i, j + 1) is used, for P2 texture pixel (i, j + 1) is used, and for P1 and P8, texture pixel (i + 1, j + 1) is used. ) Is used.
As can be seen from this example, in a total of 9 rendering processes, the texture pixel closer to the texture coordinate P is processed more frequently. That is, the drawing process is performed approximately times according to the distance from the texture coordinates. For this reason, the value of the texture pixel closer to the texture coordinate is more reflected in the cumulative result luminance.
As described above, since the cumulative result of mixing the RGB luminance and the texture data due to the reflection of light is the same as the cumulative texture data of the RGB luminance, the final result image is , Is close to a mixture of texture data obtained by interpolating the values of adjacent texture pixels. In this way, smooth texture mapping can be realized.

The reason why the predetermined weight is added and accumulated is to prevent the brightness from overflowing as a result of the accumulation process and to set a desired value.

Further, as the hardware, a weighted cumulative addition function of the created image is sufficient, and the conventional hardware often already has the same function, and the increase of the hardware physical quantity is slight.

[0020]

Embodiments of the present invention will be described below with reference to FIGS. 1 to 6, FIG. 8, FIG. 9, FIG. 13 and FIG. FIG. 2 shows the overall configuration of a system that realizes the present invention. 1 is CPU
In order to display a three-dimensional figure with a texture map,
Display processing such as coordinate conversion, clipping, and brightness calculation,
It controls the entire system such as the rendering processor 7, the frame buffer 8, the Z buffer 9, the texture buffer 10, and the DAC 11. The main memory / system bus control 2 controls access to the main memory of the CPU 1 and the system bus 4. System bus 4 is CP
Data and control signals are transferred between the U1 and the main memory 3 and the graphic bus control 5. Also, other input / output devices (not shown) are connected to the CPU 1 and the main memory 3. The graphic bus control 5 sends and receives data and control signals between the CPU 1 and the main memory 3, and also controls the graphic bus 6. The graphic bus 6 transfers data and control signals between the CPU 1 and main memory 3 and the rendering processor 7, frame buffer 8, Z buffer 9, texture buffer 10 and DAC 11 via the graphic bus control 5. The rendering processor 7 draws a span or the like in the frame buffer 8 according to an instruction from the CPU 1. The frame buffer 8 stores the image displayed on the monitor 12. The Z buffer 9 is used for hidden surface removal. The texture buffer 10 stores texture data to be texture-mapped. The DAC 11 reads out image information (digital information) on the frame buffer from the frame buffer 8 at a predetermined cycle, converts it into analog information, sends it to the monitor 12, and displays it.

FIG. 1 shows a polygon texture mapping display processing procedure according to the first method. 13 is a flowchart showing the same processing procedure. Most of this processing is CPU1
In the process of step 13f, the CPU 1
Sends command processing parameters to the rendering processor 7 via the system bus 4 and the graphic bus 6, and the rendering processor performs actual processing. In step 13a, a coordinate conversion process is performed to convert the polygonal vertex coordinates and vertex normal vectors. Step 13
In b, the brightness is calculated by the light reflection formula using the light source information, the reflection coefficient, and the normal vector.

In step 13c, clipping processing is performed to cut out the portion of the polygon outside the display range. These processes are widely and generally performed in the three-dimensional graphic display process.

In step 13d, as described above, the polygon is decomposed into spans, and the information of the end points of the span to be drawn, that is, the coordinates of both end points, the texture coordinates, and the RGB brightness due to the reflection of light are used as the vertex values. Ask from. Also DDA
RGB intensity increments, Z coordinate increments and texture coordinate increments for In step 13e, the texture pixel to be mapped is the lower left corner of the texture coordinates,
A texture pixel is designated to indicate whether it is the upper left, lower right, or upper right pixel. In step 13f, the texture coordinate data and the texture data fetched by designating the texture pixels (which are stored in advance in the texture buffer) are weighted by the decimal part of the texture coordinates, and the RGB brightness due to the reflection of light is mixed. Draw pixels on the span. In step 13g, it is determined whether or not the processing for the four neighboring texture pixels has been completed. If not completed, the processing returns to step 13e to repeat the processing. Note that the drawing is performed so that the brightness of the drawing pixel at each time is cumulatively added. In step 13h, it is determined whether drawing of all spans has been completed, and if not completed, step 13d.
Return to and repeat a series of processing.

