JPH0865672A - Processor and method for compressing character data - Google Patents

Abstract

PURPOSE: To reduce data amount than that before data are compressed by performing compression by a prescribed logical system after data are provided with redundancy. CONSTITUTION: The character data of a color pallet form are compressed and address-translated data are synthesized. When the synthesized data are less than before data are compressed, the synthesized data are compressed as they are. On the other hand, when the data are larger than before data are compressed, the character data of the color pallet form are decomposed into the data of an RGB or YIG color space system by a first data system transforming part 22, afterwards, spatial correlation is removed by performing discrete cosine transform at an axial transforming part 23, energy is concentrated to one part of coefficient and further, data are quantized by a quantizing part 24. Then, when the data are returned to the color pallet form again by an inverse quantiation part 25, inverse axial transforming part 26 and second data system transforming part 27, the data are provided with redundancy. Then, these data provided with redundancy are compressed by the prescribed logical system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character data compression processing apparatus and a character data compression method for compressing and using character data in a television game system.

[0002]

2. Description of the Related Art Generally, as a data compression method of a character data compression processing apparatus which compresses and uses character data in a television game system, an encoding method by Huffman or the like and an arithmetic operation (four arithmetic operations) by Elias or the like are used. Encoding / Complexing Method Using Lempel
-Ziv code (slide dictionary method, dynamic dictionary method) and Ben
trey-Sleator-Tarjan-Wei (B
There is a universal coding method such as STW) algorithm and a dynamic Huffman coding method. These compression methods use a reversible algorithm that does not allow information loss before and after compression, and are generally called lossless (LOSSESS) compression.

An example of a conventional television game system (hereinafter referred to as a TV game) using computer software using lossless-compressed character data is shown in FIG. 3 (first conventional example) and FIG. 4 (first conventional example). 2). In general, color data is often used for character data used in a TV game in order to reduce the amount of data. In this case, since the character data is quantized color data, even if only 1-bit data is different, the output color is greatly changed. Therefore, when compressing character data using a color palette, the lossless compression that does not cause any data error is adopted.

Normally, the lossless compression method uses a variable-length code, and therefore an address conversion table for indicating where a desired character is after being compressed is required. In the case of lossless compression, some data cannot be compressed or the compression rate is very low. In this case, since the address conversion table becomes an overhead of the character data, the compressed data including the overhead may increase more than the character data before the compression, and therefore becomes slower than the data access speed before the compression. Become.

In order to deal with such a problem in processing speed, in the first conventional example shown in FIG. 3, the address conversion table 1 for storing the address conversion data is provided separately from the character file 2 for storing the character data. When the character is compressed, the character data in the character memory 2 is losslessly compressed, and the compressed character data and the address conversion data in the address conversion table 1 are combined and stored in the compressed data file 3. In FIG. 3, 6 is an address conversion data reading unit (buffer memory) that reads address conversion data from the synthetic data in the compressed data file 3, and 7 is a character data reading unit that reads character data from the synthetic data in the compressed data file 3. Reference numeral 8 is an address conversion unit, and 9 is a data expansion unit. In the first conventional example,
At the time of character expansion, the address conversion data and the character data are individually identified, the address conversion is performed on the address conversion data, and the data expansion is performed on the character data. And the central processing unit (CPU)
Graphic controller (GRA
The data decompressed by the PHIC CONTROLLER 5 is processed.

Alternatively, in the second conventional example shown in FIG. 4, a large buffer memory (FIFO) 11 is provided,
Access speed is buffered by earning the total compression rate within a certain period of time. Reference numeral 11 in FIG. 4 is a buffer memory for temporarily storing the decompressed data.

[0007]

In the TV game system using the compression method of the first conventional example shown in FIG. 3, two separate systems of the address conversion data reading unit 6 and the character data reading unit 7 are provided. A memory is required, which goes against the demand for a reduction in memory size.

Further, in the second conventional example shown in FIG. 4, since the processing stops when the decompressed data buffer memory 11 becomes empty, it cannot be guaranteed that any combination of character data will move in real time.

