JP2002342780A - Image generation system, image generation method, and information storage medium - Google Patents

Abstract

(57) [Summary] (With correction) [PROBLEMS] To provide an image generation system, an image generation method, and an information storage medium capable of performing mapping based on texture information more simply and at high speed. SOLUTION: In image mapping processing using texture information including color information and additional information stored for each texel in a predetermined storage area, edge information in a direction oblique to the coordinate direction of the texel is used as texture information. And at least one texel is selected from a plurality of hand cells based on predetermined texel coordinate designation information and edge information, and mapping is performed using the texture information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an image generation system, an image generation method, and an information storage medium.

[0002]

2. Description of the Related Art Texture mapping is performed by transforming (mapping) texels into cells for image display, that is, pixels.

When the mapping is simply performed, the colors and lines of the image do not change smoothly due to the viewpoint and the magnification of the texel. Specifically, for example, as shown in FIG. 17A, normally, only one color of data can be stored in one texel.
As shown in (B), jaggies (step-like jaggies) stand out.

In order to prevent such jaggies, color interpolation is performed so as to smoothly change the colors and lines of an image.

In general, such interpolation techniques include:
In general, a method of mixing (interpolating) colors using texture information of four adjacent texels is used.

[0006] Transfer of texture information for 4 texels,
Complicated processing such as floating-point arithmetic is required for the interpolation, so that processing time is required.

The present invention has been made in view of the above problems, and has as its object to provide an image generation system, an image generation method, and an information storage medium capable of performing mapping based on texture information more simply and at higher speed. Is to do.

[0008]

In order to solve the above-mentioned problems, an image generation system according to the present invention uses an image using texture information including color information and additional information stored for each texel in a predetermined storage area. An image generation system for generating an image by performing a mapping process of
The texture information includes edge information representing an edge of a texture oblique to a direction of a coordinate axis of the texel, and selects at least one texel from a plurality of texels based on predetermined texel coordinate designation information and the edge information. Selecting means, means for performing a mapping process using texture information according to the selected texel,
It is characterized by including.

Further, the information storage medium according to the present invention generates an image by performing an image mapping process using texture information including color information and additional information stored for each texel in a predetermined storage area. In the computer-usable information storage medium storing the program, the texture information includes edge information representing an edge of a texture oblique to the direction of the coordinate axis of the texel, predetermined texel coordinate designation information and the edge information A program for causing a computer to implement a selecting means for selecting at least one texel from a plurality of texels based on the texel, and a means for performing a mapping process using texture information of the selected texel. .

[0010] A program according to the present invention is characterized by including a module for realizing each of the above means.

According to the present invention, since the mapping process is performed by using the edge information indicating the edge of the oblique texture, the linear interpolation process such as the bilinear interpolation process is not required, and the circuit configuration required for the mapping process and the like is eliminated. And the algorithm can be simplified, and the mapping process can be performed at high speed.

[0012] The image generation system may include means for storing the edge information in the storage area as at least a part of the additional information.

The information storage medium and the program may include a program for causing a computer to implement means for storing the edge information as at least a part of the additional information in the storage area.

According to this, by providing the means for storing the edge information, it becomes possible to dynamically store the edge information at the time of performing an image calculation or drawing, and more flexible and versatile texture data can be stored. Can be realized.

In the image generation system, the information storage medium, and the program, the texel coordinate designation information includes a two-dimensional coordinate value, and the selecting means includes
The texel may be selected based on a value of a decimal part of two numerical values of the two-dimensional coordinate value.

According to this, by selecting the texel based on the value of the decimal part of the address of the texel, a more appropriate texel can be selected, and a higher quality image can be generated.

In the image generation system, the information storage medium, and the program, the selection means may select one texel from a plurality of texels near the texel based on the two numerical values.

According to this, the mapping process can be performed at higher speed by selecting one texel instead of a plurality.

In the image generation system, the information storage medium, and the program, the selecting means may determine, based on a value of the decimal portion, an edge represented by the edge information in the texel as a base point in the decimal portion. The texel in the direction where the point to be represented is located may be selected.

According to this, by considering the value of the decimal portion, it is possible to determine, for example, which side is closer to the oblique edge crossing the square texel, that is, the contour (contour). By selecting a texel based on this determination, a more appropriate texel can be selected, and a higher quality image can be generated.

According to the image generation method of the present invention, in the image generation method of generating an image by performing an image mapping process using texture information, a plurality of texels constituting the texture information are provided. For each, assigning edge information representing an edge of a texture oblique to the direction of the coordinate axis of the texel, and performing a mapping process of mapping an image based on the assigned edge information. And

According to the present invention, since the mapping process is performed by using the edge information representing the edge of the oblique texture, the linear interpolation process such as the bilinear interpolation process is not required, and the circuit configuration required for the mapping process and the like is eliminated. And the algorithm can be simplified, and the mapping process can be performed at high speed.

[0023] The mapping processing step includes a step of acquiring a plurality of pieces of texture information corresponding to a plurality of texels, and a step of selecting one piece of texture information from the plurality of pieces of acquired texture information based on the edge information. ,
Mapping the selected texel to a pixel to be processed.

According to this, the mapping process can be performed at a higher speed by selecting one texel instead of a plurality of texels.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the drawings, taking an example in which the present invention is applied to a circuit for generating an image by performing filtering. Here, the filtering means a process of processing the texture data according to the application.

As an application of the texture data, for example, there is a texture mapping. The texture mapping means that the color and brightness of the two-dimensional texture data are adjusted and the texture data is associated with a three-dimensional polygon or the like (mapping).

Normally, since the size and the like of texture data and polygons do not correspond one-to-one, data interpolation processing is performed. Main interpolation processing includes, for example, bilinear interpolation processing, trilinear interpolation processing, and the like.

Here, the bilinear interpolation (also referred to as bilinear filtering) processing means that one pixel of data to be pasted is created by collating the original data every four pixels. That is, in the bilinear interpolation processing, one linear interpolation processing (linear interpolation processing) is performed on four values.

The trilinear interpolation (also referred to as trilinear filtering) process is a process in which original data is collated every four pixels to create one pixel of data to be pasted, on two surfaces. This means that an ideal one pixel is generated based on depth data (a value indicating how close to a predetermined plane is). That is, in the tri-linear interpolation processing, a maximum of seven linear interpolation processings are performed on eight values.

The texture data includes a plurality of pieces of texture information divided into texels.

FIG. 1 is a schematic diagram showing the relationship between pixels and texels.

When bilinear interpolation is performed, basically four texels Ta, Tb, Tc and Td are required for one pixel P as shown in FIG.

As shown in FIG. 1, each texel (actually five or more) is represented as being arranged on two-dimensional coordinates, and has a length in the Tx direction of 1 and a length in the Ty direction, respectively. 1 and are treated as being arranged adjacently in a square shape of size 1.

Therefore, the coordinates of the texel are represented by integer values. For example, the coordinates of the texel Ta are (m, n). Note that both m and n are integer values.

However, as shown in FIG. 1, since the pixel coordinates and the texel coordinates may not coincide with each other, the two-dimensional coordinate values tx0 and ty0 of the pixel P need to include decimal values.

Hereinafter, the integer value of tx0 is represented by tx_integer part,
The decimal value of x0 is tx_decimal part, and the integer value of ty0 is ty_
The description is continued with the integer part and the decimal value of ty0 as ty_decimal part.

