JPH0511733A - Character generating device - Google Patents

Abstract

PURPOSE:To reduce memory as well as to shorten processing time to facilitate LSI formation by making line width compensation simultaneously with affine transformation or curve interpolation. CONSTITUTION:While character font data 1 of a vector form are transformed coordinately or curve-interpolated in a control part 2 for line width compensation by affine transformation as well as the coordinate transformations such as affine transformation, curve interpolation, etc., the line width of a specified linear line is compensated to a specified line width. Then a flag data showing the content of compensation is added to the coordinate data to be modified. It is desirable to compensate the line width by modifying the coordinate data according to the flag data. Also it is desirable to modify the line width in vertical direction only relative to the vector facing toward the coordinate data to be processed next.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character generation method and a character generation method for enlarging or reducing contour data of a character represented in a vector format and converting the contour data into a raster format data for output. More particularly, the present invention relates to a character generation method and a character generation device for enlarging or reducing character contour data by affine transformation and then converting the contour data into raster format data for output.

[0002]

2. Description of the Related Art In recent years, as a method of outputting characters of a plurality of character sizes, the coordinates on the outline of the character data have been stored in a ROM or an external storage device as a character font, and enlarged or reduced according to a specified output size and output. There is something to do. In this method, since a calculation error becomes large in the reduction calculation for outputting a relatively small character, it is necessary to correct the line widths in the vertical and horizontal directions in order to prevent deterioration of character quality after the reduction calculation.

FIG. 12 is a block diagram of a conventional character generator.

Reference numeral 1 denotes vector format character font data, 6 is a block for executing coordinate conversion such as affine transformation or curve interpolation, and 7 is all coordinate transformation for one character. After that, a line width correction block for performing line width correction, 3 is a block for generating a line or curve, and 4 is a block for filling the inside of the contour character generated in 3.

In the figure, numeral 5 represents a raster type character pattern developed into dots.

Japanese Unexamined Patent Publication No. 62-272295 describes a method of enlarging / reducing a character pattern having a line width control function.

[0007]

In the above-mentioned conventional line width control method, as shown in FIG. 12, the line width correction is performed after all the coordinate conversion for one character is completed, so that it takes a long processing time. There was a drawback. Further, in order to perform the line width correction, it is necessary to extract the data necessary for the line width correction from the affine transformed data and hold it. Therefore, a storage device such as a RAM having a relatively large capacity for storing this data is required. Therefore, the conventional line width correction method has a line width correction function, and after performing the coordinate conversion by the affine transformation on the contour data of the character represented in the vector format, the character data represented in the raster format. It has been difficult to make a character generator for converting into a LSI into an LSI.

An object of the present invention is to provide a character generation method suitable for LSI implementation.

Another object of the present invention is to provide an LSI character generator having a line width correction function.

Another object of the present invention is to provide a line width correction system and a device for the same which can perform line width correction at high speed with a small circuit scale which can be integrated into an LSI.

Other objects of the invention will be understood from the following description.

[0012]

The outline of the representative one of the inventions disclosed in the present application will be briefly described as follows.

That is, the line width of the designated straight line is corrected to the designated line width while the coordinate transformation is performed by the affine transformation or the curve interpolation is performed. When correcting the line width of the specified straight line to the specified line width while performing coordinate conversion by affine transformation or performing curve interpolation, the coordinate data to be corrected includes flag data indicating the contents of the correction. Is added, and the coordinate data is corrected according to the correction flag data, so that the line width of the designated straight line is corrected to the designated line width.

Further, the coordinate data to be corrected is to be corrected only in the direction perpendicular to the direction of the vector toward the coordinate data to be processed next.

Further, the coordinate data to be corrected is 1
If the coordinate data that was previously corrected was corrected,
The coordinate values of the corrected coordinate data are inherited.

In each group of outline characters made up of a set of coordinate points forming the outline font data, the start point and the end point are made up of the same coordinate data,
The processed coordinate data is output in order from the second coordinate data, and finally the second coordinate data is output.

Alternatively, in each loop of outline characters formed by a set of coordinate points forming the outline font data, the start point and the end point are formed of the same coordinate data, and the permutation of the first and subsequent coordinate data is In the case of a series of data that cannot be separated, the output order of the processed coordinate data is output from the coordinate data where it can be separated, and finally, the second and subsequent series of coordinate data is output. It was done like this.

