JP3238567B2 - Image generation method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image generating method and apparatus for generating an image from compressed image data and drawing data in computer graphics. Particularly, it is suitable for a device to which computer graphics is applied, such as a video game machine and a graphic computer, where high visualization (visualization) performance is required with limited hardware resources.

[0002]

2. Description of the Related Art In computer graphics,
Generally, when a 3D (3D = 3D) graphics system is used to draw a realistic object (an object to be drawn is referred to as an object), first, the surface of the object is formed by a plurality of polygons (drawing device). Decomposes the smallest unit (triangle or quadrangle) of a figure to be handled into polygons, and stores each polygon in the frame memory (video RAM) corresponding to the monitor display screen in order.
By drawing, an image that looks three-dimensional is reconstructed.

[0005] In a general image generating apparatus of this type, a dedicated drawing device is provided between a CPU and a frame memory in order to perform processing at a high speed. When generating an image, the CPU does not directly access the frame memory, but sends a command for drawing a basic figure (polygon) such as a triangle or a rectangle (hereinafter, simply referred to as a drawing command) to the drawing device. The drawing device interprets the command sent from the CPU and draws a figure in the frame memory.

FIG. 14 shows a specific example of a drawing method. That is, when displaying a rectangular parallelepiped object having vertices A to G as shown in FIG. 14A, first, as shown in FIG. 14B, the object is divided into three rectangular polygons Pa, Pb, and Pc.
Decompose into

[0005] Then, the CPU, as shown in FIG.
Drawing command IP corresponding to each polygon Pa, Pb, Pc
a, IPb, and IPc are created. Drawing command IPa, IP
b and IPc are polygons Pa, Pb and Pc, respectively.
Each polygon P to determine the display position on the display screen of
a, Pb, Pc vertex coordinates (Ax, Ay) to (Dx, D
y), (Cx, Cy) to (Fx, Fy), (Bx, B
y) to (Gx, Gy) and the polygons Pa, Pb, P
It consists of information CODE such as the color in c.

The CPU transfers the drawing command to the drawing device. The drawing device draws an object in the frame memory based on the drawing command. If the drawing data in the frame memory is converted into an analog signal and supplied to the monitor display device, the object is displayed at the display position specified by the drawing command.

[0007] A CD-I disk has been developed as a CD-ROM disk. In addition to audio data, image data, text data, computer programs and the like are recorded on the CD-I disk. Has been recognized. This kind of CD-ROM
Image data is compressed and recorded on the disc.

Therefore, in parallel with the above-mentioned 3D graphics system, an image decompression device and the aforementioned CD-ROM
In many cases, a digital moving image reproduction system combining a secondary storage device such as a disk is mounted. This is because the digital moving picture playback system has an advantage that it is inferior to the 3D graphics system in interactivity, but has an advantage that it can easily play back an image that is difficult to express with the 3D graphics system. Often used as a complement to 3D systems.

FIG. 16 shows a configuration example of a system in which a conventional 3D graphics system and a moving image reproduction system are combined.

In FIG. 16, reference numeral 10 denotes a main bus, which includes a CPU 11, a main memory 12, and a CD-R.
The OM decoder 13 and the image decompression device unit 14 are connected. The CD-ROM decoder 13 outputs a CD-ROM disc (CD-ROM) reproduced by the CD-ROM driver 15.
I disk).

In this case, in addition to drawing commands such as computer graphics and animation data, still image data of a natural image and image data of a moving image are stored on a CD-ROM disk by DCT (Discrete Cosine Transform). Data compression is recorded.

The data decoded by the CD-ROM decoder 13 is transferred to the main memory 12 once. Then, the CPU 11 sends the drawing command from the main memory 12 to a FIFO (First In First Out).
Although the coordinates are transferred to the drawing unit 19 via the buffer, the vertex coordinate values may be changed. In this case, the coordinate calculation unit 17 is connected between the CPU 11 and the drawing unit 19 by a FIFO buffer. Is provided via That is,
The drawing command from the CPU 11 is transferred to the coordinate calculation unit 17 via the FIFO buffer 16, and a new vertex coordinate value of each polygon is obtained and rewritten if necessary.

The drawing command from the coordinate calculation unit 17 is sent to the drawing unit 1 via the FIFO buffer 18.
9 is transferred. The drawing device unit 19 includes the frame memory 2
The object is drawn at 0 according to the drawing command.

The CPU 11 sends the compressed image data in the main memory to the image data decompression device 14.
The image decompression unit 14 has a dedicated local bus and local frame memory 21 and decompresses compressed image data. The decompressed image data is stored in the local frame memory 21.

In this case, the image decompression device 14 is configured as shown in FIG.
The following three circuit blocks are provided as hardware. That is, in FIG. 17, the compressed data input through the input terminal 31 is decoded by the Huffman code decoder 32. The decoded image data is supplied to an inverse quantization circuit 33, where an inverse quantization process is performed. The output of the inverse quantization circuit 33 is supplied to an inverse discrete cosine transform circuit (inverse DCT circuit) 34,
It is T-converted and returned to decompressed data. The decompressed data is output data of an 8 × 8 matrix configuration. This output data is derived through a terminal 35.

The decompressed image data from the local frame memory 21 is sent to a multiplexer 23 via a D / A converter 22. Also, the frame memory 2
The drawing image data from 0 is sent to the multiplexer 23 via the D / A converter 24.

The multiplexer 23 selects either the drawn image data or the decompressed image data, sends it to the image monitor 25, and displays the image on the display screen.