Although it has been described that the number of texels required for one pixel is basically four, for example, tx_decimal part = 0, ty_
If the fractional part = 0, the center of the texel Ta and the center of the pixel P overlap, so that only one texel is required.
Similarly, when one of the tx_decimal part and the ty_decimal part is 0, the number of necessary texels is two. When both the tx_decimal part and the ty_decimal part are not 0, the number of necessary texels is four.

Generally, a texel having texture information is stored in a texture memory, and a bottleneck in texture mapping processing is a transfer of texture information from the texture memory.

When both the decimal parts of the pixel coordinates P are 0, the texture information for one texel may be transferred. However, this is rare, and in most cases, both the decimal parts of the pixel coordinates P are other than 0. It contains a value, and the transfer of texture information for 4 texels is required.

A first object of the present embodiment is to reduce a bottleneck at the time of transferring texture information from a texture memory and to perform a texture mapping process at high speed.

When performing trilinear interpolation processing,
A maximum of seven linear interpolation processes are performed using the texture information for eight texels. More specifically, in the tri-linear interpolation processing, two planes are selected from a plurality of (for example, eight) planes having different image definition levels, and bi-linear interpolation processing is performed on each of the planes. Based on the bilinear interpolation processing result, bilinear interpolation is further performed based on the result.

Here, the plane will be described.

FIG. 2 is a schematic diagram showing the relationship between the plane and the definition.

The surface is composed of eight surfaces, surface 0 to surface 7,
Each surface has a square shape. Assuming that the length of one side of the surface 0 is a, the length of the surface 1 is a / 2 and the length of the surface 2 is a / 4. As a result, the area is reduced by 1/4 times as the value of the surface increases to surfaces 1 and 2. That is, surface 0 has the highest definition and surface 7
Has the lowest definition.

In the trilinear interpolation processing, two continuous planes ideal for performing the interpolation processing are selected, and an ideal point on a line connecting the planes is selected based on the depth data.

Here, since the depth data is used not only for the geometry calculation but also for the Z buffer in the drawing processing after the geometry calculation, it is represented in a floating point format.

Conventionally, when a plane is selected by trilinear interpolation processing, a floating-point operation is performed using depth data. For this reason, the circuit configuration has become complicated and the circuit manufacturing cost has been high.

The pixel coordinate values tx0 and ty0 used in the bilinear interpolation processing are represented in a fixed-point format, and the provision of a circuit for floating-point arithmetic only for performing the trilinear interpolation processing requires circuit design efficiency and manufacturing. ineffective.

A second object of the present embodiment is to simplify a floating-point operation using depth data.

The above-mentioned texture information is represented by RGB
α is formed as a total of 32 bits of 8 bits each.
At present, RGB is a luminance value of Red, Green, Blue, and the like, and α is used as a value representing opacity (opacity).

However, depending on the display data, it is not always necessary to use the opacity. If the α8 bit can be used for another purpose, the storage capacity can be saved and the expressiveness of the texture can be improved. Expected.

A third object of the present embodiment is to effectively utilize α8 bits.

Further, when the texture is enlarged, the jaggy becomes conspicuous, and it is necessary to perform a bilinear interpolation process or the like in order to make the jaggy inconspicuous. Efficiency is improved.

A fourth object of the present embodiment is to perform filtering capable of making jaggy inconspicuous without performing linear interpolation.

Hereinafter, functions, processes, and the like for achieving the above-described four objects will be described.

First, functional blocks for achieving the above-mentioned four objects will be described.

(Explanation of Functional Block) FIG.
FIG. 3 is a functional block diagram of a 3D drawing unit 2 that performs filtering according to an example of the present embodiment.

As described in Iwanami Shoten's "Computer Game Technology", when performing 3D processing, the CPU, the 3D calculation section, and the 3D drawing section 2 share the processing. After the 3D calculation by the 3D calculation unit, the 3D drawing unit 2 performs 3D drawing.

The 3D drawing unit 2 includes a pixel unit data generation unit 10 for generating data in pixel units based on data input in units of polygons by the perspective transformation of the 3D operation unit, and texture information based on the pixel unit data. A texture coordinate conversion unit 20 for generating an address,
Texture information storage unit 40 in which texture information is stored
And a filtering unit 30 that performs bilinear interpolation and the like based on pixel unit data and texture information passed from the texture information storage unit 40.

The 3D drawing unit 2 includes an α conversion parameter generation unit 50 and a texture information generation unit 60. The texture information generation unit 60 generates texture information including α information, and outputs information indicating the content of the α information to the α conversion parameter generation unit 50.
The α conversion parameter generation unit 50 determines α based on the information.
An α conversion parameter indicating the content of the information is generated and output to the texture coordinate conversion unit 20 which is an interpolation processing unit. The α information and the α conversion parameter will be described later.

The texture information storage unit 40 includes a texture memory interface 42 that receives a transfer request for texture information and controls transfer of texture information and the like, and a texture memory 44 that is a storage area for texture information. ing.

The filtered data output from the filtering unit 30 is used for the subsequent 3D drawing.

(Description of Function and Operation of Texture Coordinate Conversion Unit) Next, the texture coordinate conversion unit 20 will be described in detail.

FIG. 4 is a functional block diagram of the texture coordinate conversion unit 20 according to an example of the present embodiment.

The texture coordinate conversion section 20 includes a mipmap value calculating section 210 and a mipmap interpolation area setting section 22
0, the conversion unit 230, and the bilinear interpolation area setting unit 24
0 is included.

The mipmap value calculating section 210 inputs the switching point bias value and the depth data from the pixel unit data generating section 10 in order to obtain the mipmap value used for determining the area to be interpolated. The mipmap value calculated based on the output is output to the mipmap interpolation area setting unit 220. The relationship between the switching point bias value and the depth data and the mipmap value will be described later.

The mipmap interpolation area setting section 220 inputs the mipmap interpolation area setting information from the pixel unit data generation section 10 and the mipmap value from the mipmap value calculation section 210, and based on these values, The area setting information is generated and output to the conversion unit 230.

The bilinear interpolation area setting section 240 inputs the information for setting the bilinear interpolation area (including the tx_decimal part and the ty_decimal part) from the pixel unit data generating section 10 and based on the information for setting the bilinear interpolation area. The setting information is output to the converter 230.

Here, the information for setting the bilinear interpolation area and the information for setting the trilinear interpolation area are information for determining the area to be interpolated.
This is information represented by 3). The information for setting the interpolation area is set based on, for example, the enlargement ratio of the texture, the depth data for the surface, the distance data from the viewpoint, and the like.
The setting of the interpolation area setting information is set to a specified value by the game designer.
It is appropriately controlled by a PU (game program). The purpose of setting the information for setting the interpolation area corresponds to, for example, the effect of suppressing jaggies and blurring, the effect of gradually focusing (or shifting) focus, and the like.

Next, the flow of the interpolation area setting process will be described.

(Explanation of Flow of Interpolation Area Setting Process) FIG. 5 is a flowchart for setting an interpolation area according to an example of the present embodiment. FIG. 6 is a diagram showing the relationship between the filtering coefficient and the fractional part of the mipmap value. FIG. 6A shows the case where the value of the area setting information is “11”.
FIG. 6B shows a case where the value of the area setting information is “10” and FIG.
FIG. 6C shows a case where the value of the area setting information is “01”.
(D) is a diagram showing a case where the value of the area setting information is “00”.

