JPH09325749A - Device and method for generating character pattern - Google Patents

Abstract

PROBLEM TO BE SOLVED: To decrease a frequency of access to an intermediate memory and shorten time needed to draw a character pattern, and generate the character pattern fast by omitting redundant outline information and writing the character pattern outline when drawing the character pattern outline. SOLUTION: Font data are read out of a font ROM 41, stroke by stroke, and an outline scaling circuit 42 restores and scales control point information wherein the outline of the character pattern is described. An outline drawing circuit 43 draws the outline of the character pattern in the intermediate memory 44. An outline drawing circuit 43 draws the outline of the character pattern in the intermediate memory 44. At this time, horizontal line strokes are drawn by using pieces of code information indicating the start and end of a painting- out process. A painting-out circuit 45 performs the painting-out process by decreasing access to the intermediate memory 44 by referring to difference values of W numbers of the horizontal strokes held in HRM 46.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for forming a character pattern defined by a line figure, which is used in a document creating apparatus such as a Japanese word processor, and which paints a contour drawn based on the line figure. Non-zero winding number as a crushing method
The present invention relates to a character pattern forming device and a character pattern forming method using an ing number) method.

[0002]

2. Description of the Related Art Conventionally, a document creating apparatus such as a Japanese word processor is provided with a character pattern forming apparatus for forming a character pattern in order to display characters on a display or print with a printer. A character pattern forming apparatus generally employs an outline font method capable of generating a high quality character pattern. The outline font method basically prepares character contour information for each character pattern, develops line figure information on a bitmap memory based on this character contour information, and fills this line figure to form a character pattern. To do.

1. Method of Forming Character Pattern When a character pattern is decomposed into a plurality of stroke parts which are allowed to overlap each other and stored in a font ROM and a character pattern is generated using this font ROM, (1) First, the outline of each stroke part of the character pattern is drawn in an intermediate memory (a bitmap memory corresponding to the character pattern, or a software-like array that stores the corresponding contents), and after drawing all stroke parts, ( 2) Read the contour information from the intermediate memory, fill it, generate a character pattern, and output it.
Such a method is widely known.

As a method of filling the area (pattern) defined by the line figure in this way, an even-odd method and a non-zero winding number (non-zero winding number).
er) method is well known.

The even-odd method scans a bit map memory in which a line figure is drawn in a predetermined direction (filling direction) and sequentially paints, for example, from an "odd number" point to a next "even number" point. It is a crushing method. In the even-odd method, as shown in FIG. 17B, there is a problem that when the contour line data is filled, the overlapping figure cannot be filled.

On the other hand, the non-zero winding number system is a widely known system adopted in PostScript, a page description language developed by Adobe Systems, Inc. in the United States, and as shown in FIG. Is a method of filling the left side with respect to the traveling direction of (for example, "+1" is added to the winding number, "-1" is added to the winding number in the area surrounded by the clockwise closed curve, and as a result, the winding number is Fill areas that have a non-zero value). This non-zero winding number system is disclosed in, for example, Japanese Unexamined Patent Publication No.
As disclosed in Japanese Patent No. 136890, a contour line is developed dot by dot on a bitmap memory.

Here, the bit map memory is such that one dot (one pixel) has a one-to-one storage area, and a plurality of bits corresponding to one dot are prepared. In this case, the information stored in the bitmap memory is the winding number difference information (hereinafter, referred to as W number difference information) which is the combination of the vector direction information of the contour line and the contour line overlap count information. At the time of filling, the previous bitmap memory is scanned from the scanning line direction, the previous information stored therein for each dot is obtained, and the information is accumulated to obtain the W number. This W number is not zero (non-ze
ro) Fill the dot positions in the section. With this method, it is possible to paint overlapping figures as shown in FIG. 17 (c).

Character pattern drawing methods using such a non-zero winding number method (NZWN method) include, for example, JP-A-2-266480 and JP-A-02-270019. Japanese Patent Application No. 5-135794 discloses a method of using an array as an intermediate memory instead of a bitmap memory.

In these methods, the position of the point where the W number changes and the amount of change when the character pattern is scanned in the filling direction are stored in the intermediate memory. Japanese Patent Laid-Open No. 2-2
In the case of Japanese Patent No. 66480, a WNPM (winding number pattern memory) having a plurality of bits at an address corresponding to each dot of a character pattern is used as an intermediate memory, and a point where the W number changes (that is, a character number of the character pattern is changed). The position of part of the points that make up the contour) is WNPM
The address of WNPM is W
The change amount of the number is stored.

In the case of Japanese Patent Application No. 5-135794,
Stores the coordinate value in the filling direction and the change amount of the W number 1
It has an array of a dimensional array as an intermediate memory. This one-dimensional array is prepared by the number of dots in the direction perpendicular to the filling of the character pattern, and among the position information of the points, the components in the direction perpendicular to the filling correspond to the identification information of this one-dimensional array.