[0018]

In the system shown in FIG. 16, in the part of the 3D system for drawing, the FI
Since the portion beyond the FO buffer 16 is separated from the main bus 10, it appears that the load on the main bus can be reduced at first glance, but there are the following problems.

Drawing by the drawing unit 19 and the CPU 11
Is issued asynchronously, so that a considerable number of FIFO
The number of stages must be prepared. In the worst case, the capacity of the drawing command for the polygons that compose one screen is the FIFO buffer 1
Required for 6 and 18.

The control of the CPU 11 is performed by the drawing unit 19.
, And a result of an external input such as an operation input from a user's operation input means, for example, a control pad, is not immediately reflected on the screen display. Similarly, it is difficult to change or cancel the drawing command in the middle. In particular,
This poses a serious problem in the field of video game machines because a real-time response is required to an operation input from a control pad.

Further, it is necessary for the CPU 11 to finally input the drawing command to the FIFO buffer 16, and DMA (direct memory access) cannot be used for the memory copy that occurs here. Therefore, after all, the load on the main bus is not reduced much.

The moving picture reproducing system has the following problems. In the above-described conventional system, the decompression of the compressed moving image data is all performed by the image decompression device 14 which is dedicated hardware. For this reason, in the conventional system, the hardware is required completely for image expansion completely independently of the CPU 11.

Further, the image decompression device requires a dedicated local memory 21. Furthermore, in the conventional system, since the expanded image data and other image data are switched and supplied to the image monitor device 25 by the multiplexer 23, the composite of the expanded image data and the other image is performed. Was not possible and very inconvenient. In particular, it has not been possible to perform special image synthesis such as synthesis of a moving image and an image such as computer graphics.

In view of the above, it is an object of the present invention to realize real-time image generation capable of immediately responding to an external input with a small amount of hardware resources.

Further, the present invention provides an image generation apparatus capable of easily reusing expanded image data and performing special image synthesis such as synthesis of an image such as a moving image and computer graphics. It is also an object to provide a device.

[0026]

In order to solve the above-mentioned problems, an image generation method according to the present invention comprises a CPU, an area for storing compressed image data, and data obtained by expanding the compressed image data. A main memory having an area for storing compressed image data, and an image decompression device for decompressing compressed image data are connected to a system bus,
A method of generating an image by using data of said main memory , wherein said image decompression device unit uses said system bus to compress the image data of said main memory without intervention of said CPU. And the image data decompressed by the image decompression device is transferred to the main memory using the system bus .

Further, the image generating method according to the second aspect of the present invention,
A method for generating an image using a system in which a CPU, a main memory having an area for storing a drawing command sequence, and a drawing device unit for performing drawing based on the drawing command sequence is connected to a system bus. There is no CPU
In addition, the drawing instruction sequence in the main memory is stored in the system.
Use Mubasu transferred to the drawing device portion, the drawing instrumentation
An image is generated by the placement unit using the drawing command sequence .

Claims to which the method of the invention of claim 1 is applied.
The image generating apparatus according to the third aspect of the present invention includes a reference numeral
And the system bus 41, the CPU
42, an area for storing the compressed image data, and
Area for storing data obtained by expanding compressed image data
Main memory 43 having the compressed video data
An image decompression device 51 for decompressing data,
U without the intervention of the compressed image data in the main memory.
Data using the system bus to the image decompression device section.
And the image decompressed by the image decompression device.
Transfer data to main memory using the system bus
Connected to the transfer unit 45 for transmitting
A display image is generated using the decompressed image data.
It is characterized by having made it.

The image generating apparatus according to claim 4 to which the method according to claim 2 is applied,
A CPU 42, a main memory 43 having an area for storing a drawing command sequence, a drawing device unit 61 for performing drawing based on the drawing command sequence, and the system Connecting to a transfer unit 45 for transferring to the drawing unit using a bus,
The drawing device unit uses the drawing command sequence to generate an image.
Is generated.

According to a fifth aspect of the present invention, in the image generating apparatus according to the fourth aspect, the main memory stores an area for storing compressed image data and data obtained by expanding the compressed image data. The system bus is connected to an image decompression device unit that performs decompression processing of the compressed image data, and the transfer device unit transfers the drawing command and without the intervention of the CPU. the compressed image data of said main memory sheet
Use Sutemubasu transferred to the image decompression device portion, also the image data decompressed by the image decompression device portion Shi
The data is transferred to the main memory using a stem bus .

The image generating apparatus of the present invention is provided with operation input means, and the drawing command is generated according to the operation input by the operation input means, so that a game machine can be configured.

An address on the main memory of a drawing command to be executed next is set as a part of the drawing command, and the transfer unit transfers the drawing command from the main memory to the drawing unit according to the address. It is good to adopt a configuration in which the data is sequentially transferred to

Furthermore, the decompressed image data in the main memory can be mixed with a so-called polygon drawing command by using the same data format as the drawing command.

[0034]

According to the first aspect of the present invention, the compressed image data is transferred to the CPU by the transfer unit.
Irrespective of the CPU, the image data is transferred to the image decompression device section 51 and decompressed and decoded. Thus, the compressed image data is expanded by the image expansion device section 51 and then expanded on the main memory 43 again. Therefore, the CPU 42 can reuse the expanded image data on the main memory 43.

In addition, the main memory 4 for compressed image data
3 is transferred from the image decompression device 51 to the main memory 43 regardless of the CPU 42, and the CPU 42 is involved in the transfer process. A process for displaying an image can be performed by using a portion that does not need to be performed, and high-speed processing can be performed.