The setting of the interpolation area performed by the mipmap interpolation area setting section 220 and the bilinear interpolation area setting section 240 is performed in the following procedure. Here, the processing in the mipmap interpolation area setting unit 220 will be described as an example.

The mipmap interpolation area setting section 220 acquires 2-bit interpolation area setting information set for each polygon (step S2).

When the value of the interpolation area setting information is binary "1"
In the case of "1" (step S4), the processing to the decimal part is not performed because all interpolation is performed.

In this case, as shown in FIG. 6A, the fractional part of the mipmap value and the filtering coefficient have exactly the same value, and all regions are to be interpolated.

If the value of the interpolation area setting information is "00" in binary (step S6), it is assumed that no interpolation is performed, and all bits of the decimal part are set to 0.

In this case, as shown in FIG. 6D, the filtering coefficient is "0" regardless of the value of the fractional part of the mipmap value, and all the regions are not to be interpolated.

If the value of the interpolation area setting information is other than "11" or "00" in binary, the bits of the decimal part are shifted or all bits are set to 0 (step S10). Specifically, when the value of the interpolation area setting information is “10”, if the first and second bits below the decimal point are other than “11”, all bits below the decimal point are set to “0”, and the first and second bits below the decimal point are set to “1”. If the value is "11", all bits after the third bit after the decimal point are shifted left by 2 bits.

In this case, as shown in FIG. 6B, when the fractional part of the mipmap value is 0.75 or more, it becomes an interpolation target area.

The value of the interpolation area setting information is "01".
In the case of, if the first to fourth bits after the decimal point are other than “1111”, all bits below the decimal point are set to “0”, and if the first to fourth bits after the decimal point are “1111”, all bits after the fifth bit after the decimal point To the left by 4 bits.

In this case, as shown in FIG. 6 (C), the case where the fractional part of the mipmap value is 0.875 or more is the interpolation target area, and is more limited than when the value is "10". become.

As described above, by limiting the interpolation target area, the bottleneck at the time of transferring the texture information from the texture memory, which is the first object of the present embodiment, is reduced, and the texture mapping processing is performed at high speed. It is possible to do.

(Description of Function and Operation of Conversion Unit) Next, the conversion unit 230 will be described in detail.

The conversion section 230 includes area setting information from the mipmap interpolation area setting section 220, area setting information from the bilinear interpolation area setting section 240, tx_integer, and ty_
Based on the integer part, the texel address is output to the texture information storage unit 40, and the filtering coefficient is output to the filtering unit 30.

The conversion section 230 includes a texel address generation section 232 for generating a texel address, and a filtering coefficient output section 234 for outputting a filtering coefficient.

Next, the mipmap value calculator 210 will be described in detail.

FIG. 7 is a functional block diagram of the mipmap value calculator 210 according to an example of the present embodiment. FIG. 8 is a flowchart of a mipmap value calculation according to an example of the present embodiment. FIG. 9 is an explanatory diagram of an approximation operation according to an example of the present embodiment. FIG.
FIG. 9B shows the value obtained by truncating the fraction bits.
(C) is a diagram showing a value obtained by performing bit inversion.

In this embodiment, the value of the surface is -log
₂ (1 / depth data) + log ₂ K. Here, K is a proportional constant between the depth data and the surface definition.

That is, the depth data is data for selecting an interpolation target plane to be interpolated from a plurality of planes, and is data proportional to the definition of the plane.
Also, log ₂ K is called a switching point bias value. The depth data and the switching point bias value are one type of a surface selection parameter for selecting a surface.

The depth data is represented in a floating point format. However, in general, the operation in the floating-point format requires a more complicated arithmetic circuit and a longer operation time than the operation in the fixed-point format.

In this embodiment, the arithmetic circuit is simplified by processing the depth data in the floating-point format as data in the fixed-point format.

As shown in FIG. 7, the mipmap value calculation section 210 includes a data input section 211. Data input unit 2
11 inputs the switching point bias value and the depth data from the pixel unit data generation unit 10 (step S2).
2).

The mipmap value calculating section 210 includes a data format changing section 212. Data format change unit 212
Changes the value of -log ₂ (1 / depth data) in floating-point format input by the data input unit 211 to a value in fixed-point format (step S24).

The value output from the data format change unit 212 has, for example, the data format shown in FIG.

The mipmap value calculating section 210 includes a bit limiting section 213. The bit limiting unit 213 discards unnecessary bits in order to limit the number of bits after the decimal point of the fixed-point format data from the data format changing unit 212 to a predetermined number (step S26). For example, when the number of bits after the decimal point of the switching point bias value is 10 digits, the predetermined number is preferably 10.

The value shown in FIG.
9 from the bit limiter 213 to FIG.
The value shown in (B) is output.

The mipmap value calculating section 210 includes a bit inverting section 214. The bit inverting unit 214 receives the value from the bit limiting unit 213, and inverts all bits of the value (step S28).

The value shown in FIG.
, The value shown in FIG. 9C in which all the bits are inverted is output from the bit inversion unit 214. By inverting all bits, the negative value -log ₂
(1 / depth data).

Further, the mipmap value calculation section 210
An operation unit 215 is included. The operation unit 215 is a bit inversion unit 21
The mipmap value is calculated based on the value obtained by inverting the bit in step 4 (step S30).

The mipmap value obtained by the arithmetic unit 215 is output to the mipmap interpolation area setting unit 220 (step S32).

Next, the relationship between the switching point bias value, the converted depth data, and the mipmap value will be described.

FIG. 10 is an explanatory diagram showing the relationship between the switching point bias value, the converted depth data, and the mipmap value.

The operation unit 215 calculates the sum of the value of the converted depth data whose bit has been inverted by the bit inversion unit 214 and the switching point bias value, thereby obtaining the sign bit, the overflow determination value, and the mipmap value. (Step S30).

Here, the sign bit is a bit (1 bit) indicating the validity of the value of the surface described later by positive or negative. If the value of the sign bit is 1 (negative), the value of the surface is treated as 0, and if the value of the sign bit is 0 (positive), the value of the surface described later is applied as it is.

The overflow determination value is a value (5 bits) indicating whether or not a surface value described later can be represented by 0 to 7, and is an upper bit of the surface value. If the overflow determination value is not 0, the surface value has overflowed (8 or more). In this case, the surface value is treated as 7.

The mipmap value is a surface value and a trilinear interpolation coefficient (filtering coefficient). The mipmap value is composed of an integer part of 3 bits and a decimal part of 10 bits. The three bits of the integer part become the surface value, and the ten bits of the decimal part become the trilinear interpolation coefficient.

For example, in the example shown in FIG. 10, the value of both the sign bit and the overflow determination value is 0, and the value of the surface is "010" in binary, so that the surface 2 is selected.

Conventionally, the value of a surface is obtained by performing a floating-point operation. However, as in the present embodiment, a value of a surface is determined by approximating a value of a floating-point format with a value of a fixed-point format. Accordingly, the second object, that is, the floating-point arithmetic using depth data can be simplified. It should be noted that, even when approximating the surface value or the like, the value can be obtained without lowering the precision as compared with the case of obtaining the surface value in the floating-point format.