[0018]

According to the above-mentioned means, while performing coordinate transformation by affine transformation or while performing curve interpolation,
Since the line width of the specified straight line is corrected to the specified line width, the data necessary for the line width correction is extracted from the data per character that has been affine transformed or curvedly interpolated,
The line width can be corrected without using a large-scale storage device such as a RAM for storing this data. Therefore, it is possible to reduce the scale of the entire character generator having the line width correction function.

Further, in the line width correction for correcting the line width of the designated straight line to the designated line width, the correction flag data indicating the content of the correction is added to the coordinate data to be corrected,
Since the coordinate data is corrected in accordance with the correction flag data, 1 that has been affine transformed or curvedly interpolated
Since processing such as extraction of data necessary for line width correction from data for each character is not required, high-speed line width correction can be realized.

The coordinate data to be corrected is corrected only in the direction perpendicular to the vector direction toward the coordinate data to be processed next. For example, in one set of coordinate data (x ₁ , y ₁ ) (x ₂ , y ₂ ) representing one straight line, from the coordinate point represented by _one coordinate data (x ₁ , y ₁ ) to the other coordinate data When the vector pointing to the coordinate point represented by (x ₂ , y ₂ ) is along the x-axis, only the y coordinate value of the _one coordinate data (x ₁ , y ₁ ) is corrected. On the other hand, when the vector is along the Y-axis, the coordinate data of one side (x ₁ , y ₁ )
Is modified only in the x coordinate value.

Further, the coordinate data to be corrected is as follows, when the previously processed coordinate data is corrected.
If the coordinate value of the corrected coordinate data is inherited, that is, if the previous x coordinate value is corrected, the x coordinate value is replaced as the x coordinate value of the coordinate data to be corrected. Therefore, the coordinate data (x ₁ ,
Since only _{one of} y ₁ ,), (x ₂ , y ₂ ) needs to be corrected, the registers etc. required for line width correction are 1 /
The number can be reduced to 2, and the scale of the entire apparatus can be reduced.

Also, in each loop of outline characters made up of a set of coordinate points forming the outline font data, the start point and the end point are made up of the same coordinate data, and the output order of the processed coordinate data is , The second coordinate data is output in order, and finally the second coordinate data is output.

Further, in the case of a series of data in which the permutation of the first and subsequent coordinate data cannot be separated, for example, in the case of data representing a curve, the output order of the processed coordinate data can be separated. It is possible to output a character of good quality by outputting the coordinate data where it is possible and finally outputting the second and subsequent series of coordinate data.

[0024]

EXAMPLES Examples of the present invention will be described in detail below.

FIG. 1 is a block diagram of a character generator to which the present invention is applied. Note that this device is not limited to this, and can be used as an LSI as a microcomputer peripheral device.
It may be a single element that has been realized, or a systemized device.

Reference numeral 1 represents vector-type character font data, and 2 is a control unit for realizing line width correction at the same time as coordinate conversion such as affine transformation and curve interpolation.
Is a line or curve generation unit, and 4 is a block for filling the inside of the contour character generated in 3. Further, in the figure, reference numeral 5 represents a raster-type character pattern expanded into dots.

FIG. 5 shows a character generation LSI to which the character generation device of the present invention is applied (shown as an outline font processor OFP in the same figure. Hereinafter, it will be simply referred to as OF.
An example of a system configuration using (also referred to as P) is shown. The system bus 331 is the address bus 33-
1, a data bus 33-2, and a control bus 33-3. The system bus 33 includes a host CPU 26, a system memory 27, an outline font memory 28,
The DMA controller 29, OFP 25, and page memory controller 30 are connected. A page memory controller 30 is connected to the OFP, and a page memory is connected to the page memory controller. The page memory 31 is connected to a printer or CRT system 32.

This system is mainly used for a page printer controller and a CRT system. For example, OF
P25 is DMA by a signal from the host CPU 26.
The coordinate data, which is outline font data, is received from the outline font memory 28 via C29, converted into bitmap data, and written into the page memory 31 via the page memory controller 30.

FIG. 4 shows an internal block diagram of the OFP. Although not particularly limited, each circuit block surrounded by a solid line in the figure is formed on one semiconductor chip by a well-known semiconductor integrated circuit technology.
Here, the OFP 25 includes a host interface 23 and a font input fifo FiFo 15 inside the host interface, a block 16 including an affine transformation block and a line width correction block, and a fifo FiFo 0.
(20), curve interpolation block 17, fifo FIFO
1 (21), line generation block 18, dot memory 2
2. Fill bit block transfer BitBLT
It is composed of a block 19 and a page memory interface 24. From the host interface 23, the address buses (A1 to A) connected to the system bus 33 are connected.
6), the data bus (D0 to D31), and the control signal group are input / output. The internal bus data 14 is connected to the system data bus. Further, from the page memory interface 24, an address bus (PA1 to PA24) and a data bus (P
D0 to PD15) and control signal groups are input and output.