According to the second aspect of the present invention, the drawing command sequence in the main memory 43 is provided without the CPU 42 intervening.
It is transferred to the drawing unit. Therefore, the CPU 42
For example, while generating a drawing command in response to an input operation of an operation input unit such as the control pad 71, the transfer device unit can transfer the drawing command sequence to the drawing device unit using the available space of the system bus 41. It is possible to perform real-time image generation.

In the case of the image generating apparatus according to the fifth aspect of the present invention, the image decompressing unit 51 and the drawing unit 61 are both connected to the system bus 41, and are connected to the main memory 43. The transfer of the image data or the drawing command is performed by the transfer unit 45. Therefore, the image decompression device section 51 does not require a local memory.

Further, the CPU 42 not involved in the transfer of the image data and the drawing command can perform a process for organically combining the image data with the drawing command without deteriorating the real-time property. In particular, when the decompressed image data is in the same data format as the drawing command, by combining this decompressed image data with a so-called polygon drawing command, the drawing image is synthesized between the decompressed images. be able to.

[0039]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows an example of the configuration of an image generating apparatus according to an embodiment of the present invention. This example is an embodiment of a game machine having a 3D graphic function and a moving image reproducing function.

In FIG. 1, reference numeral 41 denotes a system bus (main bus). The system bus 41 includes a CPU 4
2. Main memory 43, sorting controller 45
Is connected.

On the system bus 41, an image decompression device 51 is provided with an input FIFO buffer memory (hereinafter, referred to as an FIFO buffer memory).
The FIFO buffer memory is abbreviated as a FIFO buffer) 54 and a FIFO buffer 55 for output. Also, the CD-ROM decoder 52 has the FI
Through the FO buffer 56, the drawing device 61
Each is connected to the system bus 41 via a buffer 62.

Reference numeral 71 denotes a control pad as operation input means, which is connected to the system bus 41 via an interface 72. Furthermore, the system bus 4
1 is connected to a boot ROM 73 in which a program for starting up the game machine is stored.

The CD-ROM decoder 52 is a CD-RO decoder.
It is connected to the M driver 53 and decodes application programs (for example, game programs) and data recorded on a CD-ROM disc mounted on the CD-ROM driver 53. On a CD-ROM disk, for example, image data of a moving image or a still image compressed by a discrete cosine transform (DCT) and image data of a texture image for modifying a polygon are recorded. The application program on the CD-ROM disk contains a polygon drawing command. FIF
The O-buffer 56 has a capacity for one sector of data recorded on a CD-ROM disk.

The CPU 42 manages the entire system. Further, the CPU 42 performs a part of the processing when the object is drawn as a group of many polygons. That is, the CPU 42 creates a drawing command example for generating a drawing image for one screen on the main memory 43 as described later.

The CPU 42 has a cache memory 46, and some of the CPU instructions can be executed without fetching them from the system bus 41.
Further, the CPU 42 is provided with a coordinate arithmetic unit 44 as a CPU internal coprocessor for performing a coordinate conversion operation on a polygon when creating a drawing command. The coordinate calculation unit 44 performs three-dimensional coordinate conversion and conversion from three-dimensional to two-dimensional on the display screen.

As described above, since the CPU 42 has the instruction cache 46 and the coordinate calculation unit 44 inside,
Since the processing can be performed to some extent without using the system bus 41, the system bus 41 is easily opened.

The image decompression device 51 decompresses compressed image data reproduced from a CD-ROM disc. In this example, the image decompression device section 51 is composed of the hardware of the inverse quantization circuit 33 and the inverse discrete cosine transform circuit 34 of the image decompression device section 14 shown in FIG. In the Huffman code decoder 32 shown in FIG. 17, the CPU 42 performs the processing as software. Therefore, the configuration of the image decompression device 51 is simpler than that of the conventional image decompression device 14.

In the case of this example, the image decompression device 51
An image of one sheet (one frame) is divided into small areas of, for example, about 16 × 16 pixels (pixels) (hereinafter, referred to as macroblocks), and image expansion and decoding are performed in macroblock units. Then, data is transferred to and from the main memory 43 in macroblock units. Therefore, the FIFO buffers 54 and 55 have a capacity for a macroblock.

A frame memory 63 is connected to the drawing unit 61 via the local bus 11. The drawing device section 61 executes a drawing command transferred from the main memory 43 via the FIFO buffer 62 and writes the result to the frame memory 63. FIFO buffer 6
2 has a memory capacity for one drawing command.

The frame memory 63 includes an image memory area for storing a drawing image, a texture memory area for storing a texture image, and a table memory area for storing a color lookup table (color conversion table CLUT).

FIG. 2 shows the memory space of the frame memory 63. The frame memory 63 is addressed by a two-dimensional address of a column and a row. In this two-dimensional address space, the area AT is a texture memory area. In this texture area AT, a plurality of types of texture patterns can be arranged. AC
Is a table memory area of the color conversion table CLUT.

As described later, the color conversion table CLUT
Is transmitted from the CD-ROM to the CD-ROM decoder 5.
2 and transferred to the frame memory 63 by the sorting controller 45. The data of the texture image on the CD-ROM is decompressed by the image decompression unit 51 and transferred to the frame memory 63 via the main memory 43.

In FIG. 2, AD is an image memory area, which has an area for drawing and a frame buffer area for two surfaces of an area for display. In this example, the frame buffer area currently used for display is called a display buffer, and the frame buffer area where drawing is being performed is called a drawing buffer. In this case, while one drawing is being performed using the drawing buffer, the other is used as a display buffer, and when the drawing is completed, both buffers are switched to each other. The switching between the drawing buffer and the display buffer is performed in synchronization with the vertical synchronization at the end of the drawing.