The span drawing process in the rendering processor 7 will be described in detail below.

FIG. 3 shows an overall block diagram of the rendering processor 7. Graphic bus interface 71
Connects the graphics bus to the rendering processor. The sequencer 72 controls the operation of the rendering processor 7 such as management of the number of pixels generated by DDA. D
The DA unit 73 uses the coordinate values (X, Y,
Z), luminance (R, G, B, α, that is, α indicating RGB luminance and transparency) and texture coordinates (s, t).
To calculate. The luminance blending unit 74 mixes the RGB luminance from the DDA unit 73 and the value obtained by weighting the corresponding texture data loaded in the source buffer 75 with the decimal part (sf, tf) of the texture coordinates and the texture pixel designation. However, the result is the corresponding destination buffer 76 and frame buffer interface 7.
It is written to the corresponding address of the frame buffer 8 via 7. At this time, if necessary, the value obtained by the mixing and the value before the writing at the writing position of the frame buffer 8 may be added to perform the writing.
The destination buffer 76 holds the value of the corresponding pixel before writing in the frame buffer and the value of the Z buffer used for hidden surface removal, and also holds the brightness calculated by the brightness blending unit 74. The frame buffer interface 77 reads the value of the corresponding address of the texture buffer 10 based on the integer part (si, ti) of the texture coordinates output from the DDA unit 73 and the texture pixel designation, and loads it into the source buffer 75.

Further, the frame memory interface 77
Uses the pixel coordinates (X, Y) from the DDA unit 73 to capture the value of the corresponding portion of the frame buffer 8 and the Z buffer 9 into the destination buffer 76, and also the brightness or Z in the destination buffer 76.
The coordinates are written in the corresponding portion of the frame buffer 8 or Z buffer 9. Hereinafter, each part of the render processor 7 will be described.

FIG. 4 shows a block diagram of the DDA section 73.
The DDA unit 73 includes an XY DDA 731 and a st DDA7.
32, Z DDA733 and RGBα DDA734
These are RGBα (RGB) for calculating the generated pixel coordinate, generating the texture coordinate of the texture to be mapped, and generating the Z coordinate of the generated pixel.
It is used for calculation of α value indicating brightness and transparency.
In this way, it can be seen that various coordinates and luminance information regarding each pixel required for span drawing are generated.

FIG. 5 shows a block diagram of the luminance blending section 74. The luminance blending unit 74 is a texture synthesizing unit 741,
The texture blending unit 741 includes an α blend unit 742, a raster operation unit 743, and a Z comparison unit 744.
Then, the texture data from the source buffer 75 is added with a weight calculated by pixel selection that specifies a decimal part of texture coordinates and a texture pixel to be used among nearby texture pixels, and RGBα from the DDA unit 73.
The luminance is mixed by a predetermined method and sent to the α blending section 742. The α blending section 742 performs so-called α blending, that is, the RGBα luminance to be written and the RGBα luminance stored in the frame buffer according to the α value. Mix at a fixed rate. Raster operation unit 7
The so-called raster operation process 43 and the process of stopping the writing of pixels are performed by the output of the Z comparator 744 for the so-called raster operation process and hidden surface removal. The Z comparator 744 is the DDA unit 7
The Z coordinate value from 3 is compared with the Z coordinate stored in the Z buffer, and if the predetermined condition is not satisfied, the raster operation unit 743 is instructed to inhibit writing of pixels. In this way, in addition to α blending processing and hidden surface removal processing, pixel drawing processing with texture mapping can be realized.

FIG. 6 shows the texture synthesizer 741. The texture synthesizing unit 741 includes a luminance synthesizing circuit 741a for executing the texture synthesizing process itself, a selector 741b and a texture data weighting circuit 741c. The selector unit 741b selects the same data according to the type of texture data to be used. That is, as texture data, (1) only It, which is information on brightness, (2) It and αt, (3)
RGB brightness of texture, Rt, Gt and Bt, (4) R
Four types of texture data of t, Gt, Bt and αt are selected.