In this way, the OFP 25 in this embodiment is
Is divided into three groups, that is, a host interface, a page memory interface, and a block group for character generation processing. In particular, the block group for character generation processing is divided into four blocks: a coordinate conversion function, a curve interpolation function, a line generation function, a fill and a data transfer function,
The line width correction function is included in the coordinate transformation block or the curve interpolation block.

Further, in this embodiment, the FiFo 20 installed between the affine transformation block 16 and the curve interpolation block 17, the FiFo 21 installed between the curve interpolation block 17 and the line generation block 18, and the dot memory 22. , And the internal data bus 14 and the like, four functional blocks operate independently as described later, and pipeline processing is possible.

Further, each of the blocks 16, 17, 18 and 19 has a plurality of registers (not shown), and these registers are supplied through the internal address signal lines AI1 to AI6. Selected by the internal address signal. Internal data buses DI0 to DI15
The data supplied via the register is written to the register selected as described above, for example. A control signal is also supplied from the host interface 23 to each block. In the figure, WR1 is a Fifo FiF.
o indicates a write ready signal, and RP1 indicates a read ready signal of fifo FiFo.

In order to facilitate understanding of the present invention, first,
The characters and the data representing them will be described below.

One character is shown in FIG. As can be seen from this figure, a character can be represented by multiple closed lines. In the figure, this closed line is
It is expressed as That is, for example, the letter "day" or the numeral "day" is represented by three closed lines, loop 1, loop 2, and loop 3. When expressing characters in vector format, for example, if you specify that the left side of the vector be filled,
The loop 1 is represented by a counterclockwise vector,
Loops 2 and 3 are represented by clockwise vectors. As a result, the space between the loop 1 and the loop 2, the space between the loop 2 and the loop 3 and the space between the loop 1 and the loop 3 are filled. The character "day" (hereinafter, also including the number "day") is drawn. Of course, it may be specified that the right side of the vector is filled. However, in this case, the loop 1 is clockwise and the loop 2 is
And loop 3 are represented by a counterclockwise vector.

Since it becomes a loop, for example, if it is assumed that the coordinate data of (x _O , y _O ) starts, the coordinate data (x ₄ ,
y ₄ ) becomes the same as the initial coordinate data (x _O , y _O ).

The character "day" shown in FIG. 6 is represented in the coordinate data according to the present invention as shown in FIG. In FIG. 7, “8001 16” indicates the loop start flab, and “8004 16” indicates the end flab of one character. In this example, as described above,
It is specified that the left side is filled with respect to the vector direction.
Therefore, the coordinate data (x _O , y _O ), (x ₁ ,
In y ₁₎ between the vector coordinates de - comes at the data (x _O, y _O), and from the (x _1, y _1). This vector is
As can be seen from FIG. 6, it is a vector along the X axis. Therefore, the coordinates can be corrected in the Y-axis direction. Therefore, a correction flag (line width correction flag) is added to the coordinate data (x _O , y _O ) generated by this vector. However, since this vector serves as a reference when correcting the line width, no correction flag indicating correction or replacement is added,
A correction flag indicating a standard is added.

Since the vector originating from the coordinate data (x ₁₃ , y ₁₃ ) is also along the X axis, the coordinate data can be corrected in the Y axis direction. In this embodiment, this vector is corrected during the line width correction, so that the coordinate data
A correction flag indicating correction is added to the data (x ₁₃ , y ₁₃ ).

Further, since the vector from the coordinate data (x ₁ , y ₁ ) to the coordinate data (x ₂ , y ₂ ) is along the Y axis, the line width correction is performed in the X axis direction. Can be modified to Therefore, it is possible to add a correction flag to the coordinate data (x ₁ , y ₁ ). However, in the present embodiment, the line width correction in the X-axis direction is not shown in order to facilitate the description, and thus the correction flag is not added. Of course, a correction flag may be added to correct the line width in the X-axis direction. In this case as well, the criteria are:
A correction flag indicating that it is a reference is added, and a correction flag indicating correction or replacement is added to what is corrected. Next, the operation of the OFP will be described. The coordinate data read from the outline font memory 28 is
It is temporarily stored in the font input FiFo15 in the host interface. The affine transformation block sequentially reads the coordinate data (x, y) from the font input FiFo15 via the signal line 100 and executes the affine transformation (the following equation).