The image data read from the display buffer of the frame memory 63 is output to the image monitor 65 via the D / A converter 64 and displayed on the screen.

The sorting controller 45 is
It has the same functions as a DMA controller
Transfer of image data between the file 43 and the image decompression device 51.
Or from the main memory 43 to the drawing unit 61.
Transfers drawing instruction sequence and composes transfer device
are doing. This sorting controller 45 has a CP
Other devices such as U42 and control pad 71
Remove the gap opening the system bus 41 and remove the CPU 42
The above-described transfer processing is performed without the intervention of. In this case, the CPU
42 is a sorting control for opening the system bus 41.
Can be notified to the
The controller 45 forcibly sends the bus to the CPU 42.
Opening may be required.

The main memory 43 has a memory area for compressed image data and a memory area for decompressed image data subjected to decompression decoding for image data of moving images and still images. Further, the main memory 43 includes a memory area for graphics data such as a drawing instruction sequence (hereinafter referred to as a packet buffer).

The pallet buffer is used for setting the drawing command sequence by the CPU 42 and transferring the drawing command sequence to the drawing device unit, and is shared by the CPU 42 and the drawing device unit 61. In this example, a packet buffer for setting a drawing instruction sequence (hereinafter, referred to as a setting packet buffer) and a packet buffer for transfer are set so that the CPU 42 and the drawing device unit 61 operate in parallel. (Hereinafter referred to as an execution packet buffer). When one packet is set as a set packet buffer, the other is used as an execution packet buffer. The functions of two packet buffers are exchanged. The processing of this device will be described below.

[Fetching of Data from CD-ROM Disk] When power is turned on to the apparatus (game machine) shown in FIG. 1 and a CD-ROM disk is loaded, boot RO is executed.
A program for performing a so-called initialization process for executing the game of M73 is executed by the CPU.
Then, the recording data of the CD-ROM disc is taken. At this time, the decoding process of each user data is performed based on the identification information ID in the user data of each sector of the CD-ROM disk, and the data is checked. Based on the result of this check, the CPU 42
A process corresponding to the reproduction data having the content indicated by D is executed.

That is, the compressed image data, the drawing command, and the program executed by the CPU 42 are read out from the CD-ROM disk via the CD-ROM driver 53 and the CD-ROM decoder 52, and are sorted by the sorting controller 45 into the main memory 43. Is loaded. Then, of the loaded data, information of the color conversion table is transferred to the area CLUT of the frame memory 63.

[Decompression and Transfer of Compressed Image Data] Of the input data of the main memory 43, the compressed image data
After the PU 42 performs the decoding process of the Huffman code, the data is written into the main memory 43 by the CPU 42 again.
Then, the sorting controller 45 stores the decoded image data of the Huffman code in the main memory 4.
3 through a FIFO buffer 54 to the image decompression device 5
Transfer to 1. The image decompression device section 51 performs dequantization processing and inverse DCT processing to perform decompression decoding processing of image data.

The decompressed image data is transferred by the sorting controller 45 to the main memory 43 via the FIFO buffer 55. In this case, the image decompression device 5
As described above, No. 1 performs image data decompression processing in macroblock units. For this reason, the compressed data in units of macro blocks is transmitted from the main memory 43 to the input
The data is transferred to the O buffer 54 by the sorting controller 45. When the decompression decoding process for one macroblock is completed, the image decompression device unit 51 puts the resulting decompressed image data into the output FIFO buffer 55 and outputs the compressed data of the next macroblock from the input FIFO buffer 54. Take it out and perform decompression decoding processing.

The sorting controller 45 operates when the system bus 41 is open and the image decompression device 5
If the first output FIFO buffer 55 is not empty, the decompressed image data of one macroblock is transferred to the main memory 43, and the compressed image data of the next one macroblock is input from the main memory 43 to the image decompression device 51. The data is transferred to the FIFO buffer 54.

The CPU 42 transfers the decompressed data to the frame memory 63 via the drawing unit 61 when a predetermined amount of macroblocks of decompressed image data are stored in the main memory 43. At this time, if the decompressed image data is transferred to the image memory area AD of the frame memory 63, it is used as a background moving image without any change.
Will be displayed. Further, the data may be transferred to the texture memory area AT of the frame memory 63.
The image data in the texture memory area AT is used as a texture image for modifying a polygon.

[Processing and Transfer of Drawing Instruction Sequence] The polygons forming the surface of the object are drawn in order from the polygon located deeper in the depth direction in accordance with Z data which is three-dimensional depth information. An image can be stereoscopically displayed on the two-dimensional image display surface. The CPU 42
In this way, a drawing command sequence for causing the drawing device 61 to perform drawing is created on the main memory 43 in order from the polygon located deeper in the depth direction.

In computer graphics, a so-called Z-buffer method is used in which Z data is stored in a memory for each pixel and display priorities of polygons are determined. However, this Z-buffer method requires the use of a large-capacity memory to store the Z data. Therefore, in this example, the processing for determining the display priority of polygons is performed by the CPU 4 as follows.
2 do.

That is, in this example, as shown in FIG. 3A, the polygon drawing command IP includes the polygon drawing data P
In addition to D, a tag TG is added. In the tag TG,
The address on the main memory 43 where the next drawing command is stored is written. The polygon drawing data PD includes identification data IDP indicating the content of the drawing command and data such as vertex coordinates of the polygon. When the drawing command IP is, for example, a drawing command for a quadrilateral polygon and the inside of the polygon is mapped with one color, the identification data IDP indicates the same.

FIG. 3B shows a case of a rendering command of a quadrilateral polygon, in which coordinates of four points (X0, Y0), (X1, Y1),
(X2, Y2), (X3, Y3) and three primary color data (R,
G, B).