Further, by approximating a value in the floating-point format with a value in the fixed-point format to obtain a surface value or the like, the circuit configuration and algorithm can be simplified, and a processing circuit and the like can be realized at low cost.

(Description of Function and Operation of Filtering Unit) Next, the filtering unit 30 will be described in detail.

FIG. 11 is a functional block diagram of the filtering unit 30 according to an example of the present embodiment.

The filtering unit 30 includes an α conversion unit 310 for inputting texture information and α conversion parameters, an output from the α conversion unit 310, and a filtering coefficient output unit 23.
4 and the filtering information and texture information (R
Contains GB information. ), And a trilinear interpolation unit 340 that performs a trilinear interpolation process by a plurality of bilinear interpolation processes using the bilinear interpolation unit 330.

The filtering section 30 includes a control section 320 for generating various types of control information. Part of this control information is input to the trilinear interpolation section 340 and used for trilinear interpolation processing.

The filtering unit 30 includes a contour texture unit 350 described later. Further, the filtering unit 30 includes a contour texture unit 350,
Bilinear interpolation unit 330 or trilinear interpolation unit 34
And a selector 360 for selecting which output to use based on the output from 0.

The selection result of the selection section 360 is used as filtered data in 3D drawing after filtering.

Next, the α conversion unit 310 will be described in detail.

FIG. 12 is a functional block diagram of the α conversion unit 310 according to an example of the present embodiment.

The α conversion section 310 includes an image generation information extraction section 312 for extracting image generation information from texture information from the texture information storage section 40 based on α conversion parameters from the α conversion parameter generation section 50. As the image generation information, for example, various flags such as the opacity, the drawing mode, and whether or not the fade effect is applied correspond.

That is, as described above, when generating the texture information including the α data (α information), the texture information generation unit 60 does not set only the opacity as the α data as in the related art. In order to effectively utilize the α8 bit, which is the third object, the drawing mode and the flag are set in the α data as needed. And α
The conversion parameter generation unit 50 includes the texture information generation unit 6
Based on the information from 0, an α conversion parameter indicating what information is included in the α data is generated, and the α conversion parameter is output to the filtering unit 30 as necessary.

Next, the relationship between α data and α conversion parameters will be described.

FIG. 13 is a schematic diagram showing the relationship between α data and α conversion parameters. FIG. 14 is a schematic diagram illustrating a relationship between the opacity determination value and the opacity output. FIG. 15 is a schematic diagram showing the relationship between the drawing mode determination value and the drawing mode output. Also, FIG.
Judge values of G_A (flag A) to FLG_E and FLG_A
It is a schematic diagram which shows the relationship with -FLG_E output.

The α data included in the texture information is α7
It is composed of 8 bits of? 0. On the other hand, the α conversion parameter is composed of 10 bits β9 to β0. Note that α
7, β9 are the most significant bits (MSB). Of these, three bits β9 to β7 are used as opacity bit information, β6 is used as a bit for determining the presence or absence of opacity, β5 is used as a bit for determining the presence or absence of a drawing mode, and β4 to β7. β0 is 5
It is used as a bit for determining the presence / absence of two flags (FLG_A to FLG_E).

Here, the opacity bit information indicates the number of effective bits of the opacity data in the α data −1, and the number of effective bits of the opacity data is eight at the maximum. Therefore, the opacity bit information is set to 0 as shown in FIG.
~ 7.

In this embodiment, a contour mode described later is used as the drawing mode. Also,
The flag can be used to represent, for example, a fade effect invalid, a specular invalid, an environment mapping invalid, and the like.

The α conversion section 310 includes an α information conversion section 314 to which the value of the image generation information extraction section 312 is input.

[0126] The α information conversion unit 314 is configured to generate 1
It outputs 0-bit opacity data, 1-bit drawing mode valid / invalid judgment value, 4-bit drawing mode, and 5-bit flag.

Here, assuming a more specific case, α
The mutual relationship between data, α conversion parameters, and the like will be described.

(Case 1: When all of α7 to α0 as α data are used as opacity) In this case, FIG.
Are all 1 (ON), and the value of β6, which is the opacity validity / invalidity determination value, is also 1 (opacity validity). The α information conversion unit 314 uses all of the α data as opacity, and as shown in FIG.
Opacity data expanded from α7 to α0 to become 10 bits (for example, when the opacity determination value is 7, α7α6
α5α4α3α2α1α0α7α6) are output. The α information conversion unit 314 outputs a value in which the validity / invalidity determination value of the drawing mode and the value of the drawing mode itself are all set to 0 (invalid), and the flags are all set to 0.

(Case 2: Case of Using Drawing Mode as Part of α Data) In this case, β5 indicating the validity / invalidity determination value of the drawing mode is 1 (valid).

In case 2, if β6 indicating the opacity validity / invalidity determination value is 0 (invalid), four bits α7 to α4 are used as the drawing mode as shown in FIG.

In case 2, if β6 indicating the validity / invalidity judgment value of opacity is 1 (valid), the α information conversion unit 3
14 determines bits to be used as opacity among α8 bits based on the values of β9 to β7 which are opacity bit information.

For example, as shown in FIG.
Is 0 (“the number of bits of opacity−1” in FIG. 15), the opacity is 1 bit (α).
7) is used, and 4 bits α6 to α3 are used as the drawing mode.

If the value obtained by deciphering β9 to β7 into 3 is 3, 4 bits (α7 to α4) are used as the opacity, and 4 bits α3 to α0 are used as the drawing mode.

Further, the value obtained by evolving β9 to β7 into 10 is 4
In the above case, α3 and later are also used as the opacity, and since 4 bits out of α8 bits are not used as the drawing mode, the α information conversion unit 314 sets all values of the drawing mode 4 bits to 0, Output.

(Case 3: When using a flag as a part of α data) When any one of β4 to β0, which is a bit for determining the presence or absence of a flag (FLG_A to FLG_E), is 1 (present), Apply the image effect indicated by the flag.

As shown in FIG. 16, for example, FLG_A
If β4 = 1 indicating the presence or absence of
It is determined that FLG_A is used, and the most significant bit (α_α) of α7 to α0 not used in the opacity and the drawing mode is determined.
A) is applied as FLG_A.

For example, if β3 = 1, the α information conversion unit 314 determines that FLG_B is used, and sets α7 to α7 that are not used in opacity, drawing mode, and FLG_A.
The most significant bit (α_B) of 0 is applied as FLG_B.

Further, for example, if β0 = 1, the α information conversion unit 314 determines that FLG_E is used, and
Opacity, drawing mode, FLG_A, FLG_B, F
The most significant bit (α_E) among the unused α0 to α7 in LG_C and FLG_D is applied as FLG_E.

As described above, according to the present embodiment, the contents of the α data can be grasped by using the α conversion parameters indicating the contents of the α data, including the image generation information other than the opacity in the α data. . Since image generation information such as a drawing mode can be extracted from the α data and used based on the α conversion parameter, there is no need to provide a dedicated storage area for the image generation information such as the drawing mode. Accordingly, the third object of the present invention, ie, the effective use of α8 bits can be achieved.