FIG. 2 illustrates a conventional filling technique in the case of using WNPM as an intermediate memory, taking the character of "earth" as an example. However, the overall size is reduced to 20 dots × 20 dots for the sake of simplicity, but the horizontal line width is set to 3 as a general operation example.
The dot width is set. In this example, the filling is performed in the Y-axis direction (from top to bottom). This is moving the head with the dot printing mechanism lined up in the vertical direction in the horizontal direction,
This is because the whole is optimized by matching the character pattern data transfer direction of the OA device printer that prints the character pattern.

The letter "Sat" consists of three stroke parts, a horizontal bar of the first stroke, a vertical bar of the second stroke, and a horizontal bar of the third stroke. FIG. 2 shows a state in which they are sequentially drawn and the outline of the character pattern (different from the actual outline) is stored in WNPM. In Japanese Patent Laid-Open No. 2-266480, the content of WNPM (code information indicating the W number of each dot) is represented by "-3" to "+3" and "±" according to the maximum number of overlaps (3 times). As expected, in this example,
Since only three values "+1", "-1", and "±" appear, they are represented by "+", "-", and "±".

The "+" mark indicates that the W number of the point of interest is larger by 1 than the W number of the point immediately before it (immediately before the filling point, which is above the point of interest here). Is. Therefore, when the attention point is "+", only 1 is added as the change amount of the W number.

The "-" mark indicates that the W number of the point of interest is smaller by 1 than the W number of the point immediately before it. Therefore, when the point of interest is "-", 1 is subtracted as the change amount of the W number.

The mark "±" is a point where "+" and "-" overlap, and there is no difference in the W numbers immediately before and after that point, but this point should be printed in black. Shown (JP-A-2-
(See page 7 of 266480).

Also, the left and right contours of the vertical bar of the second stroke of the letter "Sat" are shown by "●", but this is actually W.
Not drawn in NPM. The "●" mark is a mark that is added to the character pattern to make it easier to see, and is actually in the same state as a dot with no mark.
For example, the left side contour of the vertical bar is filled from the "+" mark at the point (7,0) to the "-" mark at the point (7,17),
By filling in from the "+" mark at the point (8,0) to the "-" mark at the point (8,17), it is natural to compare it with the left unfilled white background. Appears as a contour in. That is, the contours along the fill direction need not be stored in WNPM.

Further, the overlapping portion of the horizontal and vertical bars, for example, the point (8,16) becomes "+2" when the W numbers are added up by the filling process, but this point is also a character pattern. A black dot is formed like the point of +1 ".

2. Facsimile Image Coding Method A conventional technique similar to the embodiment of the present invention described later will be described.

In a facsimile, the image to be transmitted is scanned in the horizontal direction, and white or black dots are regarded as columns. In the point sequence, the positions of the points that have changed from white to black or from black to white are sequentially transmitted, so that the receiving side can restore the same image as the transmitting side. As one of the methods of expressing the position of the change point, there is a vertical mode coding method.

This is because, as shown in FIG. 16, the pattern of the line immediately before the scan line currently being encoded (referred to as a reference line) is stored on both the transmitting side and the receiving side.
When the change point of the reference line is a reference point, the change point of the coding line is represented by a code "V" when it is directly below the reference point, and by a code "VR (1)" when it is one dot right from the reference point. Codes from VL (3) to VR (3) can be used.

It is known that there are many cases in which the code can be made shorter than when the position of the point is expressed as a numeric value as it is or when the run length code is used.

[0022]

In the character pattern drawing method using the non-zero winding number method as described above, conventionally, as shown in FIG. 2, the dots forming the upper side of the horizontal line and the contour line on the output side are formed. Had to be stored in intermediate memory. It should be noted that it is not necessary to store the vertical line portion of the outline of the horizontal line stroke in the intermediate memory. For example,
This is the point (19, 17) in FIG.

Generally, of the time required for the entire character pattern drawing process, the ratio of the time required for drawing the outline of the character pattern in the intermediate memory is large, and the ratio is 1/2.
There are many cases that exceed. An important factor that influences the contour drawing time is the number of accesses to the intermediate memory.

Referring to FIG. 2, "+" or "-"
In many cases, the code information of the W number indicating the change value of is uniformly written at consecutive positions having the same Y coordinate value.
For example, the points are (3, 7) and (4, 7) to (15, 7). Therefore, it can be seen that a large number of memory cycles are consumed to write the redundant information in the intermediate memory.

The present invention has been made in view of the above points, and draws the outline of a character pattern without writing redundant information in the intermediate memory, thereby reducing the number of accesses to the intermediate memory and drawing the character pattern. An object of the present invention is to provide a character pattern forming apparatus and a character pattern forming method capable of forming a character pattern at high speed by shortening the time required for.