In view of the above problems, the present invention provides a character data compression processing apparatus and a character data compression method for guaranteeing an access speed before compression by reducing the amount of compressed data without using a separate system table. The purpose is to

[0010]

According to a first aspect of the present invention, there is provided a character table for storing character data in a color palette format, and the character data in the character table is stored in a predetermined color space system. A first data system conversion unit for converting into data, and a discrete cosine conversion is performed on the data of the predetermined color space system provided from the first data system conversion unit to remove spatial correlation and An axis transforming unit that concentrates energy on a coefficient, a quantizing unit that quantizes the cosine function given by the axial transforming unit, and inverse quantization is performed on the information quantized by the quantizing unit. A dequantization unit that performs code assignment, an inverse axis conversion unit that converts the data assigned the code by the dequantization unit into data of the predetermined color space method, and a conversion performed by the inverse axis conversion unit A second data format conversion section for converting the data of the predetermined color space format into a color palette format again, and a first compression section for compressing the data given from the character table by a predetermined logical format, A second compression unit that compresses the data given from the second data system conversion unit by the predetermined logical system, compressed data from the first compression unit, and compressed data from the second compression unit. A selection unit for selecting any one of the above, a switching unit for instructing switching of the selection of the selection unit, an address conversion table for storing address conversion data, address conversion data of the address conversion table, and the selection unit selected by the selection unit. The data switching unit includes a data synthesizing unit for synthesizing any of the compressed data, and a synthetic data storage file for storing the data synthesized by the data synthesizing unit. Is an initialization unit that initializes the equal distribution circuit so as to select the first compression unit at an initial stage, and the character data in which the data amount of the combined data combined by the data combining unit is quantized. Data determining means for determining whether or not the data amount is greater than the data amount, and the data determining means determines when the data amount of the composite data is greater than the quantized data amount of the character data. And an input switching means for switching the input of the selection unit so as to select the second compression unit.

The problem solving means according to claim 2 of the present invention is
A first compression step of compressing character data in the color palette format by a predetermined logical method, and in parallel with the first compression step, the quantized character data is converted into data of a predetermined color space method. A first conversion step, a second conversion step of performing discrete cosine conversion on the data of the predetermined color space method to remove spatial correlation and concentrate energy in some coefficients, and the second conversion step. A third transform step for quantizing the data that has been discrete cosine transformed in the step, a fourth transform step for dequantizing the data quantized in the third transform step again, and a fourth transform step In the fifth conversion step of converting the dequantized data into the data of the predetermined color space method, the data converted into the data of the predetermined color space method in the fifth conversion step is converted into A sixth conversion step for converting to a palette format, a second compression step for compressing the data converted in the sixth conversion step by the predetermined logical method, compressed data obtained in the compression step, and A first synthesizing step of synthesizing the address-converted data corresponding to, and whether the data amount of the synthetic data synthesized in the first synthesizing step is larger than the quantized data amount of the character data. And a data determination step for determining whether or not the data amount of the combined data is smaller than the quantized data amount of the character data in the data determination process, and stores the combined data in a predetermined file. The second compression when the data amount of the combined data is determined to be larger than the data amount of the quantized character data in the first storing step and the data determining step. A second synthesizing step of synthesizing the compressed data obtained in the above step and address conversion data corresponding to the compressed data, and the synthetic data synthesized in the second synthesizing step in the first synthesizing step. And a second storage step of storing the synthesized data in the predetermined file.

[0012]

In the character data compression processing apparatus according to claim 1 of the present invention and the character data compression method according to claim 2, in the initial stage, the initial setting means of the switching section selects the first compression section. Initialize the equal distribution circuit.
Then, the character data in the color palette format is compressed and the address conversion data is combined. When the combined data is smaller than that before the compression, the combined data is compressed as it is. On the other hand, when the data becomes larger than that before compression, the character data in the color palette format is decomposed into RGB or YIQ color space format data in the first data format conversion section, and then the discrete cosine conversion is performed in the axis conversion section. Is applied to remove spatial correlation, concentrate energy in some coefficients, and further quantize it in the quantizer. Then, when the data is returned to the color palette format again in the inverse quantizer, the inverse axis converter, and the second data method converter, the data has redundancy. If the data having such redundancy is compressed by a predetermined logical method, the degree of compression is dramatically improved as compared with the case of directly compressing the character data. In this case, the addition of the redundant component of the data using the visual characteristics can be suppressed as much as possible by finally including the address conversion data so that the data amount does not become larger than that before the compression so that the deterioration of the image can be suppressed as much as possible. . Further, as compared with the conventional example, an address conversion table of another system and a large buffer memory for buffering the access speed are unnecessary.