In the above, an example in which linear interpolation (bilinear interpolation, trilinear interpolation) is performed has been described. However, in order to further improve processing efficiency, it is necessary to perform filtering without performing linear interpolation.

In order to achieve the fourth object, in this embodiment, a texture image having oblique information is used.

(Description of Texture Image Having Oblique Information) FIG. 17 is a diagram showing an example of conventional texture mapping. FIG. 17A shows an original texture image, and FIG. FIG. 9 is a diagram showing an image after mapping.

The texture image shown in FIG. 17A is composed of a total of 16 texels, that is, 4 pixels vertically and 4 pixels horizontally. Each texel is represented by a square of one color.

For this reason, when the texture image is enlarged to a size of 8 × 8 as shown in FIG. 17B, the color change is conspicuous, and the viewer sees the image as a so-called jaggy.

As a method for solving such a problem, linear interpolation processing such as the above-described bilinear interpolation is performed.
In order to further increase the processing efficiency, a method other than the linear interpolation processing is required.

In the present embodiment, a method of making jaggy less noticeable by associating oblique information with a texture image is adopted.

FIG. 18 is a diagram showing an example of texture mapping according to an example of the present embodiment. FIG. 18A shows an original texture image, and FIG. 18B shows an image after texture mapping. FIG. FIG. 19 is an enlarged view of a part of the original texture image shown in FIG. 18A, and FIG.
9 (B) is an image of a 63-degree contour line, and FIG.
It is a figure which shows the image of the outline of a degree.

As can be seen from a comparison with FIG.
A texel corresponding to an edge portion in the four-by-four-by-four texture image shown in FIG. 18A has a portion represented by two colors. Hereinafter, the edge part of the texture image will be described as a contour (contour).

For example, the texture image is defined by coordinates (vertical position,
When expressed by a horizontal position, the texel at coordinates (2, 3) shown in FIG. 19A is represented by a 45-degree contour line.

The coordinates (1, 1) and the coordinates (1, 1)
The texel of 2) is represented by a contour line of 63 degrees with respect to the vertical direction as shown in FIG.

Further, the coordinates (3, 4) and (4,
The texel of 4) is represented by a contour line of 27 degrees with respect to the vertical direction as shown in FIG.

In the contour texture processing, four texels Ta to Td near the desired point P shown in FIG. 1 are used without performing linear interpolation.

For example, the above-described 4-bit drawing mode is used as a contour mode. Also, a contour flag 1 bit is provided for each polygon. When the contour flag is ON, the contour texture processing is performed. When the contour flag is OFF, the contour texture processing is not performed. As the contour flag, a flag β5 indicating the presence or absence of the above-described drawing mode is used.

As described above, the drawing mode is represented by 4 bits. Therefore, the drawing mode is from “0000” to
Values up to “1111” can be taken.

FIG. 20 is a diagram showing the contour mode values corresponding to each texel shown in FIG.

First, 45 degrees, 63 degrees, and 2 degrees within one texel.
A texel requiring oblique information such as 7 degrees will be described. For example, the value of the contour mode corresponding to the texel at the coordinates (2, 3) shown in FIG.
101 ", which is" 5 "in hexadecimal. FIG.
The values of the contour mode corresponding to the texel of coordinates (1, 1) (1, 2) shown in FIG. 9 (B) are “1001” and “1010” in binary, “9” in hexadecimal, It becomes "A". Also, the values of the contour mode corresponding to the texel of coordinates (3, 4) (4, 4) shown in FIG. 19C are “1110” and “1” in binary, respectively.
101 ", which are" E "and" D "in hexadecimal.

For a texel that does not require oblique information in one texel, the value of the drawing mode corresponding to the texel is “0000” in binary and “0” in hexadecimal.

The value of the contour mode is "000".
In the case of "0" to "0011" (0 to 3), the texel to be used is selected using only the contour mode. When the value of the drawing mode is "0100" to "1111" (4 to F), the contour is selected. Mode and tx decimal, ty
Select the texel to use using the decimal part.

For example, if the value of the contour mode is “000”
The texel Ta is used for “0”, the texel Tc is used for “0001”, the texel Tb is used for “0010”, and the texel Td is used for “0011”.

In the case of representing a 45-degree outline, the values of the drawing mode are “0100” to “0111” (4
To 7) are used. Further, when representing other contour lines, the values of the drawing mode are set to “1000” to “111”.
1 "(8 to F).

Hereinafter, the case shown in FIGS. 18 and 19 will be described in detail with reference to an example.

For example, in the case of the texel corresponding to the coordinates (1, 1), the value of the contour mode is “1001”. In this case, if tx decimal part / 2 exceeds ty decimal part, texel Tb is selected, and if tx decimal part / 2 is less than ty decimal part, texel Tc is selected.

For example, in the case of the texel corresponding to the coordinates (1, 2), the value of the contour mode is “1010”.
It is. In this case, tx decimal part / 2 is ty decimal part-1 /
If it exceeds 2, texel Tb is selected and tx decimal part /
If 2 is equal to or smaller than the ty decimal part -1/2, the texel Tc is selected.

For example, in the case of the texel corresponding to the coordinates (3, 4), the value of the contour mode is “1110”.
It is. In this case, if the tx decimal part is larger than the ty decimal part / 2, the texel Tb is selected, and if the tx decimal part is less than the ty decimal part / 2, the texel Tc is selected.

For example, in the case of the texel corresponding to the coordinates (4, 4), the value of the contour mode is “1101”.
It is. In this case, tx decimal part is ty decimal part / 2 + 1 /
If it is larger than 2, the texel Tb is selected, and if the tx decimal part is less than or equal to the ty decimal part / 2 + 1/2, the texel Tc is selected.

As described above, the texture mapping can be performed without performing the linear interpolation process by using the contour mode, the tx decimal part, and the ty decimal part by including the oblique information in the texture image.

Therefore, jaggies can be reduced without performing the linear interpolation processing, and the filtering processing efficiency can be improved. That is, the fourth object can be achieved by using the contour texture.

The functional blocks and the like to which the present invention is applied have been described. The above-described functional blocks and the like are embodied in, for example, a circuit, a program for causing a computer to realize the above-described units, a computer-usable information storage medium 4 storing the program, and a carrier wave for realizing the above-described functions. It can be realized using information to be converted. In addition, as the information storage medium 4, for example, a CD-ROM, a DVD-ROM, a RAM, an HDD, or the like can be applied.

Hereinafter, a case where the above-described functions are realized using a circuit for realizing each function will be described.

(Description of Circuit Block) FIG.
FIG. 3 is a block diagram of a coordinate conversion circuit according to an example of the present embodiment.

The coordinate conversion circuit is a circuit for realizing the function of the texture coordinate conversion unit 20 shown in FIG. Note that the pixel unit data generation unit 10 can be similarly implemented using a circuit. Pixel unit data generator 1
A circuit that realizes 0 is a pixel unit data generation circuit.

A part of the data input to the pixel unit data generation circuit in units of polygons is converted into pixel unit data by the pixel unit data generation circuit.

As shown in FIG. 21, as data in units of polygons, for example, depth data, tx_integer part, tx
_Decimal part, ty_integer part, ty_decimal part, mipmap decimal part.

The data in pixel units include, for example, a switching point bias value, information for setting a mipmap interpolation area, and information for setting a bilinear interpolation area.