In particular, in order to solve the above-mentioned object, the present invention provides a method in which a character pattern can be reduced until simple horizontal line segments overlap each other, and a stroke pattern does not need to be divided into simple horizontal line segments and others. The purpose is to do.

[0027]

[Means for Solving the Problems]

(1) The character pattern forming apparatus according to the present invention performs the filling process in the first coordinate axis direction of the character pattern based on the font data in which the character pattern is divided into a plurality of parts and described. Is the right angle second
Is a character pattern forming device for forming the character pattern by repeating in the direction of the coordinate axis, and storing means for storing control information indicating the start and end of the filling processing in the direction of the first coordinate axis; A generator having a function having the same function as the stored control information is generated each time the filling process is repeated in the direction of the second coordinate axis, and the filling is performed based on the signal obtained by the generator. A character pattern forming device characterized by performing processing.

With such a configuration, the number of times information for filling processing is written in and read from the intermediate memory (storage means) can be greatly reduced, and the processing speed can be increased. In addition, the storage capacity of the intermediate memory can be reduced.

(2) Further, in the character pattern forming method of the present invention, the filling processing in the first coordinate axis direction of the character pattern is performed based on the font data in which the character pattern is divided into a plurality of parts and described. A character pattern forming method for forming the character pattern by repeating in the second coordinate axis direction, which is a direction perpendicular to the first coordinate axis direction, and showing the start and end of the filling processing in the first coordinate axis direction. The control information is stored in a memory, and a signal having a function equivalent to that of the control information stored in the memory is generated by repeating the filling process in the second coordinate axis direction. The feature is that the filling processing is performed based on the signal.

According to such a method, when the processing capability of the microprocessor system is sufficiently high with respect to the required character pattern forming speed, the cost can be reduced by implementing all the functions by software.

[0031]

Embodiments of the present invention will be described below with reference to the drawings.

First, the principle of the present invention will be described with reference to FIG. FIG. 1 is a conceptual diagram for explaining the filling method of the present invention corresponding to FIG. 2 (conventional). It should be noted that, in order to preferably implement the present invention, it is preferable to use a method using an array as a base depending on a method using WNPM as an intermediate memory (described later). But for the sake of simplicity,
First, the principle of the present invention will be described based on the method using WNPM.

In FIG. 1, although nothing is written in the WNPM (intermediate memory), the point that is the outline of the character pattern is
It is indicated by the marks "●" and "○". The mark "●" has the same meaning as in FIG. That is, in practice WNP
It is a point that is not drawn on M, and is a mark added to make the character pattern easy to see in this figure.

The mark "○" is indicated by "+" or "-" in FIG.
This means that nothing is written in this embodiment. That is, it indicates that the access to the WNPM is reduced by the number of “◯” marks. In the example of FIG. 1, “+”,
Since the number of “◯” marks is larger than the number of “−” and “±” marks, it is shown that the number of WNPM accesses is reduced to half or less by implementing the present invention. This means that the speed of character pattern drawing is greatly increased.

Further, in FIG. 1, "+ s" and "-s"
Mark has the same function as "+" and "-" in FIG.
It shows that the effect is repeatedly effective to the right. That is, the difference value of the same W number is repeatedly generated at the same Y-axis position in the right direction of the X-axis from the position where “+ s” and “−s” marks are present. The repetition thus designated continues until the "+ e" and "-e" marks appear at the same Y-axis position.

In this case, "+ s" is + in the W number.
Code information indicating the start position of the component, "+ e"
Is code information indicating the end position. Similarly,
“-S” is code information indicating the start position of the − component in the W number, and “-e” is code information indicating the end position. Here, it is assumed that the filling direction is the Y-axis direction (from the top to the bottom), and the filling is performed one column at a time from the left. By repeating the difference value of the W numbers designated by "+ s" and "-s" up to "+ e" and "-e" for each filling process of each column, the points marked with "○" naturally. It becomes the outline of the character pattern.

A block diagram of a character pattern forming apparatus to which the present invention is applied is shown in FIG. In FIG. 4, an HRM (horizontal repeat memory) 46 is a memory for implementing the present invention.

The font ROM 41 stores character patterns after data compression. Contour scaling circuit 42
The control point information describing the contour line of the character pattern is restored, and the coordinate value and line width information are scaled so that the character size becomes a predetermined character size. The contour drawing circuit 43 draws the contour of the character pattern in the intermediate memory 44. Filling circuit 45
Is to fill each row of the character pattern from top to bottom based on the contents set in the intermediate memory 44.
The output image memory 47 stores the filled character pattern as a finally output character pattern.

Here, HRM (horizontal pattern memory) 4
Reference numeral 6 holds the W number difference value of the horizontal stroke included in the column immediately before (immediately left) the column being filled. Therefore, the HRM 46 has the same number of words (the number of bytes) as the length (the number of dots) of the output character pattern in the filling process direction. Each word is at least "-3"
It is necessary to have the number of bits (that is, 3 pits or more) enough to represent eight values of "+3" and "± 0".