The texture data weighting circuit 741c weights the texture data by the fractional part (sf, tf) of the texture coordinates and the pixel selection. In order to obtain this weight, a weight value table 741d for storing the weight values calculated in advance for a series of values of the pixel selection and the decimal part of the texture coordinates is provided. The weight value can be calculated at high speed simply by referring to this table. FIG. 13 shows the processing of the texture data weighting circuit 741c. First, by the pixel selection, a weight value table provided for each pixel selection is selected. Predetermined bits are cut out from the decimal point of the DDA for calculation of the texture coordinates s, t, sf ', tf' are obtained and merged, and the selected weight value table is drawn using this value as an index to obtain the weight value. ..

The luminance synthesizing circuit 741a performs the synthesizing process shown in the table of FIG. 8 such as addition and multiplication.

That is, as the mixing method, there are multiplication and addition, and RG depends on the type of texture data.
The output values of B brightness and α are as shown in the table of FIG. Here, the texture data is the texture data weighting circuit 74.
It is a value weighted by 1c.

When it is not necessary to apply the texture map, the brightness is not mixed, and the input information (RGBα) from the DDA unit 73 is output as it is.

From the above description, it is clear that the rendering processor 7 can draw spans with and without texture maps.

Further, considering the processing flowchart shown in FIG. 1, the value of the texture data at the point designated by the texture coordinates calculated for each display pixel is a weighted average of the values of neighboring texture pixels, that is, the interpolation. I think it became clear that I could do what I did.

For example, the following interpolation is possible.
That is, the decimal parts sf and tf of the texture coordinates are used, and regarding the values of the texture pixels adjacent to the texture coordinates, the lower left pixel is Ilb, the upper left pixel is Ilt, the lower right pixel is Irb, and the upper right pixel is Irt. , Expressed by the equation 1,
Interpolation commonly called bilinear interpolation can be performed. This is for performing linear interpolation in both directions of s and t, and is considered to be effective in the interpolation of texture data.

[0037]

## EQU00001 ## I = (1-sf) (1-tf) Ilb + (1-sf) tfIlt + sf (1-tf) Irb + sftfIrt In order to realize this interpolation, the weight value table 741
The corresponding pixel selection area of c is (1-sf) (1-t
f), (1-sf) tf, sf (1-tf), and sftf may be set to a series of values and the process of the flowchart 13 may be performed.

In the case of the second means, the rendering processor 7 is the conventional one, that is, the texture synthesizing unit 74 does not have the texture data weighting circuit 741c, and the frame buffer interface 77 always has the texture pixel closest to the texture coordinate. Is taken from the texture buffer 10.

The processing procedure is shown in the flowchart 14 of FIG. Step 14a is coordinate conversion, step 14b is brightness calculation, step 14c is clipping processing, and step 14d is span end point information calculation, which are the same as the processing of the flowchart shown in FIG.

In step 14e, a minute displacement is added to the texture coordinates of the span end point, and in step 14f, span drawing is performed. At this time, the brightness drawn in the frame buffer 8 is obtained by dividing the brightness obtained by texture mapping by the number of times of repeated drawing, that is, a weighted value is added to the contents of the frame buffer. This processing is performed by the function of the α blending unit 742 of the luminance blending unit 74, that is, by multiplying the texture-mapped luminance obtained by the texture synthesizing unit 741 by a constant, adding the result to the content of the frame buffer 8, and writing the result in the frame buffer 8. It can be realized by using the function.

In step 14g, it is determined whether the number of times of drawing with a given displacement has reached a prescribed number. If not, the process returns to step 14e to repeat a series of processes. In step 14h, it is determined whether or not the drawing processing of all the spans of the polygon is completed. If not completed, the processing returns to step 14d to repeat the series of processing.

From the above description, it has been clarified that the conventional hardware enables smooth texture mapping by the second means.

[0043]

As described above, according to the present invention, it is possible to perform mapping with a value obtained by interpolating from the value of a texture pixel in the vicinity even when the texture coordinate of the display pixel is not the center of the texture pixel with a relatively small-scale hardware. It is possible to display various texture mappings.