[0039]

## EQU1 ## x '= ax + by + e y' = cx + dy + f where a, b, c and d are affine matrix data constants and e and f are constants. These constants are determined based on the data regarding the amount of expansion or reduction supplied from outside the OFP 25 for expansion or reduction, and are set in a register (not shown) in the affine transformation block. To be done. In addition, x ′ and y ′ represent coordinate data that has been transformed by affine transformation, that is, enlarged or reduced. In this specification, the coordinates are expressed in a rectangular coordinate system.

The operation of the affine transformation block 16 includes an affine transformation mode and a transfer mode. In the transfer mode, coordinate data is sent to the next stage without affine transformation. On the other hand, in the affine transformation mode, the affine transformation operation as shown in FIG. 15 is performed. The affine transformation operation will be described below with reference to FIGS. 4 and 15.

The font input FiFo15 stores flag data (including a line width correction flag) in addition to the coordinate data described above. As will be described later with reference to FIG. 2C, the flag data is given an extended identifier for identifying it. The data supplied from the font input FiFo15 is based on this extended identifier,
A determination is made as to whether it is coordinate data or flag data (step 51). The affine transformation shown in the above equation is performed on the data determined to be the coordinate data (step 52).

Affine transformed coordinate data (x ',
It is determined whether y ') is the first coordinate data of the coordinate data loop forming one character (step 53).
This determination is made by, for example, a loop start flag “8001 1
This can be executed by determining whether or not the coordinate data immediately after 6 "(see FIGS. 6 and 7).

When it is determined that the coordinate data is the first coordinate data, the line address calculation is performed by the affine transformation block 16.
Done in. Here, the line address calculation means that when the line generation block 18 receives the coordinate data and generates a line by a dot image between the coordinate points, each group of the coordinate data forming one character. Is a calculation for obtaining an address in the dot memory 22 corresponding to the first coordinate data. That is, it is a calculation for obtaining an address in the dot memory 22 corresponding to the first coordinate data. Since the address corresponding to the first coordinate data is found for the second and subsequent coordinate data, the address (dot memory 22) corresponding to the second and subsequent coordinate data is used as the displacement for this address. Address)
Can be asked. Therefore, the line generation block 18 can handle the second and subsequent addresses by managing the addresses.

In the present embodiment, the affine transformation block 16 corrects the coordinate data (x ', y') designated by the correction flag in order to execute the line width correction. Modified coordinate data (x ", y")
Is output to FiFo20. This line width correction function will be described in detail later. Further, as described above, the line width correction function may be included in the curve interpolation block 17.

The curve interpolation block 17 reads the coordinate data (x ', y') or (x ", y") temporarily stored in the FiFo 20, and in the case of the curve data, new coordinate data representing the curve. Is generated and output to the FiFo 21. The curve interpolation block 17 generates interpolation point data that forms a smooth curve using four coordinates by a general method such as a Bezier curve interpolation method.

The line generation block 18 uses the FiFo 21
The coordinate data temporarily stored in is read, a line is generated as a dot image between the coordinate points, the outline of the character is created, and it is written in the dot memory 22 as bitmap data. As a method of generating this line, for example, the well-known Bresenham algorithm is used. Fill
Bit BLT Black reads the bitmap data representing the outline for each character from the dot memory 22, fills the inside of the outline, and passes through the page memory interface 24,
Bit map data is written in the page memory 31 while being block-transferred.

Before the description of the line width correction function, the pipelining of each functional block for the character generation processing shown in FIG. 4 will be described. As is clear from the above description, each functional block 16, 17, 18, 19
For each data transfer, FiFo20,
21 or the dot memory is built in, so that if necessary data is prepared in the FiFo 20, 21, or the dot memory 22, they can operate independently. Here, the signal indicating that the data is prepared is the signal RR1, and since the FiFo 20, 21 and the dot memory have a finite size, it is confirmed that the number of stored data exceeds the size. The indicated signal is the WR
It is 1.