In such a structure, the font RO
The font data of the designated character (character contour information for forming the line figure) is read from M41, and the control point information describing the contour line of the character pattern is restored by the contour scaling circuit 42 to obtain a predetermined character size. The coordinate values and line width information are scaled so that Then, in the contour drawing circuit 43, the contour of the character pattern is stored in the intermediate memory 44 along with the code information of "+" and "-", and "+ s", "-s", "+ e" for the horizontal line trooke. , "-
Drawing is performed using the code information of e ″ (control information indicating the start and end of the filling process).

After the information of all the strokes forming the character is stored in the intermediate memory 44, the next step is the filling circuit 45.
At, the contours of the character patterns set in the intermediate memory 44 are filled from the top to the bottom for each column, and the final character patterns are output to the output image memory 47. At that time, the difference value of the W number of the horizontal stroke held by the HRM 46 (“+ s”, “−s”, “+” in FIG. 1).
e "and" -e ")
The filling process is performed by generating a signal indicating the same difference value every time the filling process is repeated. As a result, it is possible to reduce the access to the intermediate memory 44 and perform the filling. The filling and the character pattern output are repeated from the leftmost column to the right.

FIG. 3 shows the HRM (horizontal repeat memory) corresponding to FIG. 1 and the manner in which its contents change as the filling process progresses. The HRM in FIG. 3 is the same as the length in the filling direction (from top to bottom) of the character pattern in FIG.
Has 20 dots. The details will be described later. At the start of filling, all the contents are cleared to 0 (herein, white is represented by 0 and black is represented by 1) (not shown).

FIG. 3A shows the state of the HRM when the filling process has proceeded to column 1 in FIG. Figure 1 for HRM
The W number difference value +1 at the position of the Y coordinate 16 designated by “+ s” at the point (0, 16) is stored. The value of this HRM is read in synchronization with the intermediate memory, and both are input for W number filling. That is, in the example of column 1, the W number difference value of +1 is read from the HRM at the Y coordinate value 16 and the Y coordinate value 17 is read as (1, 1 in the intermediate memory).
"-" Of 7) is read. As a result, the dots from the Y coordinate 16 to the Y coordinate 17 become black.

W read from the intermediate memory as usual
There is no distinction between the difference value of the number and the difference value read from the HRM, and the WN difference value is sequentially added from y = 0 to y = 19 as usual, and the points where the accumulated value is not zero (the accumulated value is zero) It is the same as the conventional method to make black) (including points changed from other than zero).

After performing the filling processing in the seventh row, the twelfth row
Figure 3 shows the contents of the HRM before starting to fill the columns.
(B). “+ S” at point (2, 7), (4,
9) point “-s”, (0, 16) point “+ s”, (7,
It can be seen that the W number difference values indicated by "-s" at point 17) and "-s" at point (2,18) are stored at the respective Y coordinate positions.

The contents of the HRM during the filling processing in the 18th column are shown in FIG. 3 (c). The "-" at the (12,17) point in FIG.
By e +, “+ e” at the (15,7) point and the (15,16) point, the W number difference value stored at the corresponding Y coordinate is canceled, and the position of y = 18 It can be seen that only the difference value remains.In this way, the character pattern of "Sat" is created by the action of the intermediate memory and the HRM having the contents as shown in FIG.

In FIG. 1, from the (7,17) point to (1
The part up to the 2,17) point is expressed as a pair of "-s" and "-e", but the part from the (7,0) point to the (9,0) point is "+ s". Instead of expressing it as a pair of "+ e",
The reason why it is expressed by the conventional simple "+" connection will be described below.

The portion from the (7,17) point to the (12,17) point (the end cut-off and decoration of the vertical bar stroke) is the outline that is a horizontal line in the font data, and is therefore drawn using HRM. It is expressed as a pair of "-s" and "-e". On the other hand, the part from the (7,0) point to the (9,0) point is originally a curve, and when the points representing the curve happen to be placed on the dot grid, they are lined up on the same horizontal line. And, because it is, it is rendered as usual.

Further, points (0, 16) and (2, 7) in FIG.
The points are stored together with the "+ s" code and the "-" code. This is the reason why it is difficult to carry out the present invention based on the filling method using WNPM as described above.

That is, in the case of the conventional WNPM, since a simple W number difference value is stored therein, it is necessary to store two or more W number difference values at the same position. In addition, the sum of those difference values should be added and stored as one difference value. However, the difference value with the s flag and the e flag of the present invention cannot be directly added to the conventional simple difference value. Therefore, in the following, an example of implementing the present invention on the basis of a method using an array for the intermediate memory will be described.

FIG. 5 shows a hardware block diagram as an embodiment of the present invention. In FIG. 5, reference numeral 51 is a CPU that performs overall control. A main memory 52 stores various information. A font ROM 54 stores font data in a compressed form. Reference numeral 56 denotes an HRM for storing the horizontal repeating pattern, which is a very small memory. 57 is a bus.