The CPU 42 calculates the movement of the object and the viewpoint based on the user's operation input from the control pad 71, and creates a polygon drawing command sequence on the main memory 43. Next, the tag of the polygon drawing command sequence is rewritten in the display order according to the Z data. At this time, the address of each drawing command on the main memory 43 is not changed, and only the tag is rewritten.

When the drawing command sequence is completed, the sorting controller 45 sequentially follows the tag TG of each drawing command and transfers the tag TG from the main memory 43 to the drawing device unit 61 for each drawing command. Therefore, the FIFO buffer 6
2 only needs to have a capacity for one drawing command.

In the drawing device section 61, since the sent data has already been sorted, as shown in FIG. 4, polygon drawing commands IP1, IP2, IP3,
, IPn are replaced by their tags TG1, TG2, TG3,.
The execution is sequentially performed according to TGn, and the result is stored in the drawing area AD of the frame memory 63.

The processing of the CPU 42 at the time of polygon drawing will be described with reference to the flowchart of FIG.

First, the CPU 42 sets the frame memory 63
Frame buffer area A of the image memory area AD
A command is issued to the drawing unit 61 to output the image data (which is a display buffer) to the image monitor 65 (step 101). Next, the operation input of the control pad 71 is read (step 102), and in accordance with the operation input, the coordinate value of the drawing instruction sequence A of one packet buffer (which is a set packet buffer) of the main memory 43 is updated. At the same time, the tag of each drawing command in the drawing command sequence A is rewritten (step 103).

In steps 101 to 103, the drawing command sequence B of the other packet buffer (which is an execution packet buffer) of the main memory 43 is
Is stored in the image memory area AD of the frame memory 63 via the drawing device 61 by the sorting controller 45.
Is transferred to the other frame buffer area B (which serves as a drawing buffer), and drawing is performed by the drawing device section 61 in real time using the drawing command sequence B.

Next, the process waits until the rendering by the rendering command sequence B is completed (step 104). That is, it is determined whether or not the transfer of the drawing command sequence B from the main memory 43 has all been completed and the drawing has been completed.

When the execution of the drawing instruction sequence B is completed,
Using the frame buffer area B of the frame memory 63 as a display buffer, the drawing image data is read out from this,
The drawing device unit 6 outputs the image data to the image monitor device 65.
An instruction is issued for 1 (step 105). At this time,
At the same time, the frame buffer area A of the frame memory 63
Is switched to the drawing buffer.

Next, the operation input of the control pad 71 is read (step 106), and the coordinates of the drawing command sequence B of the other packet buffer (which is a set packet buffer) of the main memory 43 are read in response to the operation input. The value is updated and the tag of each drawing command in the drawing command sequence B is rewritten (step 107).

Then, from Step 105 to Step 10
7, the drawing instruction sequence A of one packet buffer (which is an execution packet buffer) of the main memory 43 is transmitted by the sorting controller 45 to the one frame buffer area A of the frame memory 63 via the drawing device unit 61. (Which is a drawing buffer), and the drawing device section 61 performs drawing in real time by the drawing command sequence A.

Next, the process waits until the rendering by the rendering instruction sequence A is completed (step 108). That is, it is determined whether the transfer of the drawing instruction sequence A from the main memory 43 has all been completed and the drawing has been completed.

When the execution of the drawing instruction sequence A is completed,
Using the frame buffer area A of the frame memory 63 as a display buffer, the drawing image data is read out from this,
The drawing device unit 6 outputs the image data to the image monitor device 65.
An instruction is issued for 1 (step 10109). At this time, the frame buffer area B is simultaneously switched to the drawing buffer. Then, returning to (Step 102), the above processing is repeated. Perform the above operations 30 times
By repeating at 60 times / second, an image with a motion can be displayed.

At this time, as is apparent from the above description, the CPU 42 and the drawing device section 61 operate in parallel. That is, the CPU 42 sets the address value of the tag of each drawing command of the drawing command sequence of the setting packet buffer of the main memory 43 to the address value of the main memory 43 in which the next drawing command indicated by the arrow in FIG. 6A is stored. Rewrite sequentially. At the same time, the sorting controller 45 reads out each drawing command from the drawing packet buffer of the main memory 43 by following the tag of each drawing command as shown by an arrow in FIG. 6B, and transfers it to the drawing device unit 61. The drawing device section 61 executes drawing according to the drawing command sequence.

In this case, as shown in FIG.
While the drawing command sequence is being created, drawing by the drawing command sequence created earlier is performed by the drawing device unit 61, and after the drawing is completed, drawing by the currently created drawing command sequence is performed. Will be executed.

At the time of this polygon drawing, the data is sent to the gradient calculating unit of the drawing device section 61, where the gradient is calculated. The gradient calculation is a calculation for calculating the inclination of the plane of the mapping data when filling the inside of the polygon with the mapping data by polygon drawing. In the case of texture, polygons are filled with texture image data, and in the case of Gouraud shading, polygons are filled with luminance values.

When a texture is pasted on a polygon constituting the surface of the object, the texture data of the texture area AT is subjected to a two-dimensional mapping conversion. For example, texture patterns T1, T2, T as shown in FIG.
3 is converted into coordinates on a two-dimensional screen so as to fit the polygon of each surface of the object as shown in FIG. 8B.
The texture pattern T1, thus mapped,
T2 and T3 are attached to the surface of the object OB1, as shown in FIG. 8C. And this is the image memory area AD
And is displayed on the display screen of the image display monitor 65.