[0013]

【Example】

{First Embodiment} <Structure> FIGS. 1 and 2 show the T of the first embodiment of the present invention.
1 shows a character data compression processing device in a V game, FIG. 1 showing a character compression circuit, and FIG. 2 showing a character expansion circuit.

In FIG. 1, reference numeral 21 is a character table for storing character data, 22 is a first data format conversion unit for converting character data in a color palette format (hereinafter referred to as color palette data) to television data, and 23. Is a two-dimensional axis transform unit (discrete cosine transform unit: 2D-DCT), 24 is a quantizer unit (Q), 25 is an inverse quantizer unit (-Q), and 26 is a two-dimensional inverse discrete cosine transform unit (2D). -IDCT), 27 is a second data format conversion unit for converting television data back into color palette data again, and 28a and 28b are lossless for performing the same lossless compression as in the first conventional example and the second conventional example. (LOSS
LESS compression unit, 29 is a selection unit (multiplexer: M
UX), 30 is an address conversion table for storing address conversion data, 31 is a data synthesizing unit for synthesizing the address translation data of the address translation table 30 and the output signal from the selecting unit 29, and 32 is a data synthesizing unit. A switching unit for comparing the data amount of the data (combined data) with the data amount of the color palette data in the character table 21 and switching the input of the selection unit 29 based on the comparison result, 33 is the data combining unit 31. It is a combined data storage file for storing combined data.

The character table 21 stores character data in a color palette format as quantized color data having a small amount of data.

The first data system conversion unit 22 converts the color palette data in the character table 21 into the YSQ three-dimensional color space axis system of NTSC system which is a general system of color television broadcasting in Japan. The data is converted to. Here, the YIQ three-dimensional color space axis is a color space axis composed of a luminance signal Y and color difference signals I and Q of 8-bit data, respectively, and the luminance signal Y is brightness (luminance), The color difference signal I mainly means orange to cyan hues and saturations, and Q mainly means magenta to yellow and green hues and saturations, and the luminosity characteristics of three primary color (red, green, blue) signals are shown. Assuming R, G, and B respectively, Y = 0.30R + 0.59G + 0.11B I = 0.74 (R−Y) −0.27 (B−Y) Q = 0.48 (R−Y) +0. It can be represented by 41 (BY). The luminance signal Y and the color difference signals I and Q are 8-bit data.

The discrete cosine transform unit 23 includes the first cosine transform unit 23.
Each component (orthogonal definition) of the vector of the YIQ three-dimensional color space axis system data given from the data system conversion unit 22 of FIG. It is expressed by a function and has a function of performing a matrix operation by using a high speed algorithm such as a fast Fourier transform.

The quantizing section 24 approximates the discrete cosine transform section 23 or the value of compressed data described later at a discrete level, and performs a kind of waveform approximating operation.

The inverse quantizing unit 25 includes the quantizing unit 24.
Inverse quantization is performed again on the waveform approximation information quantized in (1) to assign a binary code (code assignment).

The inverse discrete cosine transform unit 26 uses the inverse matrix of the discrete cosine transform unit 23 to transform the code given from the inverse quantization unit 25 into a YIQ three-dimensional color space axis.

The selection unit 29 includes the data synthesizing unit 31.
Which of the two lossless compression units 28a and 28b is used as the address conversion data to be given to is selected.

The address conversion table 30 stores address conversion data indicating which portion of the compressed data as a variable length code has a desired character (that is, becomes an overhead of the character data). .

The data synthesizing unit 31 synthesizes the compressed data as a variable length code after the address translation data from the address translation table 30 is used as an overhead.

The switching section 32 initializes the equal distribution circuit 29 so as to select one of the lossless compression sections (first compression section) 28a directly connected to the character table 21 at the initial stage. 40, a data judging means 41 for judging whether or not the data amount of the data (combined data) combined by the data combining section 31 is larger than the data amount of the color palette data, and the data judging means 41. An input switching unit 42 for switching the input of the selection unit 29 when it is determined that the data amount of the combined data is larger than the data amount of the color palette data.

Further, in FIG. 2, reference numeral 51 is a reading section for reading the composite data stored in the composite data storage file 33, 52 is a data expansion section for expanding the data in the read section 51, and 53 is a data expansion section 52. Is a graphic controller that receives the expanded data and performs graphic processing, 54 is an address conversion unit, and 55 is a CPU.