Reference numeral 53 is a painting circuit for performing painting processing of the character pattern. This filling circuit 53
This circuit is similar to the conventional circuit for realizing the filling with the non-zero winding number, and the details will be described later. There are various conventional methods for realizing the filling circuit depending on the target performance, cost, and the like. Since the present invention relates to the principle, only the dot serial processing type, which is the simplest implementation method, will be described. The above-mentioned intermediate memory (array) 55 used for the filling processing is stored in the main memory 52.

FIG. 6 shows the structure of the intermediate memory realized as an array. The data format of each element of the array in this case is shown in FIG. The correspondence between the arrangement of FIG. 6 and the conceptual diagram of FIG.
This is exactly the same as the case where the conventional non-zero winding number filling is realized by an array (for example, Japanese Patent Application No. 5-135794).

That is, a one-dimensional array having the same number as the number of dots of the target character size is provided in the direction (horizontal direction in the example of the drawing) perpendicular to the direction of the filling process. In each one-dimensional array, the coordinate values in the filling direction of the outline of the character pattern and the W number difference value at that point are sequentially combined and stored. The state of the arrangement of FIG. 6 corresponds to that of FIG. 2 at point (0,16) in Figure 1
Although one code is assigned, it is realized by assigning two elements to the array of column 0 in FIG. 6 and storing the same Y-axis coordinate value 16 together.

FIG. 7 is a diagram showing the data format of array elements. When an array is used as the intermediate memory, a Y coordinate table corresponding to the X coordinate of each point is prepared individually. In this case, one point is represented by 16-bit data as shown in FIG. Of these, the lower 0 to 9 bits are the portions that have been conventionally used, and the remaining 10 to 16 bits are empty. This time, 2 of the vacant parts
Bits are used for the s flag and e flag.

When R = 00, it is exactly the same as the conventional one. In other words, bit 7-0 has the Y coordinate value, and bit 9
It has a difference value of W number at -8. D = + indicates that, including that point, the W number of each point after that point (below in this example) is incremented by 1 from the W number of the immediately preceding point. D
= -In the same way indicates that it is decremented by 1. D = ± is W
The number does not change and indicates that the dot should be black. D = ± indicates that D = + and D = − overlap in the same dot. Handle D = ± in this way so that the extremely thin black line does not disappear. This is as has been done conventionally.

Here, the code of R = s and R = e, and D =
The invention is realized by a combination with + or D =-.
This is "+ s", "+ e", "-s", "-e" in FIG.
Corresponds to the code information expressed as. R = s indicates that the W number difference value should be repeated thereafter (ie to the right) at the same Y coordinate value. R =
The e indicates that the iteration should end at that column.

In this embodiment, R = s or R =
Do not use e in combination with D = ±. If a horizontal line with a width of 1 dot frequently appears, D = ± and R = s
Alternatively, if a code combining R = e is also used, the amount of data stored in the array is reduced. In the present embodiment, for example, D =
Two codes of +, R = s and D =-, R = s are set to the same Y
Such a case is realized by storing the coordinate values.

Next, FIG. 8 shows the configuration of the HRM realized as an array in the main memory. The HRM is a one-dimensional array having the same size as the number of dots in the filling direction of the target character pattern. Here, each element of HRM has 6 bits, and it is used by dividing it into fields HC and RW of 3 bits each.

The field RW corresponds to FIG. 3 and stores the W number difference value rw repeated in the horizontal direction. The range of the value of rw is "-3" to "+3". Field HC
Stores the number hc of horizontal contour lines of the character pattern at that position. The range of the value of hc is “0” to “7” here. This indicates that the characters can be reduced until the seven horizontal contours are gathered at the same dot position.

The reason why the HC is provided is to prevent the horizontal line from disappearing even if the horizontal line becomes extremely thin and has a width of 1 dot. Details thereof will be described later. This will be described with reference to FIG.

The state of the HRM of FIG. 8 corresponds to the state of FIG. 3 (b). In the HRM, the element at the corresponding Y coordinate position is updated as the filling process proceeds downward from y = 0. The updating method is shown in FIG.

FIG. 9 is a diagram showing an updated state of the HRM.
The code on the left side of this figure shows the code (combination of R and D in FIG. 7) read from the intermediate memory corresponding to the dot position during the filling process. hc and rw are H
The contents of C and RW are shown.

In FIG. 5, the font ROM 54
In the step of restoring the contour of the character pattern while sequentially reading and writing it in the array 55 of the intermediate memory (the step of the program processing corresponding to 42 in FIG. 4), the arrays are stored in the order of appearance. After the information of the entire character pattern is stored in the array 55, the contents of the array 55 are sorted by the Y coordinate value for each column. At the stage of the filling process (corresponding to 43 in FIG. 4), all the dots having the same Y coordinate value as the dot being filled are read from the column of the array 55 and processed. This is the same as the conventional one and is described in detail in Japanese Patent Application No. 5-135794.