A process of outputting the character string "ABC" by the pipeline process of the OFP 25 will be described with reference to FIG. In FIG. 14, rectangles represent the operating time of each functional block. In order to simplify the explanation,
It is assumed that the time required to process one character takes a unit time T in any functional block. In the conventional character generator, each functional block can operate only after the processing of the preceding functional block is completed.
Took a long time to process. When the character generator of this embodiment is used, the data of the next block is started at the same time as the data required for the next block is processed in the functional block of the previous block. The data required in the next stage is, for example, in the case of the curve interpolation block 17, data for four coordinates expressing one curve, or one word for data other than the curve (flag data and straight line data). Further, if there is 18 between line generation blocks, it is data for the coordinates representing one line. However, the filled and Bit BLT block 19 is
The operation is started after the line generation block 18 finishes processing all the data for one character. Therefore, the string, "A
The time required to output BC "is 4T, and processing performance three times that of the conventional character generator can be obtained.

FIG. 2 shows character fifo data. 2A shows the conventional contour data, and FIG. 2B shows the data to which the correction flag 8 according to the present invention is added. Figure 2
(C) shows a detailed format of the correction flag 8. In order to correct the line width, the correction flag 8 is added to the predetermined coordinate data in advance, as described above. The contents of the correction flag to be added will be described later. However, in order to identify whether or not the data taken into the block 16 via the signal line 100 is a correction flag, the correction flag has an extension identifier indicating that the 12th to 15th bits are the correction flag. Note that is set (see Figures 2c and 4).

FIG. 3 is a detailed block diagram of the line width correction unit 2 shown in FIG. That is, FIG. 3 shows main blocks in the line width correction block of FIG. In FIG. 3, 10 is a number (0 to n: integer) represented by bits 0 to 7 of the correction flag 8 shown in FIG. 2C and coordinate data (x) after affine transformation to which the correction flag is added. ′,
This is a RAM (memory) for storing coordinate value data of either y ′) or coordinate data (x ″, y ″) corrected by line width correction. Reference numeral 11 denotes coordinate data (x ′,
This is a calculation unit for correcting y '). Reference numeral 13 is a designated line width register for setting the line width to an arbitrary value. Although not shown in the figure, the designated line width register 13 is supplied with data from the internal data buses DI0 to DI15 of FIG. The line width is set by this data.

The input side of the RAM (10) is provided with the number of times of selection (not shown), and the data or affine transformation block (or curve interpolation block) from the arithmetic unit 11 is selected according to the contents of the correction flag. RA from the data from 9
Tell M (10).

When the line width correction block is provided after the affine transformation block, RA of 12 in FIG.
M2 corresponds to FiFo0 (20) in FIG. 4, block 9 in FIG. 3 corresponds to an affine transformation block, and block 3 in FIG. 3 corresponds to a curve interpolation block. When the line width correction block is provided after the curve interpolation block, the block 9 corresponds to the curve interpolation block, the block 3 corresponds to the line generation block, and the R
AM ² corresponds to FiFo 1 (21).

3 through FIG. 11 through FIG.
The line width correction method performed in 1 will be described.

FIG. 6 is a diagram schematically showing an outline font composed of coordinate data, taking the character "day" as an example. Actually, it includes curves etc.,
Here, in order to facilitate the explanation, it is assumed that the line is composed of horizontal lines and vertical lines. In this example, it is composed of three horizontal lines and two vertical lines. In order to simplify the explanation, only the horizontal line will be line width corrected. The coordinate data (x _O , y _O ) to (x ₁₄ , y ₁₄ ) are the coordinate data after the affine transformation. Actually, the affine-transformed coordinate data is sequentially supplied to the line width correction unit. Here, the coordinate data after the affine transformation is shown for ease of explanation.

Numbers (0), (1) and (2) are attached to the horizontal lines. A number corresponding to this number is added to the correction flag in FIG. That is, the number of the horizontal line is represented by bits 0 to 7 of the correction flag, which are added to the coordinate data that respectively emit the two vectors in the X-axis direction forming the horizontal line. For example, the coordinate data that emits the vector forming the horizontal line (0) are (x _O , y _O ) and (x ₈ , y ₈ ), and the bits 0 to 7 of the correction flag added to them are ( 0) is set.

A case will be described below in which line width correction is performed so that all of these horizontal lines have a line width of 2 bits.