[Brief description of drawings]

FIG. 1 is a flowchart showing a polygon texture map display processing procedure by a first means.

FIG. 2 is a configuration diagram of the entire system.

FIG. 3 is a block diagram of a rendering processor.

FIG. 4 is a block diagram of a DDA unit.

FIG. 5 is a block diagram of a luminance blending unit.

FIG. 6 is a block diagram of a texture synthesizing unit.

FIG. 7 is a diagram showing an example of texture data.

FIG. 8 is a diagram showing the details of luminance mixing processing.

FIG. 9 is an explanatory diagram of polygon filling processing.

FIG. 10 is an explanatory diagram of a light reflection model.

FIG. 11 is a diagram showing a conventional texture mapping device.

FIG. 12 is an explanatory diagram of a texture data calculation method.

FIG. 13 is an explanatory diagram of processing in a texture data weighting circuit.

FIG. 14 is a diagram showing an example of how to apply a minute displacement to texture coordinates.

FIG. 15 is a flowchart showing a polygon texture map display processing procedure by the second means.

[Explanation of symbols]

1 ... CPU, 2 ... Main memory / system bus control, 3 ... Main memory, 4 ... System bus, 5 ... Graphic bus control, 6 ... Graphic bus, 7 ... Rendering processor 8 ... Frame buffer, 9 ... Z buffer, 10 ... Texture buffer, 11 ... DAC, 12 ... Monitor, 13 ...
The flowchart which shows the processing procedure by the 1st means of this invention, 14 ... The flowchart which shows the processing procedure by the 2nd means of this invention.

Claims (6)

[Claims] 1. Geometrical information of a shape relating to a display object,
Based on the texture data that specifies the surface pattern, the texture coordinates that show the correspondence between each part of the display object and the texture data, and the viewpoint information that is virtually given to set the displayed situation. By mapping the data to the object surface, in the three-dimensional graphic display that expresses the details of the object surface, the value of the texture pixel near the position indicated by the texture coordinates in the display pixel is mapped to obtain the brightness of the display pixel. A feature is that weighting is added to the values, cumulative addition is performed, and this series of processing is repeated so as to spread over a plurality of neighboring texture pixels, and the cumulative result is obtained as the brightness of the display pixel to realize smooth texture mapping. Texture mapping device. 2. The texture mapping device according to claim 1, wherein the weight is determined by a value of a deviation from a position indicated by texture coordinates to a center of the texture pixel. 3. When obtaining the weight from the deviation,
The texture mapping device according to claim 2, further comprising a table for storing a weight value calculated for a series of deviations in advance, and referring to the table. 4. A texture pixel that is considered to be closest to a position indicated by the texture coordinate is obtained by adding a small displacement to the original texture coordinate in order to obtain the texture pixel in the vicinity. Texture mapping device. 5. Geometrical information about the shape of the display object,
Based on the texture data that specifies the surface pattern, the texture coordinates that show the correspondence between each part of the display object and the texture data, and the viewpoint information that is virtually given to set the displayed situation. By mapping the data to the object surface, in a three-dimensional graphic display that represents the details of the object surface, a texture that gives a small displacement to the original texture coordinates that show the relationship between the displayed object and the texture data and maps it to the display pixel. As a value of, the texture mapping is performed using the value of the texture pixel that is considered to be closest to the position indicated by the texture coordinates corresponding to the display pixel, and the image is drawn, weighted and accumulated, and this processing is performed. Repeat the display for various small displacements to obtain a display image and perform a smoother texture mapping display. A texture mapping method characterized by revealing. 6. Geometrical information about the shape of the displayed object,
Based on the texture data that specifies the surface pattern, the texture coordinates that show the correspondence between each part of the display object and the texture data, and the viewpoint information that is virtually given to set the displayed situation. By mapping the data to the object surface, in the three-dimensional graphic display that represents the details of the object surface, the value of the texture data of the display pixel is interpolated from the texture pixel near the position indicated by the texture coordinate of the pixel. Therefore, the texture mapping device is characterized in that the mapping is repeated by the values of the texture pixels in the vicinity, and the result is weighted and cumulatively added to obtain a display image.