In the case of a still image texture, the texture pattern on the main memory 43 is stored in the drawing device 61.
Through the texture area A on the frame memory 63
Transferred to T. The drawing device section 61 pastes this on the polygon. Thereby, the texture of the still image is realized on the object. The data of the texture pattern of the still image can be recorded on a CD-ROM disk.

Further, the texture of a moving image is possible.
That is, in the case of the moving image texture, the compressed moving image data from the CD-ROM disk is once read into the main memory 43 as described above. And
The compressed image data is sent to the image decompression device 51. The image data is expanded by the image expansion device 51.
At this time, as described above, a part of the decompression process is performed by the CPU.
42 bears.

Then, the expanded moving image data is sent to the texture area AT on the frame memory 63. Since the texture area AT is provided in the frame memory 63, the texture pattern itself can be rewritten for each frame. As described above, when a moving image is sent to the texture area AT, the texture is dynamically rewritten and changed every frame. If texture mapping to polygons is performed using the moving image in the texture area, the texture of the moving image is realized.

As described above, if the image data expanded by the image expansion device section 51 is sent to the image memory area AD on the frame memory 63, the moving image of the background image can be displayed on the screen of the image monitor 65. Yes, CPU
Image memory area AD only with the drawn image created in step 42
Can be embedded and drawn on the screen of the image display monitor 65. Also, in the image memory area AD, the CD
It is also possible to draw an object by polygon drawing by the CPU 42 on a still image obtained by expanding image data from a ROM disk.

As described above, the sorting controller 45 transfers drawing commands and image data through the gap where the system bus 41 is open without the intervention of the CPU 42. An explanatory diagram of this state is shown in FIG.

FIG. 9A shows the system bus 41 of the CPU 42.
In the hatched portions, the work using the system bus 41 is performed, and blank sections 81 and 8 are shown.
2,... 88 indicate that the CPU 42 has opened the system bus 41.

FIGS. 9B and 9D show the F of the image decompression device 51.
The state of the I / O buffers 54 and 55 is shown, and the shaded portion indicates the state where data is accumulated. FIG. 9C shows the decompression processing timing in the image decompression device unit 51.

FIG. 9E shows the FIF of the drawing unit 61.
The state of the O-buffer 62 is shown, and the hatched portion indicates the state where data is accumulated. FIG. 9F
The drawing execution timing in the drawing device unit 61 is shown.

When the FIFO buffers 54 and 62 become empty, respectively, they issue a transfer request to the system bus 41. Further, the FIFO buffer 55 issues a transfer request when this becomes full. Sorting controller 45
Are the times 81 to 8 during which the CPU 42 has opened the bus 41.
At 8, the transfer is performed from or to the FIFO buffer that has issued the transfer request.

When transfer requests overlap, it is determined which transfer request is to be executed according to a predetermined priority. When transfer requests from buffers having the same priority are overlapped, the order is determined by a so-called round-robin method so that the transfer request with the highest priority when the previous transfer request has the lowest priority. In the example of FIG.
The transfer requests from the buffers 54 and 55 are FIFO
The priority is higher than the transfer request of the buffer 62.

In the example of FIG. 9, the bus opening times 81, 85,
At 87, since the transfer request is issued because the FIFO buffer 54 is empty, the sorting controller 45 transfers the compressed image data from the main memory 43 to the FIFO buffer 54 as shown in FIG. 9B. When the FIFO buffer 54 becomes full, as shown in FIG. 9C, when the decompression decoding of the previous data is completed, the image decompression unit 51 fetches the compressed data from the FIFO buffer 54 and starts decompression decoding processing. . The FIFO buffer 54 issues a transfer request to the bus 41 again.

On the other hand, when the image decompression decoding is completed, the image decompression device section 51 puts the decompressed data into the FIFO buffer 55. The FIFO buffer 55 issues a transfer request to the bus 41 when it is filled with the decompressed data. In the example of FIG.
The sorting controller 45 includes a bus open section 83,
At 86, a transfer for this request is performed. That is, the decompressed data is transferred to the main memory 43.

Further, in the example of FIG. 9, the transfer request from the FIFO buffer 62 of the drawing device section 61 is executed in the bus open sections 82, 84 and 88 as shown in FIG. Device section 6
1 is transferred to the drawing command. Then, the transferred drawing command is executed by the drawing device section 61 after the execution of the previously transferred drawing command is completed, as shown in FIG. 9F.

When the image data decompressed and decoded by the image decompression device section 51 is transferred from the main memory 43 to the frame memory 63, the following transfer command is used in this example. The CPU 42 converts the decompressed image data into the transfer command format in this manner.

That is, FIG. 10 is a diagram showing the structure of this transfer instruction. This transfer command has substantially the same format as the drawing command, and is provided with a tag TG at the beginning, followed by identification information IDP. The tag TG includes an address value of the main memory 43 in which the next drawing command or transfer command is stored, similarly to the drawing command. IDP includes:
A code indicating that this is a transfer instruction is described.

In FIG. 10, the next data "H" and "W" indicate the height and width of the decompressed data area to be transferred. The height and width of this transfer area are 1
It corresponds to the area of the screen for the frame. Also,
The data “X” and “Y” indicate the coordinates of the transfer destination.
These coordinates indicate the upper left coordinates of the rectangular area since the transfer area is rectangular. If the transfer destination is in the image memory area AD of the frame memory 63, the coordinates are in the area AD, and the texture area A
If it is in T, the coordinates are in the area AT.