As shown in FIG. 9, the W number difference value to be repeatedly applied to each column in the right direction is stored in RW according to the contents of the array of the intermediate memory. The content of HC stores the total number of difference codes repeated at that position, that is, the number of horizontal contour lines, regardless of the difference value of the W number. The updated contents of the HRM are written to the same position of the HRM when the corresponding dot filling process is completed.

Next, the filling circuit (53 in FIG. 5) will be described.

First, a conventional filling circuit will be described with reference to FIG. The W number difference value read from the intermediate memory is added to the accumulator 101, and the W number at the next dot position is output from the accumulator 101. The output of accumulator 101 is applied to zero decision circuit 102. The zero determination circuit 102 sends a logical value “1” to the OR circuit 103 when the W number is other than zero. Y that is the same as the Y coordinate value of the dot
When the coordinate value is in the data read from the intermediate memory, a logic value "1" is sent from a circuit (not shown). When the output of the OR circuit 103 becomes “1”,
That is, when the W number has a value other than zero or when the difference value change information of the dot position is in the array, the filling result of the corresponding dot is made black.

The signal "with data" has a function of blackening the dots given the code of "±". Since the accumulator used as an embodiment is of a type in which the accumulation result is obtained at the timing corresponding to the next dot, this “with data” also has the function of blackening the dot at the filling start point. There is.

FIG. 11 shows the filling circuit of the present invention. Basically, it is a circuit for calculating the W number of each dot by adding the W number difference value stored in the HRM to the conventional circuit of FIG. The contents of the decoder 111 are shown in FIG.

That is, the W number difference value stored without the R flag or with the R = s flag in the array becomes the decoder output. The same value as the conventional simple W number difference value is added to the W number difference value rw read from the HRM, and becomes the output of the adding circuit 112.

Zero decision circuit 114 after accumulator 113,
The operation of the OR circuit 115 is almost the same as that of the conventional example, but here, a zero determination circuit 116 is added to make the output dot black even when the value of hc of HRM is not zero.

The purpose of the zero decision circuit 116 is to correctly output a horizontal line having a width of 1 dot. That is, in FIG. 11, rw represents the W number difference value in the original non-zero winding method, while
hc indicates the number of "s", that is, the number of horizontal contour lines.
In the W number, it becomes "0" when "+" and "-" overlap. At that time, the W number is "0", but if there is a contour, it must be black. There is hc to see it. If you don't consider overlapping, you don't need such a circuit.

More specifically, for example, "soil" in FIG.
Consider the case where the line width of the first stroke becomes thin and both the upper contour and the lower contour become y = 8 and overlap. Since there is no data in the columns 5 and 6 and the W number is also zero, the horizontal bar will be cut after all without the zero determination circuit 116 of hc. Here, since hc at column 5 and column 6 at y = 8 becomes 2, this part becomes
Printed correctly with black dots. In addition, since hc is also returned in the columns 18 and 19, the dot of y = 8 is correctly a white dot.

Next, the above-mentioned R is added to the array of the intermediate memory.
A method for writing the W number difference value with a flag will be described.

As a method therefor, it is simple to store it as a part of the font data in the font ROM of FIG. This is self-explanatory and requires no explanation.
The contents of the font ROM remain unchanged, and the contour restoration / scaling portion (contour scaling circuit 4 in FIG.
In many cases, it is desirable to be able to respond by changing 2).
This is because, in general, the font ROM is often used in common for various products.

Next, a case where the present invention is applied to a core font will be described. Here, as an example, a case where the present invention is applied to a conventional example in which a character stroke is described based on the shape of the core line M and the width W from the core line M to the stroke contour as shown in FIG. explain.

The basic data structure of conventional font data is shown in FIG. 13 (b). A stroke is expressed as a connection of segments, and FIG. 13B shows the data structure of one segment. A simple stroke such as the horizontal bar in FIG. 13A is expressed as one segment. The segment type field has a portion indicating horizontal line, vertical line, diagonal line, or curved line. In this data structure, in addition to the fields showing the core line shape and line width, there are fields showing clipping and decoration. Core wire M and wire width W
End point of the stroke contour line determined from
The positions of the four points of Crs, Cre, Cls, and Cle shown in (a) are indicated by clipping fields, and the shape of the portion beyond that is indicated by decoration fields.

Conventionally, in the contour restoring unit, two straight lines from Crs to Cre and Cls to Cle and two decorative contours from Cls to Cres and Cre to Cle are converted and the target character is converted. It was scaled to the size and passed to the contour drawing section.

Here, the basic idea of applying the present invention to a core font will be described. As a similar method,
There is a method of dividing a stroke pattern into a rectangular portion (horizontal line pattern) surrounded by a contour parallel to the filling direction and a contour orthogonal to the filling direction and a portion (a method of storing the horizontal line pattern), This method does not require such division of strokes.