The coordinate conversion circuit includes a mipmap value calculation circuit 1210 for realizing the function of the mipmap value calculation section 210 shown in FIG. 7 and a mipmap interpolation area processing circuit 1222 for realizing the function of the mipmap interpolation area setting section 220. And a mipmap determination circuit 1224, a conversion circuit 1230 for realizing the function of the conversion unit 230, a texture interpolation area processing circuit 1242, 1244, and a trilinear determination circuit 1246, 1248 for realizing the function of the bilinear interpolation area setting unit 240. It is comprised including.

The mipmap value calculation circuit 1210 receives the switching point bias value (A) and the depth data (B) from the pixel unit data generation circuit. Further, the mipmap value calculation circuit 1210 simplifies the above-described calculation (does not perform floating-point calculation) based on the switching point bias value and the depth data, and performs a mipmap determination circuit 1224.
, And outputs a mipmap decimal part (MM_DEC) to the mipmap interpolation area processing circuit 1222.

Mipmap interpolation area processing circuit 1222
Inputs the mipmap decimal part (D_IN) from the mipmap value calculation circuit 1210 and the mipmap interpolation area setting information (S) from the pixel unit data generation circuit. Also, a mipmap interpolation area processing circuit 1222
A mipmap decimal part (D_OUT) obtained by processing the mipmap decimal part as necessary based on the mipmap interpolation area setting information.
Output to

The mipmap determination circuit 1224 inputs the mipmap value integer part (INT_IN) from the mipmap value calculation circuit 1210 and the processed mipmap decimal part (DE) from the mipmap interpolation area processing circuit 1222.
C_IN).

Then, the mip map determination circuit 1224
Outputs the mipmap value integer part (INT), a value obtained by adding 1 to the mipmap value integer part (INT + 1), and the mipmap decimal part (DEC) to the conversion circuit 1230.

The texture interpolation area processing circuit 124
2 inputs tx_decimal part (D_IN) and bilinear interpolation area setting information (S) from the pixel unit data generation circuit. The texture interpolation area processing circuit 1242
Corrects the tx_decimal part based on the bilinear interpolation area setting information represented by 2 bits, as described in the description of the bilinear interpolation area setting section 240 described above.
Then, the texture interpolation area processing circuit 1242 outputs the tx_decimal part (D_OUT) corrected as necessary to the trilinear determination circuit 1246.

The trilinear decision circuit 1246 outputs the tx_integer (INT_I) from the pixel unit data generation circuit.
N) and the corrected tx_decimal part (DEC_IN) from the texture interpolation area processing circuit 1242.

Then, the trilinear decision circuit 1246
Outputs to the conversion circuit 1230 a tx_integer part (INT), a value obtained by adding 1 to the tx_integer part (INT + 1), and a tx_decimal part (DEC).

The texture interpolation area processing circuit 124
Reference numeral 4 inputs ty_decimal part (D_IN) and bilinear interpolation area setting information (S) from the pixel unit data generation circuit. Also, the texture interpolation area processing circuit 1244
Corrects the ty_decimal part based on the bilinear interpolation area setting information represented by 2 bits, as described in the description of the bilinear interpolation area setting unit 240 described above.
Then, the texture interpolation area processing circuit 1244 outputs the ty_decimal part (D_OUT) corrected as necessary to the trilinear determination circuit 1248.

The tri-linear decision circuit 1248 outputs the ty_integer (INT_I) from the pixel unit data generation circuit.
N), and the corrected ty_decimal part (DEC_IN) from the texture interpolation area processing circuit 1244.

Then, the trilinear decision circuit 1248
Outputs a ty_integer part (INT), a value obtained by adding 1 to the ty_integer part (INT + 1), and a ty_decimal part (DEC) to the conversion circuit 1230.

The conversion circuit 1230 outputs the mipmap integer part (MM_INT) from the mipmap determination circuit 1224,
The mipmap integer part + 1 (MM_INT + 1) and the mipmap decimal part (MM_DEC) are input.

The conversion circuit 1230 receives the tx_integer part (TX_INT), the tx_integer part + 1 (TX_INT + 1), and the tx_decimal part (TX_DEC_IN) from the bilinear decision circuit 1246.

Further, the conversion circuit 1230 outputs the ty_integer (TY_INT) from the bilinear judgment circuit 1248,
The ty_integer part + 1 (TY_INT + 1) and ty_decimal part (TY_DEC_IN) are input.

Further, the conversion circuit 1230 inputs the contour information (CONTOUR_ON) indicating whether to use the contour from the pixel unit data generation circuit.

The conversion circuit 1230 calculates tx_decimal part, ty
_ Based on the value of the decimal part, 1 to 4 texel addresses (TEXTEL_ADD) are stored in the texture information storage unit 40.
(Texture memory 44).

More specifically, conversion circuit 1230 operates as tx_
When the values of the decimal part and ty_decimal part are both 0, one texel address (27 bits) is output to the texture information storage unit 40. Note that this texel address is 2
When converted into dimensional coordinates, coordinates (tx_integer part, ty_integer part) are obtained.

When the value of tx_decimal part is 0 and the value of ty_decimal part is not 0, conversion circuit 1230 outputs two texel addresses to texture information storage unit 40. When the texel address is converted into two-dimensional coordinates, the coordinates (tx_integer part, ty_integer part) and the coordinates (tx_tx part)
_Integer part, ty_integer part + 1).

If the value of tx_decimal part is not 0 and the value of ty_decimal part is 0, conversion circuit 1230 outputs two texel addresses to texture information storage unit 40. When the texel address is converted into two-dimensional coordinates, the coordinates (tx_integer part, ty_integer part) and the coordinates (tx_tx part)
_Integer part + 1, ty_integer part).

If both the values of tx_decimal part and ty_decimal part are not 0, conversion circuit 1230 outputs four texel addresses to texture information storage unit 40.
When this texel address is converted into two-dimensional coordinates, coordinates (tx_integer part, ty_integer part), coordinates (tx_integer part, ty_integer part + 1), and coordinates (tx_integer part + 1, t
y_integer) and coordinates (tx_integer + 1, ty_integer + 1).

As described above, the tx_decimal part and the ty_decimal part are corrected, and the probability of becoming 0 becomes higher than before, so that the number of texel addresses to be used is reduced. The transfer amount of texture information can be reduced.

Further, conversion circuit 1230 includes a mipmap_decimal part (MM_DEC) and a tx_decimal part (TX_D).
EC) and ty_decimal part (TY_DEC) are output to the filtering circuit.

Next, details of the filtering circuit will be described.

FIG. 22 is a block diagram of a filtering circuit according to an example of the present embodiment.

The filtering circuit includes an α conversion circuit 1310 for realizing the function of the α conversion unit 310, an overall control circuit 1320 for realizing the function of the control unit 320, and a trilinear interpolator for realizing the function of the trilinear interpolation unit 340. 13
40 to 1346.

The α conversion circuit 1310 receives the α conversion parameter (PARA) transmitted from the pixel unit data generation circuit.
_IN) and texture information (D_IN (31) from the texture information storage unit 40 (texture memory 44).
-0)), the α data (D_IN (31-24), α_IN) of the 24th to 31st bits are input.