The coordinates “X”, “Y” from the tag TG
Are the headers of the transfer instructions, and are followed by the decompressed image data PIX0, PIX1, PIX2,.
IXn is included in this transfer instruction. Then, in units of this transfer command, the sorting controller 45
The decompressed image data is transferred from the main memory 43 to the drawing device section 61.

As described above, the image decompression unit 51 divides an image of one frame into macroblocks consisting of 16.times.16 pixels, and performs decompression decoding in macroblock units. Now, for example, assuming an image in which one frame is composed of horizontal × vertical = 320 × 240 pixels, as shown in FIG. 11, one frame is divided into 300 macroblocks.

In transferring these 300 macroblocks to the drawing device section 61, if a transfer command is created in macroblock units, the overhead of the header portion becomes too large. Accordingly, in this example, a plurality of (five in FIG. 12) macroblocks in one column in the vertical direction are connected, and this is used as a unit to be transmitted by a transfer instruction.

FIG. 13 shows an example of the first transfer instruction of one frame.
Shown in That is, in FIG. 13, the coordinates “X” and “Y” are “0” and “0”. In the next transfer command, the coordinates “X” and “Y” become “16” and “0”.

As described above, since the decompressed image data is also converted into the transfer command format similar to the drawing command, the tag TG
Is used, the polygon drawing command and the transfer command are mixed and transferred by the sorting controller 45, and the frame memory 6 by the drawing device unit 61 is used.
3 can be executed.

According to the apparatus of the above embodiment, the following effects can be obtained. That is, the drawing command sequence on the main memory 43, the compressed image data and the decompressed image data
Since the data is transferred without the intervention of the CPU 42 through the gap where the U42 opens the system bus 41, the system bus 41 can be used efficiently and in a time-division manner.

Further, since all the drawing instruction sequences are stored in the main memory 43, the CPU 42 can directly control the system at any time, so that immediate control according to an external input from the control pad 71 or the like is possible. Therefore, the reaction speed of screen display can be improved.

Similarly, all the moving image data is once
Since the moving image data is stored in the main memory 43, the CPU 42 can always directly control the moving image data. Therefore, immediate control according to an external input from a control pad or the like is possible, and the response speed of screen display can be improved.

The FIFO associated with the drawing unit 61
Since the number of stages in the buffer is sufficient for one drawing command, the circuit scale of the drawing device can be reduced.

Since the image decompression device 51 uses the main memory as a buffer, it is not necessary to have a local memory. Moreover, the FI attached to the image decompression device section
Since the number of stages of the FO buffer is sufficient for one macroblock, the circuit scale of the image decompression device can be reduced.

Since the address of the next drawing command to be executed in the main memory is incorporated in the drawing command, when the drawing order changes, only the address value in the drawing command is stored. Just by changing, it is not necessary to rearrange the drawing command itself in the main memory at the changed address position. Therefore, the system bus 41
Can be reduced correspondingly.

Further, when an image that moves continuously is generated, the content of the drawing command sequence developed on the main memory rarely changes greatly between adjacent frames. Therefore, the address value is rarely changed, and in practice, it is often necessary only to change the coordinate value to the drawing command sequence of the previous frame, and the control is easy.

In the above example, the image data and the application program are recorded on the CD-ROM. However, other recording media such as a magnetic disk and a semiconductor memory such as a memory card may be used as the recording medium. Can also.

As a method for compressing image data, D
Although CT was used, various other image data compression methods can be used.

[0114]

As described above, according to the present invention, the drawing instruction sequence in the main memory is transferred without the intervention of the CPU through the gap where the system bus is open. 41 can be used efficiently and in a time-sharing manner. The transfer of the compressed image data and the decompressed image data on the main memory is performed in the same manner, so that the system bus can be used efficiently.

Further, since all the drawing instruction sequences are stored in the main memory, the CPU can directly control the system at any time, and thus can perform immediate control according to an external input from a control pad or the like. Therefore, the response speed of screen display can be improved, and a game machine with high real-time property can be easily realized.

Further, since the image decompression device uses the main memory as a buffer, it is not necessary to have a local memory. Moreover, a FIFO attached to the image decompression device section
The number of stages in the buffer may be equal to the expansion processing unit data. Further, by making the CPU bear a part of the image data decompression processing, the hardware scale as the image decompression processing circuit can be reduced.

The number of stages of the FIFO buffer associated with the drawing device section is sufficient for one drawing command. Therefore, the circuit scale of the system can be reduced.

Also, since the address value of the next drawing command to be executed in the main memory is incorporated in the drawing command, only the address value in the drawing command is changed when the drawing order changes. Just by changing, it is not necessary to rearrange the drawing command itself in the main memory at the changed address position. Therefore, the load on the system bus can be reduced.

Further, since the expanded image data of the compressed and transmitted image data is expanded in the main memory, any part of the image data in the main memory can be extracted and reused. And, for example, CPU
Can be combined with the image created in step (1).

At the time of data decompression processing, the main memory can be used as a processing buffer memory, so that a dedicated local memory for decompression processing as in the prior art is not required.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is a diagram for explaining a memory area in one embodiment of the present invention.

FIG. 3 is a diagram showing an example of a polygon drawing command in one embodiment of the present invention.

FIG. 4 is a diagram for explaining a polygon display order in one embodiment of the present invention.

FIG. 5 is a diagram illustrating a C in a drawing process according to an embodiment of the present invention;
It is a flowchart for demonstrating the process of PU.

FIG. 6 is a diagram for explaining a drawing command process of a CPU on a main memory and a drawing execution process in a drawing device unit according to an embodiment of the present invention;

FIG. 7 is a diagram for explaining parallel processing of a drawing command process of a CPU on a main memory and a drawing execution process in a drawing device unit according to an embodiment of the present invention;

FIG. 8 is a diagram for explaining texture mapping.