<Operation> The operation of the character data compression processing apparatus having the above configuration will be described. First, in the character compression circuit shown in FIG. 1, first, the color palette data in the character table 21 is directly sent to the lossless compression unit 28a and compressed according to a predetermined algorithm (first compression step). Then, the compressed data (compressed data) is transmitted to the data synthesizing unit 31 through the selecting unit 29. The data synthesizer 31 uses the address conversion table 3
The address conversion data from 0 and the compressed data from the selecting unit 29 are combined (first combining step). Here, the data amount of the combined data (combined data) and the data before compression, that is, the amount of color palette data in the character table 21 are compared and judged (data judgment process), and the data amount of the combined data is determined. However, if it is smaller than the amount of the original color palette data, the compressed data at this time is directly adopted and stored in the composite data storage file 33 (first storing step: A in FIG. 1).

On the other hand, when the data amount of the combined data becomes larger than the data amount of the color palette data before compression, the data judging means 41 of the switching section 32 judges that fact, and based on the judgment result, the input switching means. At 42, the input of the selection unit 29 is switched. Then, the character compression circuit will select the route of B in FIG. 1 (second storage step). In the route B, first, the color palette data (character data) in the character table 21 is converted into the NTSC YIQ (8 bits each) three-dimensional color space axis system data by the first data system conversion unit 22. Yes (first conversion step). The discrete cosine transform unit 23 performs a two-dimensional discrete cosine transform on such data, and transforms it from the time axis to the spatial frequency axis to concentrate the power distribution of the data (second transformation step).
Then, the quantizer 24 quantizes the data (third conversion step). In the quantization operation in the third conversion step, considering the visual characteristics, low frequency spatial frequency components are not quantized so much, and high frequency spatial frequency components are largely quantized. Further, the flow up to this point is reversed, the inverse quantization unit 25 performs inverse quantization (fourth conversion process), the inverse discrete cosine conversion unit 26 performs two-dimensional inverse discrete cosine conversion (fifth conversion process), and Second data method converter 2
YIQ-color palette conversion (sixth conversion step) is performed. After that, the lossless compression unit 28b (second compression unit) again performs lossless compression on the color palette data (second compression process). Flow B
The operation of is always executed in parallel with the flow A. Then, the compressed data obtained in the second compression process and the address conversion data corresponding to the compressed data are combined by the data combining unit 31 (second combining process), and combined by the second combining process. The combined data is output to the combined data storage file 33 (output step). Here, since the flow B and the flow A are always executed by parallel processing, the time from selection by the selection unit 29 to storage of any compressed data in the composite data storage file 33 should be minimized. You can

With this processing, since the color palette data can be provided with redundancy with almost no deterioration in visual characteristics, the compression rate of lossless compression can be increased, and even if address conversion data is included. It is possible to suppress the amount of data before compression to be equal to or less than that.

Next, the expansion operation will be described. When the color palette data is requested from the CPU 55, the graphic controller 53 generates a character address for the color palette data. Address conversion unit 54
Calls the conversion address of the compressed data for this character address from the reading unit 51. The address of the compressed data is calculated based on the conversion address in the address conversion unit 54, and the compressed data is sequentially read. Then, the compressed data read by the data decompression unit 52 is decompressed,
It is sent to the graphic controller 53.

As described above, according to the present embodiment, the data amount of the compressed data stored in the composite data storage file 33 can be made smaller than the data amount before compression even if the address conversion data is included. As described above, the character expansion / contraction circuit has a simple structure that does not require a separate system table or FIFO, and the access speed before compression can be reliably guaranteed. In this case, the addition of the redundant component of the data using the visual characteristics can be suppressed as much as possible by finally including the address conversion data so that the data amount does not become larger than that before the compression so that the deterioration of the image can be suppressed as much as possible. .

{Modification} (1) As shown in FIG. 1, in the first embodiment, the color palette data is converted into NTSC by the first data system conversion unit 22.
The data was converted into data of the YIQ three-dimensional color space axis system of the system, but other systems such as PAL system or SECA
You may convert into the RY of an M system color, and BY (RGB) system.

[0032]

According to the first and second aspects of the present invention, when the data amount of the compressed data obtained by directly compressing the character data in the color palette format is larger than the data amount of the original character data, the color space method is adopted. , The discrete cosine transform, and the quantization are performed, and then the inverse quantization, the inverse discrete cosine transform, and the conversion to the color palette format are performed to make the data redundant, and then compressed by a predetermined logical method. Therefore, even if the address conversion data is included, the data amount is always smaller than the data amount before compression. Therefore, as shown in FIG. 2, the character decompression circuit does not require a separate system table or FIFO as in the first conventional example (see FIG. 3) and the second conventional example (see FIG. 4), and is simple as a whole. It becomes a composition. Moreover, there is an effect that the access speed before compression can be surely improved.