The α-conversion circuit 1310 extracts various flags (FLAG),
Whether to apply contour mode (CONTOUR_ON),
It outputs the contour mode (CONTOUR_MD) and the opacity (OPACITY).

Further, the overall control circuit 1320 receives the flag (FLAG_IN), the application of the contour mode (CONTOUR_ON), and the contour mode (CONTOUR_MD) from the α conversion circuit 1310.

Further, the overall control circuit 1320 outputs a flag (FLAG_OUT) to a subsequent circuit. This flag is used for α blending and the like.

Further, the overall control circuit 1320 transmits control information (CTRL) for controlling the trilinear interpolators 1340 to 1346 to the respective trilinear interpolators 1340 to 134.
6 is output.

Each trilinear interpolator 1340 to 1346
Receives the control information (CTRL) from the overall control circuit 1320 and the tx_decimal part (tx_dec), ty_decimal part (ty_dec), and mipmap decimal part (mm_dec) from the conversion circuit 1230 shown in FIG.

The tri-linear interpolator 1340 calculates α
The opacity (D_IN) is input from the conversion circuit 1310. Then, the trilinear interpolator 1340 outputs the opacity to the subsequent circuit based on the input opacity, tx_decimal part, ty_decimal part, mipmap decimal part, and control information.

The trilinear interpolator 1342 outputs the 16th to 23rd bits (D_IN (23-1) of the texture information extracted from the texture memory 44.
6)) is input. And the trilinear interpolator 134
2 is input texture information, tx_decimal part, ty_
R based on decimal part, mipmap decimal part and control information
ED (red value of RGB values) is output to a subsequent circuit.

The tri-linear interpolator 1344 calculates the eighth to fifteenth bits (D_IN (15-8)) of the texture information extracted from the texture memory 44.
Enter Then, the trilinear interpolator 1344 calculates
GREEN based on input texture information, tx_decimal part, ty_decimal part, mipmap decimal part, control information
(The green value of the RGB values) is output to the subsequent circuit.

[0209] The trilinear interpolator 1346 inputs the 0th to 7th bits (D_IN (7-0)) of the texture information extracted from the texture memory 44. Then, the trilinear interpolator 1346 generates a BLUE (RGB) based on the input texture information, tx_decimal part, ty_decimal part, mipmap decimal part, and control information.
The blue value of the value) is output to the subsequent circuit.

Next, the trilinear interpolators 1340 to 1330
46 will be described in detail.

FIG. 23 is a block diagram of the trilinear interpolators 1340 to 1346 according to an example of the present embodiment.

Each trilinear interpolator 1340 to 1346
Are selectors 354 and 362 and linear interpolators 332 to 3
42.

In FIG. 23, seven registers 3
80 to 386 temporarily hold the values, and 9 registers 3
90 to 398 are enable signals XA_EN, XB_E
N, XC_EN, XD_EN, XTA_EN, XTB_
The value is held until EN and XEN are turned ON.

First, the operation of the circuit at the time of linear interpolation processing by ordinary bilinear interpolation and trilinear interpolation will be described.

The linear interpolator 332 has two texels T
a, Tb texture information and tx decimal part (tx_d
Enter ec). The linear interpolator 332 selects appropriate texture information from among a plurality of pieces of texture information based on the value of the tx decimal part, and outputs it as an output Y.

The linear interpolator 334 has two texels T
c, texture information of Td and tx decimal part (tx_d
Enter ec). The linear interpolator 334 selects appropriate texture information from a plurality of pieces of texture information based on the value of the tx decimal part, and outputs the selected texture information as an output Y.

The linear interpolator 336 has an output Y (A) from the linear interpolator 332 and an output Y (A) from the linear interpolator 334.
(B), ty decimal part (K) is input. Linear interpolator 33
6 selects appropriate texture information from the texture information A and B based on the value of the ty decimal part and outputs it as an output Y.

By the processing of the linear interpolators 332 to 336, bilinear processing of the first plane is performed.

Further, by performing the processing of the linear interpolators 332 to 336 again, bilinear processing of the second surface is performed.

The result of the bilinear processing on the first surface is temporarily stored in the register 397, and the result of the bilinear processing on the second surface is temporarily stored in the register 398.

The linear interpolator 342 inputs the first surface bilinear processing result (A), the second surface bilinear processing result (B), and the mipmap decimal part (K) from the conversion circuit 1230. The linear interpolator 342 performs a trilinear interpolation process based on the information.

Next, the operation of the contour texture circuit 352 for performing the above-described contour texture processing will be described.

The contour texture circuit 352 includes a selector 354, a selector 362, and registers 380 to 380.
383 are included.

The selector 354 has four texels T
a, Tb, Tc, Td texture information A, B, C,
D, Contour mode (CONTO as drawing mode)
UR_MD), tx decimal part (tx_dec) and ty
Enter the decimal part (ty_dec).

The selector 354 functions as a 4to1 selector. That is, as described above, the contour mode values 0 to F (hexadecimal notation), tx decimal part, t
One piece of texture information is selected from the four pieces of texture information A to D based on the value of the y decimal part, and output as an output Y.

The selector 362 receives the texture information (A) selected by the selector 354, the texture information (B) subjected to the trilinear interpolation processing by the linear interpolator 342, and the application of contour (S). The selector 362 selects the texture information (A) selected by the selector 354 if the contour application / non-application (CONTOUR_ON) is ON, and performs trilinear interpolation processing by the linear interpolator 342 if the contour application / non-application is OFF. Selected texture information (B).

The selector 362 outputs the result of this selection to the subsequent circuit as filtered data. The subsequent circuit generates a final display image using the filtered data.

As described above, the present invention can be realized using a circuit.

(Modifications) The preferred embodiments to which the present invention is applied have been described above. However, the application of the present invention is not limited to the above-described embodiments, and various modifications are possible.

For example, although the above-described drawing mode is used as the contour mode, the drawing mode is not limited to the contour mode, and various drawing modes can be applied. As described above, when a plurality of types of drawing modes are applied, a drawing mode to be applied for each texture information may be provided.

In this case, as a method of selecting a drawing mode, for example, tx decimal part, t
The drawing mode may be selected using the y decimal part. More specifically, for example, as shown in FIG.
If both y decimal parts are 0.5 or less, the drawing mode Ma associated with Ta is selected, and the tx decimal part exceeds 0.5,
If the ty decimal part is 0.5 or less, the drawing mode Mb associated with Tb is selected, and the tx decimal part is 0.5 or less,
When the ty decimal part exceeds 0.5, the drawing mode Mc associated with Tc is selected. When both the tx decimal part and the ty decimal part exceed 0.5, the drawing mode Md associated with Td is selected. do it.

As described above, by selecting the drawing mode based on the filtering coefficient, the optimum drawing mode can be selected.

Similarly, the above-mentioned flag is also
ON of the appropriate flag based on the flag for 4 texels,
OFF may be determined.

As a method of determining this flag, for example,
The majority method can be adopted. As shown in FIG. 1, flags corresponding to texels Ta, Tb, Tc, and Td are respectively set to F
a, Fb, Fc, Fd, Fa, Fb, Fc, F
If the sum of d exceeds 2, the flag is regarded as ON,
When the sum of Fa, Fb, Fc and Fd does not exceed 2, the flag is regarded as OFF. When the sum of Fa, Fb, Fc and Fd is 2, ON or OF is determined depending on the function of the flag.
F is determined.