FIG. 9 is a diagram for explaining transfer control by a transfer device unit in one embodiment of the present invention.

FIG. 10 is a diagram illustrating an example of a data structure at the time of transferring image data according to an embodiment of the present invention.

FIG. 11 is a diagram illustrating an example of an image of one frame.

FIG. 12 is a diagram for explaining a transfer unit of image data in one embodiment of the present invention.

FIG. 13 is a diagram illustrating an example of a data structure when transferring image data according to an embodiment of the present invention.

FIG. 14 is a diagram for describing an example of a drawing method.

FIG. 15 is a diagram illustrating an example of a drawing command.

FIG. 16 is a diagram illustrating a configuration example of a system in which a conventional 3D graphics system and a moving image reproduction system are combined.

FIG. 17 is a block diagram of an example of a conventional image decompression device.

[Explanation of symbols]

 41 System Bus 42 CPU 43 Main Memory 44 Coordinate Operation Unit 45 Sorting Controller 46 Cache Memory 51 Image Decompression Unit 52 CD-ROM Decoder 53 CD-ROM Driver 54, 55 FIFO Buffer 61 Drawing Unit 62 FIFO Buffer 63 Frame Memory 65 Image display monitor 71 Control pad AD Image memory area AT Texture area

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Makoto Furuhashi 8-1-22 Akasaka, Minato-ku, Tokyo Inside Sony Computer Entertainment Inc. (72) Inventor Masayoshi Tanaka 8-1-1 Akasaka, Minato-ku, Tokyo No. 22 Sony Computer Entertainment Inc. (56) References JP-A-5-274417 (JP, A) JP-A-63-91787 (JP, A) JP-A-4-294684 (JP, A) JP-A-4-278678 (JP, A) JP-A-2-255991 (JP, A) JP-A-7-85308 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 1 / 20

Claims (8)

(57) [Claims] 1. A main memory having a CPU, an area for storing compressed image data, and an area for storing data obtained by expanding the compressed image data with respect to a system bus; An image decompression device for performing decompression processing of the image data and transferring the compressed image data in the main memory to the image decompression device using the system bus without the intervention of the CPU. in decompression processing and transfer device unit for the image data has been completed using the system bus for transferring to said main memory is connected, by using the expanded image data of said main memory, to generate a display image To
The main memory further includes an area for storing a drawing command.
For the system bus, based on the drawing command,
A drawing device unit for performing drawing is further connected, and the
The transfer unit is configured to execute the main menu without the intervention of the CPU.
Using the system bus to draw the drawing instructions of Mori
Transferred to the drawing device section, and the drawing device section
The image is generated using the image command, and each of the image commands is executed next to a part thereof.
Including the address of the drawing command to be
Only, the transfer device section performs the transfer according to the address.
Drawing commands are sequentially sent from the main memory to the drawing device.
An image generating apparatus characterized in that the image is transferred. 2. An expanded image data written in the main memory, the drawing command and is the same format data, according to claim 1, characterized in that the drawn command and can be mixed Image generation device. 3. An operation input means is connected to the system bus, and the drawing command is generated by the CPU in response to an operation input by the operation input means, thereby forming a game machine. 3. The image generation device according to 1 or 2 . Against 4. A system bus, a CPU, drawing
A main memory having an area for storing instructions ;
A drawing device unit that performs drawing based on a command and includes a buffer memory; and a transfer device unit that transfers the drawing command of the main memory to the drawing device unit using the system bus without the intervention of the CPU. And are connected,
By the imaging device section, the drawing command together are adapted to generate an image using, each of said drawing command, the drawing instruction to be executed next in a part, an address on the main memory The image generation apparatus, wherein the transfer device unit sequentially transfers the drawing command from the main memory to the buffer memory of the drawing device unit according to the address. 5. The image processing apparatus according to claim 1, wherein the main memory stores a compressed image data.
Data storage area and the compressed image data
An area for storing lengthened data, wherein the system bus
Includes an image to be subjected to a decompression process of the compressed image data.
A decompression device is connected, and the transfer device is connected to the CPU
Compressed image data in the main memory without the intervention of
Data to the image decompression device using the system bus.
Transfer, and the decompression process is completed by the image decompression unit.
The main image data using the system bus.
5. The method according to claim 4, wherein the data is transferred to a memory.
An image generation device according to claim 1. 6. expanded image data written in the main memory, the drawing command and is the same format data, according to claim 5, characterized in that the drawn command and can be mixed Image generation device. 7. A CPU and stores compressed image data.
Area to be compressed and data obtained by expanding this compressed image data.
Main memory with an area for storing data
An image decompression device that performs decompression processing of image data
Connected to the system bus and
A method of generating an image by using the compressed data of the main memory without the intervention of the CPU.
Image data using the system bus.
The image decompression device, and decompresses the image by the image decompression device.
Image data that has been processed using the system bus.
The data is transferred to the main memory, and the main memory further includes an area for storing a drawing command.
For the system bus, based on the drawing command,
A drawing device unit for drawing, and without the CPU,
The drawing command of the main memory is used by using the system bus.
A transfer device unit for transferring to the drawing device unit using
Connected to use the drawing command by the drawing device unit.
To generate an image, and further, each of the drawing instructions is executed next to a part thereof.
Including the address of the drawing command to be
The drawing command according to this address.
From the memory to the drawing unit
An image generation method characterized by the following. 8. A decompressed image written in said main memory.
The image data is data in the same format as the drawing command.
And can be mixed with the drawing command.
The image generation method according to claim 7.