As a line width correction method, as shown in FIG. 9A, the correction of the coordinate point (xn, yn) is performed by modifying the next coordinate point (x
It is assumed that only the vertical direction can be performed toward (n + 1, yn + 1) (with respect to the vector direction). In addition, FIG.
As shown in (b), the corrected coordinate point (xn, yn)
When the processing shifts to the next coordinate point (xn + 1, yn + 1), the corrected coordinate value is inherited. That is,
When yn is corrected to yn ″ at the coordinate point (xn, yn), yn + of the next coordinate point (xn + 1, yn + 1)
1 becomes yn ″ which is inherited from yn ″ as it is.

Here, there is a correction of the coordinate point and a replacement of storing the corrected coordinate value and replacing it with the coordinate value stored when the new coordinate point is corrected. As shown in FIG. 11, when a horizontal line or a vertical line is composed of a plurality of coordinate points, correction points and replacement points occur. That is,
In the first correction, the correction is performed, and in the next correction, the replacement is performed using the coordinate values used for the correction. This is because the same correction process is not repeated.

FIG. 13 shows a flowchart of the line width correction by the above method.

In order to execute the line width correction according to the above method, the outline font data shown in FIG. 6 is looped to the line width correction font data, that is, the data to which the correction flag 8 is added. Process in the order of 1, loop 2 and loop 3. In the coordinate data forming the horizontal line, for example, in the horizontal line (0), (xo, yo), (x ₁ , y ₁ ),
(X ₄ , y ₄ ), (x ₅ , y ₅ ), (x ₈ , y ₈ ), (x ₉ ,
y ₉ ), and the data that appears first when the data string is formed as shown in FIG. 7 is the reference point.

That is, the coordinate data (xo, yo) is set as the reference point, and the line segment to the next data (x ₁ , y ₁ ) is set as the reference line (0).

When the direction of the vector of this line segment is compared with the coordinate axis parallel to it, if the direction of the vector is the increasing direction of the coordinate value (positive direction), the coordinate data (xo, yo)
The bit 11 of the correction flag 8 added to is set to "0". On the other hand, if the direction of the vector is the decreasing direction of the coordinate value (negative direction), the coordinate data (xo, yo)
The bit 11 of the correction flag 8 of 1 is set to "1".
The correction of the correction point is performed by adding or subtracting the designated line width value (for example, the value set in the designated line width register 13) to the coordinate value of the reference line.
Whether to add or subtract is uniquely determined by the vector direction of the reference line. When the left side of the vector direction is filled and the line width is corrected only in the y-axis direction as in the present embodiment, when the vector direction of the reference line is the positive direction, the Y of the reference line is Y. The correction point can be obtained by adding the line width value to the coordinate value. On the other hand, when the vector direction of the reference line is a negative direction, the correction value is obtained by subtracting the line width value from the y coordinate value of the reference line.

For example, since the correction point (x ₁₃ , y ₁₃ ) in FIG. 6 is above the reference line (2) (in the direction in which the Y coordinate value increases), the line width corresponds to the y coordinate value Y6 of the reference line (2). The corrected point Y13 ″ is obtained by adding the values. At this time, the vector direction of the reference line (2) is in the positive direction with respect to the coordinate axes.

On the other hand, since the correction point (x ₁₁ , y ₁₁ ) is below the reference line (1) (in the direction in which the y coordinate value decreases), the line width value is changed from the y coordinate value Y2 of the reference line (1). The corrected point can be obtained by subtracting. The vector direction of the reference line (1) at this time is a negative direction with respect to the coordinate axes.

As described above, the correction point can be corrected by applying the line width value to the coordinate value of the reference line and the operation (addition or subtraction) determined by the vector direction of the reference line.

The coordinate data (x ₁ , y ₁ ) is the source of the vector directed in the y-axis direction. Therefore, the correction in the x-axis direction is possible. That is, since the coordinate data (x ₁ , y ₁ ) is corrected in the x-axis direction, it has no attribute in the case of horizontal line width correction only. In other words, the correction flag may not be added. (X ₂ , y ₂ ) becomes the reference point of the horizontal line (1).

(X ₃ , y ₃ ) has no attribute.
(X ₄ , y ₄ ) is the same as the reference point (xo, yo),
Moreover, since the reference point is a fixed point, it has no attribute.

Next, regarding the same horizontal line (0), the coordinate point which is the data that appears next in the data row of FIG. 7 and can be corrected in the y-axis direction is the correction point. In this example, de - data (x _8, y ₈₎ are de appeared this second - a motor, a fix point. By the way, since (x ₅ , y ₅ ) is a point corrected in the X-axis direction, the second appearing data is
It's not. Further, although not appearing in this example, the data appearing after the third becomes the replacement point.