[Brief description of drawings]

FIG. 1 is a diagram showing a character compression circuit of a character data compression processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a character expansion / contraction circuit of the character data compression processing apparatus according to the first embodiment of the present invention.

FIG. 3 is a diagram showing a configuration of a compression / expansion system in a first conventional TV game.

FIG. 4 is a diagram showing a configuration of a compression / expansion system in a second conventional TV game.

[Explanation of symbols]

 21 character table 22 first data system conversion unit 23 discrete cosine conversion unit 24 quantization unit 25 dequantization unit 26 inverse discrete cosine conversion unit 27 second data system conversion unit 28a first compression unit 28b second compression Part 29 Selection part 29 Equal distribution circuit 30 Address conversion table 31 Data combining part 32 Switching part 33 Composite data storage file 40 Initial setting means 41 Data judging means 42 Input switching means 51 Reading part 52 Data expansion part 53 Graphic controller 54 Address conversion part 55 CPU

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location G09G 5/06 9377-5H H03M 7/30 A 9382-5K H04N 1/41 B

Claims (2)

[Claims] 1. A character table for storing character data in a color palette format, a first data system conversion unit for converting the character data in the character table into data of a predetermined color space system, and the first data system conversion unit. An axis conversion unit that performs a discrete cosine transform on the data of the predetermined color space method provided from the data system conversion unit to remove spatial correlation and concentrate energy in a part of the coefficients; And a dequantization unit that dequantizes the information quantized by the quantization unit and performs code assignment, and a dequantization unit that dequantizes the information. An inverse axis conversion unit for converting the data assigned the code into the data of the predetermined color space method, and the data of the predetermined color space method converted by the inverse axis conversion unit is again subjected to color parsing. A second data method conversion section for converting the data given from the character table in a predetermined logical method, and a second data method conversion section A second compression unit that compresses data according to the predetermined logical method; a selection unit that selects one of the compressed data from the first compression unit and the compressed data from the second compression unit; A switching unit for instructing switching of selection of units, an address conversion table for storing address conversion data, and a data synthesizing unit for synthesizing the address conversion data of the address conversion table and one of the compressed data selected by the selecting unit. And a combined data storage file for storing the data combined by the data combining unit, and the switching unit selects the first compression unit at an initial stage. Initializing means for initializing the equal distribution circuit, and data determining means for determining whether or not the data amount of the synthesized data synthesized by the data synthesizing unit is larger than the quantized data amount of the character data. And switching the input of the selection unit so as to select the second compression unit when the data determination unit determines that the data amount of the combined data is larger than the data amount of the quantized character data. Character data compression processing device comprising input switching means for performing the above. 2. A first compression step of compressing character data in a color palette format by a predetermined logical method, and the quantized character data in a predetermined color space method in parallel with the first compression step. First to convert to data
And a second conversion step of performing discrete cosine transformation on the data of the predetermined color space method to remove spatial correlation and concentrate energy in some coefficients, and in the second conversion step A third transform step for quantizing the discrete cosine transformed data, a fourth transform step for dequantizing the data quantized in the third transform step, and an inverse step for the fourth transform step. A fifth conversion step of converting the quantized data into the data of the predetermined color space method, and the data converted into the data of the predetermined color space method in the fifth conversion step into the color palette format. A sixth conversion step for conversion, a second compression step for compressing the data converted in the sixth conversion step by the predetermined logical method, and compressed data obtained in the compression step and corresponding thereto A A first combining step of combining the dress-converted data, and determining whether the data amount of the combined data combined in the first combining step is larger than the quantized data amount of the character data. And a first data-storing step for storing the synthesized data in a predetermined file when it is determined that the data amount of the synthesized data is smaller than the quantized data amount of the character data in the data determination process. A storage step, and the compressed data obtained in the second compression step when it is determined in the data determination step that the data amount of the composite data is larger than the data amount of the quantized character data, and A second combining step of combining the corresponding address conversion data, and the combined data combined in the second combining step into the combined data combined in the first combining step. Instead, a second storage step of storing the character data in the predetermined file, the method for compressing character data.