In this method, it is obvious that the contour line from Cls to Cle is a horizontal line, so that simply "+ s" code information is provided at a position corresponding to Cls and "+ e" is provided at a position corresponding to Cle. Write the code information of. In this case, it is not necessary to write anything in the position corresponding to each dot of the contour line between them. This greatly reduces the number of times the array is accessed, and therefore allows faster processing. Similarly, the code information of "-s" is written in the position corresponding to Crs, the code information of "-e" is written in the position corresponding to Cre, and what is written in the position corresponding to each dot of the contour line in the meantime. Also do not write.

The process of storing the contour information of the horizontal bar stroke body in the array is completed in this way, regardless of the character size, by writing the array four times. The contours from Cls to Crs and from Cre to Cle, that is, the contours of the decorative portion of the horizontal bar stroke are processed in the same manner as in the conventional case. Vertical bar strokes and curved strokes are processed in the same manner as in the past. In the present embodiment, the horizontal line segment forming the outline of the decorative portion is also stored in the array using the difference code with the R flag (portions (7, 17) to (12, 17) in FIG. 1).

FIG. 14 shows a flowchart of the entire character pattern forming process according to the embodiment of the present invention. Here, the process of reading the character font data from the ROM and expanding the data is collectively performed (steps A11 and A1).
2), reading and extension may be performed for each stroke as shown by the dotted line in FIG.

When the compressed font data is expanded, the outline of the stroke is restored, the coordinate value is scaled to the target character size, and then the outline is written in the intermediate memory (step A13). At that time, as described above, drawing is performed using the code information of “+” and “−” and the code information of “+ s”, “-s”, “+ e”, and “-e” for the horizontal stroke. To do.

When the processing for all strokes forming the character pattern is completed (step A14), the information for one column in the intermediate memory (array) is sorted by the coordinate value (step A15), and the filling circuit as hardware is set. It is activated and the array pointer is initialized at that time (step A1).
6). Then, the elements pointed to by the array pointer are sequentially sent to the filling circuit to perform the filling (steps A17 and A18). This is repeated for all columns (step A19). In this case, if "+ s" or "-s" is found, then "+ e" or "-e"
Until a appears, a signal indicating the difference value of the same W number is generated each time the filling process is repeated, so that it is possible to reduce the access to the intermediate memory and perform the filling.

In the above embodiment, the filling processing circuit is realized by hardware and used, but the filling processing itself can be realized by software. In this case, all the processing in FIG. 11 may be realized by software and combined with the processing in FIG.

Further, instead of dividing the HRM into HC and RW as shown in FIG. 9, it may be configured as a counter for separately counting the number of + and the number of-. The structure at this time is shown in FIG. In this case, an adder for calculating "pc-mc" is added to FIG. 11, and the output is put in the adder circuit instead of rw. As an alternative to the signal of hc ≠ 0, (pc ≠
0) V (mc ≠ 0) is input to the OR circuit.

As described above, in the method of the present invention, assuming that the filling direction of the character pattern is the Y-axis direction (from top to bottom), the control information ("" indicating the start and end of the filling process in the Y-coordinate axis direction is used. + S "," -s "," + "
e "," -e ") is stored in the intermediate memory, a signal having a function equivalent to this control information is generated every time the filling process is repeated in the X coordinate axis direction, and based on the generated signal. By doing so, the number of times of writing and reading information for the filling process to the intermediate memory can be significantly reduced, the processing can be speeded up, and the storage capacity as the intermediate memory can be reduced. can do.

Further, by using the method of the present invention, even for a reduced character in which simple horizontal line segments overlap each other, for example, the overlapping portion can be correctly filled to form a high quality character pattern. Also,
Unlike the method of storing horizontal line patterns, it is not necessary to divide the stroke pattern into simple horizontal line segments and others.

As shown in FIG. 5, in a character pattern forming apparatus that includes a microprocessor and a main memory and connects them by a bus, a part of the main memory is used for the above-mentioned "+ s", "-s", "+ e". By using as a storage area for code information indicating ","-e "and using a bus for writing / reading the code information, the following effects can be obtained.

In a microprocessor system, the processing time is often determined mainly by the number of bus accesses, so the number of bus accesses can be greatly reduced.

Further, when used in a printer having a print head for printing a plurality of dots in parallel, the following effects can be obtained by making the filling processing directions coincide with each other in the dots of the print head. can get.

Buffering is facilitated. The effect of the present invention is particularly great when the filling direction matches the upright (or upside down) kanji. The reason is that there are many horizontal lines in Chinese characters, and there is almost no processing to write and read each dot of the upper contour and the lower contour in the intermediate memory (regardless of the stroke length, the stroke body part does not matter. 4 times).