Similarly, looking at all the coordinate data, (x ₂ , y ₂ ) becomes the reference point for the horizontal line (1), and (x ₁₁ , y ₁₁ ) is corrected. It becomes a point. For the horizontal line (2), (x ₆ , y ₆ ) becomes the reference point, and (x _6
13 and y ₁₃ ) are the correction points.

Although the above-described line width correction is made only for horizontal lines, the line widths of horizontal and vertical lines can be corrected at the same time by adding attributes to the coordinate data for vertical lines as well. . At this time, one coordinate data has only one attribute of horizontal line width correction and vertical line width correction.

FIG. 7 shows a font data structure for line width correction, which is added to the coordinate data having the attribute by setting the attribute in the correction flag with respect to the coordinate data shown in FIG. As described above, “8001 16” is the loop start flag,
“8004 16” is a one-character end flag. The line width correction is executed according to the correction flag while sequentially reading the line width correction font data.

FIG. 8 shows coordinate data stored in the RAM 1 (10) shown in FIG. Here, the correction flag number (represented by bits 0 to 7) is
It is used as an address of the RAM 1 and the coordinate data to which the correction flag is added is stored at the address indicated by the correction flag number. FIG. 8A shows (xo, y
It is a state of the RAM 1 at the time of processing from (o) to (x ₆ , y ₆ ), and the coordinate value of the reference point corresponding to the number is written. This is processed based on that the contents of the 8th and 9th bits of the correction flag are "00" (reference point).
That is, the 8th and 9th bits of the correction flag indicate that the coordinate data to which these correction flags are added is the reference point. Therefore, the coordinate values of yo, y ₂ and y ₆ remain unchanged by the selection circuit (not shown).
It is written in the addresses (0), (1) and (2) of M1.

[0073] FIG. 8 (b), (x _8, y ₈₎ is state of the RAM1 of the time of the treatment, the 8,9-bit contents of the correction flag assigned to the (x _8, y ₈₎ "01" (correction point) 10-bit content is "1" (correction / replacement of Y coordinate)
Based on the content of 11 bits being "0" (the reference line direction is positive), yo at the address (0) is rewritten to y ₈ ". That is, this coordinate data is 8 and 9 bits of the correction flag. Is the correction point, and at the 10th bit, y
It is indicated that the coordinates are corrected / replaced, and the 11th bit indicates that the reference line is positive. Therefore, the line width correction value is added to the coordinate value of yo, and the obtained y
₈ ″ is written to the RAM address (0) represented by the address of the correction flag. Here, yo of the reference point is erased, but this is no longer necessary after the correction, and the processing of the replacement point of the same number is performed. This is because sometimes the rewritten y ₈ ″ is used.

FIG. 8C shows the state of the RAM 1 when all the coordinate data have been processed. It can be seen that the capacity of RAM1 is half the total of the correction points for X and Y.
In this example, since the horizontal lines (0) and (1) have a width of 2 bits before the line width is corrected, y ₈ ″ = y ₈ , y
₁₁ ″ = y ₁₁ .

Further, at (x ₁₄ , y ₁₄ ), the pre-correction point (x ₁₃ , y ₁₃ ″) y ₁₃ ″ is inherited, and y ₁₄ → y ₁₄ ″.
= Y ₁₃ ″. In order to carry out this inheritance, although not particularly limited, the arithmetic unit 11 (see FIG. 3) is provided with one X-system register and one Y-system register (not shown). This register holds the modified result if one coordinate data emitting the vector is modified, and the held value is the other coordinate data to which the vector is directed. This is output when the data is modified so that the inheritance as described above can be realized.
The system is provided to enable inheritance in the X-axis direction and inheritance in the Y-axis direction.

FIG. 10 shows the RAM2 (12) of FIG. 3 or the FiFo 0 of FIG. 4 in the line width correction method of the present invention.
(20) or the output order to FiFo 1 (21). FIG. 10A shows a case where line width correction is executed and output in the same order as the input coordinate data P ₁ . In this case, P ₁ may remain as dust. Therefore, as shown in FIG.
If the start of a straight line starts with a straight line, P ₁ is not output and P ₂
Output from the outputs P ₂ at the end, P ₂ is the starting point,
A new end point is P ₆ . In the case where the start of the loop as shown in FIG. 10 (c) is a series of data that cannot be separated by the curve generation control data, P ₁ , P ₂ and P ₃ are not output, It outputs from P ₄ , outputs P ₂ , P ₃ , and P ₄ at the end, P ₄ becomes the starting point, and P ₉ newly becomes the ending point. With the above output method, a high-quality loop can be generated.