Further, at the time of this determination, the tx decimal part is 0.
In the case of, Fb and Fd are each regarded as 0 (OFF), and when the ty decimal part is 0, Fa and Fc are each 0.
(OFF). If tx decimal part is 0, T
a or Tc is selected, Tb and Td and the flags Fb and Fd corresponding to the texel have no meaning. When the ty decimal part is 0, Tb or Td is always selected, and Ta and Tc are selected.
This is because the flags Fa and Fc corresponding to the texel have no meaning.

In the contour mode described above, diagonal information is represented by using two texels, but one texel represents diagonal information (always 45 degrees).
The present invention can also be applied to an oblique display using one or more texels.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing the relationship between pixels and texels.

FIG. 2 is a schematic diagram showing a relationship between a surface and a definition.

FIG. 3 is a functional block diagram of a 3D drawing unit that performs filtering according to an example of the embodiment.

FIG. 4 is a functional block diagram of a texture coordinate conversion unit according to an example of the present embodiment.

FIG. 5 is a flowchart of setting an interpolation area according to an example of the embodiment.

6A and 6B are diagrams showing a relationship between a filtering coefficient and a mipmap value decimal part. FIG. 6A shows a case where the value of the area setting information is “11”, and FIG. 6B shows a value of the area setting information. Is “10”, FIG. 6C shows that the value of the area setting information is “0”.
In the case of “1”, FIG. 6D shows that the value of the area setting information is “00”.
FIG.

FIG. 7 is a functional block diagram of a mipmap value calculation unit according to an example of the embodiment.

FIG. 8 is a flowchart of a mipmap value calculation according to an example of the present embodiment.

9A and 9B are explanatory diagrams of an approximation operation according to an example of the present embodiment. FIG. 9A is an original value, FIG. 9B is a value obtained by truncating a fractional bit, and FIG. It is a figure showing the value which performed bit inversion.

FIG. 10 is an explanatory diagram showing a relationship between a switching point bias value, converted depth data, and a mipmap value.

FIG. 11 is a functional block diagram of a filtering unit according to an example of the embodiment.

FIG. 12 is a functional block diagram of an α conversion unit according to an example of the embodiment.

FIG. 13 is a schematic diagram showing a relationship between α data and α conversion parameters.

FIG. 14 is a schematic diagram illustrating a relationship between an opacity determination value and an opacity output.

FIG. 15 is a schematic diagram illustrating a relationship between a drawing mode determination value and a drawing mode output.

FIG. 16 shows determination values of FLG_A to FLG_E and FLG_
It is a schematic diagram which shows the relationship with the output of A-FLG_E.

FIG. 17 is a diagram showing an example of conventional texture mapping. FIG. 17 (A) shows an original texture image, and FIG.
FIG. 7B is a diagram showing an image after texture mapping.

18A and 18B are diagrams illustrating an example of texture mapping according to an example of the present embodiment. FIG. 18A illustrates an original texture image, and FIG. 18B illustrates an image after texture mapping. .

19A is an enlarged view of a part of the original texture image shown in FIG. 18A, FIG. 19A is an image of a 45-degree contour, and FIG. 19B is a 63-degree contour. Image, FIG. 19
(C) is a figure which shows the image of the contour line of 27 degrees.

20 is a diagram showing a contour mode value corresponding to each texel shown in FIG.

FIG. 21 is a block diagram of a coordinate conversion circuit according to an example of the present embodiment.

FIG. 22 is a block diagram of a filtering circuit according to an example of the present embodiment.

FIG. 23 is a block diagram of a trilinear interpolator according to an example of the present embodiment.

[Explanation of symbols]

 Reference Signs List 4 information storage medium 10 pixel unit data generation unit 20 texture coordinate conversion unit 30 filtering unit (interpolation processing means) 40 texture information storage unit 44 texture memory 210 mipmap value calculation unit 212 data format change unit 213 bit limit unit 214 bit inversion unit 215 arithmetic section 220 mipmap interpolation area setting section (correction means) 232 texel address generation section (selection means) 240 bilinear interpolation area setting section (correction means) 310 α conversion section 312 image generation information extraction section 314 α information conversion section 320 Control unit 330 Bilinear interpolation unit 340 Trilinear interpolation unit 350 Contour texture unit 360 Selection unit

Claims (12)

[Claims] 1. An image generation system for generating an image by performing an image mapping process using texture information including color information and additional information stored in a predetermined storage area for each texel, wherein the texture information Selecting means for selecting at least one texel from a plurality of texels based on predetermined texel coordinate designation information and the edge information, the edge means including edge information representing an edge of a texture oblique to a direction of a coordinate axis of the texel. Means for performing a mapping process using the texture information of the selected texel. 2. The image generation system according to claim 1, further comprising means for storing the edge information in the storage area as at least a part of the additional information. 3. The texel coordinate designation information according to claim 1, wherein the texel coordinate designation information includes a two-dimensional coordinate value, and the selection unit determines a texel coordinate based on a fractional part of two numerical values of the two-dimensional coordinate value. An image generation system, wherein the texel is selected. 4. The image generation system according to claim 3, wherein the selection unit selects one texel from a plurality of texels near the texel based on the two numerical values. 5. The point according to claim 3, wherein the selection unit is configured to determine a point represented by the decimal part based on an edge represented by the edge information in the texel based on a value of the decimal part. An image generation system for selecting a texel in a direction in which the texel is located. 6. An image generation method for an image generation system for generating an image by performing an image mapping process using texture information, wherein a plurality of texels constituting the texture information are provided for each texel. Image generation characterized by including a step of assigning edge information representing an edge of a texture oblique to a direction of a coordinate axis of a texel, and a mapping processing step of performing an image mapping process based on the assigned edge information. Method. 7. The mapping process according to claim 6, wherein the mapping process comprises: acquiring a plurality of pieces of texture information corresponding to a plurality of texels; and obtaining one piece of texture information from the plurality of pieces of acquired texture information based on the edge information. And a step of mapping the selected texel to a pixel to be processed. 8. A computer usable program storing a program for generating an image by performing an image mapping process using texture information including color information and additional information stored in a predetermined storage area for each texel. In the information storage medium, the texture information includes edge information representing an edge of a texture oblique to a direction of a coordinate axis of the texel, and at least one of a plurality of texels is determined based on predetermined texel coordinate designation information and the edge information. An information storage medium including a program for causing a computer to execute: selecting means for selecting one texel; and means for performing mapping processing using texture information of the selected texel. 9. The information storage medium according to claim 8, further comprising a program for causing a computer to store the edge information in the storage area as at least a part of the additional information. 10. The texel coordinate designation information according to claim 8, wherein the texel coordinate designation information includes a two-dimensional coordinate value, and the selecting unit determines the texel coordinate information based on a fractional part of two numerical values of the two-dimensional coordinate value. An information storage medium, wherein the texel is selected. 11. The information storage medium according to claim 10, wherein the selection unit selects one texel from a plurality of texels near the texel based on the two numerical values. 12. The point according to claim 10, wherein the selection unit is configured to select a point represented by the decimal part based on an edge represented by the edge information in the texel based on a value of the decimal part. An information storage medium characterized by selecting a texel in a direction in which the texel is located.