Further, the above-mentioned code is provided by using a memory (HRM) having a capacity equal to or larger than the number of dots in the first coordinate axis direction (filling direction) of the character pattern, without using a bus. The following effects are obtained by repeatedly generating a signal having the same function as information in the second coordinate axis direction.

Since the intermediate memory must have a certain number of bytes of the total number of dots included in the character pattern, it is placed in the main memory, and the HRM has a capacity for one column of the character pattern. If you can update without going through the bus, the performance will be higher and the cost will not be too high. If the HRM is also placed in the main memory, the cost will be reduced, but the number of bus accesses will be significantly increased. While the intermediate memory can access only the elements actually used, the HRM needs to write and read the total number of dots of the character pattern regardless of the character pattern.

The HRM has two fields (HC and RW, or PC and MC) and is updated according to the contents of the intermediate memory (array) by different updating methods, so that the following effects can be obtained. .

Even if the horizontal line is narrowed to 1 dot width, it can be printed correctly without disappearing as a character pattern. H
This cannot be achieved unless the RM has multiple fields.

[0097]

As described above, according to the present invention, the control information ("+ s", "-s", "+ e", indicating the start and end of the filling process of the character pattern in the first coordinate axis direction,
"-E") is stored in the intermediate memory, and a signal having a function equivalent to this control information is stored in the second portion of the filling processing.
Since it is generated every time the coordinate axis is repeated and the filling process is performed based on the generated signal, the number of times of writing and reading the information for the filling process to the intermediate memory can be significantly reduced, Can be speeded up. In addition, the storage capacity of the intermediate memory can be reduced.

Further, by using the method of the present invention, even for a reduced character in which simple horizontal line segments overlap each other, for example, the overlapping portion can be correctly filled to form a high quality character pattern. , It is not necessary to divide the stroke pattern into simple horizontal line segments and others.

[Brief description of drawings]

FIG. 1 is a conceptual diagram for explaining a filling method of the present invention.

FIG. 2 is a conceptual diagram for explaining a conventional filling method corresponding to FIG.

FIG. 3 is a diagram showing how the HRM corresponding to FIG. 1 and its contents change as the filling process progresses.

FIG. 4 is a block diagram showing the configuration of a character pattern forming device to which the present invention has been applied.

FIG. 5 is a block diagram showing a hardware configuration as an embodiment of the present invention.

FIG. 6 is a diagram showing a configuration of an intermediate memory realized as an array.

FIG. 7 is a diagram showing a data format of array elements in an intermediate memory.

FIG. 8 is a diagram showing the configuration of an HRM implemented as an array in the main memory.

FIG. 9 is a diagram showing an updated state of HRM.

FIG. 10 is a block diagram showing a configuration of a conventional filling circuit.

FIG. 11 is a block diagram showing the configuration of a filling circuit according to the present invention.

12 is a diagram showing the contents of a decoder in FIG.

FIG. 13 is a diagram showing an expression of a horizontal bar stroke as a basic data structure of a core font.

FIG. 14 is a flowchart of the entire character pattern forming process according to the embodiment of the present invention.

FIG. 15 is a diagram showing a configuration of another embodiment of the HRM.

FIG. 16 is a conceptual diagram for explaining a facsimile encoding method.

FIG. 17 is a conceptual diagram for explaining a conventional filling method.

[Explanation of symbols]

 41 ... Font ROM 42 ... Contour scaling circuit 43 ... Contour drawing circuit 44 ... Intermediate memory 45 ... Filling circuit 46 ... HRM (horizontal repeating memory) 47 ... Output image memory 111 ... Accumulator 112 ... Addition circuit 113 ... Accumulator 114 ... Zero decision circuit 115 ... OR circuit 116 ... Zero decision circuit

Claims (2)

[Claims] 1. A method of filling a character pattern in a direction of a first coordinate axis based on font data in which a character pattern is divided into a plurality of parts and described in a direction perpendicular to the direction of the first coordinate axis. A character pattern forming device for forming the character pattern by repeating in the direction of the coordinate axis No. 2, and storage means for storing control information indicating the start and end of the filling processing in the direction of the first coordinate axis, and the storage means. Generating means for generating a signal having a function equivalent to that of the control information stored in, every time the filling process is repeated in the second coordinate axis direction, and based on the signal obtained by the generating means. A character pattern forming device characterized by performing a filling process. 2. The filling process in the first coordinate axis direction of the character pattern is performed in a direction perpendicular to the first coordinate axis direction based on font data in which the character pattern is divided into a plurality of parts and described. 2 is a character pattern forming method for forming the character pattern by repeating in the direction of the coordinate axis, wherein control information indicating the start and end of the filling processing in the direction of the first coordinate axis is stored in a memory, A signal having a function equivalent to the control information stored in is generated every time the filling process is repeated in the second coordinate axis direction, so that the filling process is performed based on the generated signal. Character pattern forming method.