[0077]

In the character generator according to the present invention, the line width correction can be performed simultaneously with the affine transformation or the curve interpolation, so that the processing time can be shortened. Further, in the conventional method, all the correction points had to be stacked on the memory, but this is not necessary and the memory can be largely erased.

[Brief description of drawings]

FIG. 1 is a block diagram of a character generator according to an embodiment of the present invention.

FIG. 2 is a diagram showing correction data.

FIG. 3 is a detailed block diagram of a line width correction unit.

FIG. 4 is an internal block diagram of the OFP.

FIG. 5 is a system configuration diagram of an OFP.

FIG. 6 is a diagram for explaining a line width correction method.

FIG. 7 is a diagram showing a structure of line width correction font data.

FIG. 8 is a diagram for explaining a line width correction method.

FIG. 9 is a diagram for explaining a line width correction method.

FIG. 10 is a diagram for explaining a line width correction method.

FIG. 11 is a diagram for explaining a line width correction method.

FIG. 12 is a block diagram of a conventional character generator.

FIG. 13 is a flowchart of line width correction.

FIG. 14 is a timing chart for explaining the operation of the OFP.

FIG. 15 is a flowchart for explaining the operation of the OFP.
FIG.

[Explanation of symbols]

1 is character font data, 2 is a line width correction unit, 8 is a correction flag, 10 is a RAM 1, 11 is a calculation unit, 13 is a designated line width register, 3 is a line or curve generation unit, 4 is a fill unit, 5 Indicates an enlarged / reduced character pattern, and 25 indicates an OFP.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Wakisaka Shinji             292 Yoshida-cho, Totsuka-ku, Kanagawa Co., Ltd.             Hitachi Microelectronics Equipment             In development lab (72) Inventor Hiroshi Wada             292 Yoshida-cho, Totsuka-ku, Kanagawa Co., Ltd.             Hitachi Microelectronics Equipment             In development lab (72) Inventor Hiroyuki Koizumi             5-20-1 Kamimizuhonmachi, Kodaira-shi, Tokyo Stock             Ceremony company Hitachi Ltd. Musashi factory (72) Inventor Susumu Kaneko             5-20-1 Kamimizuhonmachi, Kodaira-shi, Tokyo Stock             Ceremony company Hitachi Ltd. Musashi factory (72) Inventor Fumihiro Takahashi             5-22-1 Kamimizuhonmachi, Kodaira-shi, Tokyo Stock             Inside the Hitachi Microcomputer System (72) Inventor Toshio Asai             5-22-1 Kamimizuhonmachi, Kodaira-shi, Tokyo Stock             Inside the Hitachi Microcomputer System

Claims (6)

[Claims] 1. A contour data of a character represented in a vector format.
Is a character generator that generates a character by performing coordinate conversion by affine transformation and then converting it to character data represented in a raster format. While performing coordinate conversion by affine transformation, or A character generator which corrects the line width of a specified straight line to a specified line width while performing curve interpolation. 2. The correction of the upper air line width is performed by adding correction flag data indicating the contents of the correction to the corrected coordinate data and correcting the coordinate data according to the correction flag. The character generation device according to claim 1, wherein the line width of the designated straight line is corrected to the designated line width. 3. The coordinate data to be modified is modified only in the direction perpendicular to the direction of the vector toward the next processed coordinate data. Item 2. The character generator according to item 2. 4. The coordinate data to be modified is that the coordinate value of the modified coordinate data is inherited when the previously processed coordinate data has been modified. The character generation device according to claim 1, 2 or 3, which is characterized. 5. In each of the groups of outline characters made up of a set of coordinate points constituting the outline font data, the start point and the end point are constituted by the same coordinate data, and the processed coordinate data are processed. The character generator according to claim 1, wherein the data is output in order from the second coordinate data, and finally the second coordinate data is output. 6. In each loop of an outline character made up of a set of coordinate points constituting the outline font data, the start point and the end point are constituted by the same coordinate data, and the first and subsequent ones. In the case of a series of data in which the permutation of data cannot be separated, the output order of the processed coordinate data is such that the coordinate data which can be separated is output first, and finally, The character generating device according to claim 1 or 5, wherein the second and subsequent series of coordinate data are output.