GB2186163A - Display apparatus for Hangul characters - Google Patents

Abstract

Korean (Hangul) characters each comprise elements representing an initial consonant, a vowel and possibly a final consonant. Apparatus for displaying such characters offers six different display shape formats (Fig. 2), one of which is selected according to whether the vowel is 'vertical', 'horizontal' or 'composite' and on whether or not a final consonant is present. Each character element is stored in a 12x10, 24x10 or 24x20 dot matrix and is displayed normal size, double height, double width or half height in combination with at least one other to form a character occupying 24x20 dots of the display. <IMAGE>

Description

SPECIFICATION Data display apparatus This invention relates to data display apparatus of a type for displaying data represented by digital codes, the displayed data being composed of characters the shapes of which are determined by selected dots of a dot matrix which constitutes a character format for the characters.

Such data display apparatus can be used in a television receiver which is adapted to receive either a teletext or a videotex transmission, or both.

In the adapted television receiver there would be provided a decoder which can acquire transmitted digital codes representing selected data, store these digital codes and use them to produce character generating signals for driving the CRT of the television receiver to produce the data display. It is known to use in the decoder a character memory in which is stored character information identifying all the available character shapes which the system can display. This character information is selectively addressed by character generator means in accordance with the stored digital codes to produce the character generating signals for the data display, this selective addressing being effected synchronously with the scanning action of the CRT.

To facilitate this selective addressing, it is convenient to store the character information in a manner in which, for each character shape, the pattern of discrete dots that define the shape in the dot matrix character format are represented by respective data bits which are located spatially in accordance with the character shape in a character memory cell which is composed of a matrix of elemental storage areas. With this form of storage, the dot pattern of a character shape as displayed in a television frame can have a one-to-one correspondence with the stored bit pattern for the character.

The teletext transmissions in the United Kingdom are all in one language, namely English, and the character generator employed for these transmissions produces words using discrete English language characters derived from the acquired digital codes. When teletext transmissions are provided in a country whose national language uses additional characters to those of the English language or discrete characters which differ to English language characters, it is usually a simple matter to provide for these transmissions a character memory for the alphabet of the country concerned.

Such a simple solution, however, does not fit the needs of a country whose language does not conform to a discrete character format for constructing words.

In particular, the authorities in the Republic of Korea have shown great interest in the introduction of a teletext service for that country. A major difficulty, however, is the writing system for the Korean language, which is rather different from anything encountered hitherto in teletext development work. Known as Hangul, it is a syllabic writing system. Each Hangul character represents a syllable, which is composed of elements representing the vowel and associated consonants in a logical but quite complex way.

The elements making up Hangul characters, together with examples of such characters, are shown in Figs. 1 a to 1 n of the accompanying drawings.

There are 21 vowels of which, 10 are single vowels as shown in Fig. 1a, 9 are double vowels (diphthongs) as shown in Fig. 1 b and 2 are triple vowels (triphthongs) as shown in Fig. 1c. There are also 19 initial or final consonants as shown in Figs. id, and a further 11 final consonant or consonant combinations as shown in Fig. 1e. Therefore in theory over 12,000 Hangul characters are possible. In practice about 1500 are though necessary.

From the foregoing it can be seen that Hangul character shapes can be quite complex and a character dot matrix of 12 x 10 dots, which has currently been considered as a standard for teletext and videotex systems, has been found to give insufficient resolution for a legible display.

As mentioned above, Hangul is a phonetic writing system in which each character represents a syllable. A syllable consists of one vowel, preceded by an initial consonant and possibly followed by a final consonant. Each of these elements has a recognisable shape and they are combined into a composite character representing the syllable. Depending on the shape of the vowel, the initial consonant is added to the left of the vowel or above it.

For example, as shown in Fig. if, vowel (ya)+initial consonant (k) gives (kya), and as shown in Fig. 1g, vowel (a)+initial consonant (r) gives (ra).

The vowel can be simple, as in the examples above, or a diphthong or triphthong.

These are composed using two or three of the simple vowels, often giving a 'composite' vowel shape which has a space at the top left for the initial consonant. For example, as shown in Fig. 1h, diphthong (we)+initial consonant (r) gives (rwe), and triphthong (wae)+ initial consonant (h) gives (hwae).

For the consonants, the symbol for (ng) represents an 'absent' consonant used in the same way as the others. So the symbol for (ng+a) represents the syllable with the simple vowel sound 'a'. It is possible to have double consonants for a long or vocalised sound, for example (ss) or (bb), leading to complex character shapes, as shown in Fig. lj, for (sswi) and (bbae). As an added complication, if a final consonant is present the previously composed symbol is reduced in height and the final consonant placed underneath. Examples for (chum), (rur) and (natch) are shown in Fig.

1k.

The final consonant may be doubled as before, or may be certain combinations of two consonants, for example (rm) and (ps), as shown in Fig. 11, giving syllables for (charm) and (bops), as shown in Fig. 1m.

These composite characters representing syllables are grouped together to form words with the usual spaces between. In printing, the space occupied by each character is roughly the same however many elements are used in its construction. There are occasions when a given sound combination as represented by Hangul may be ambiguous in meaning and sometimes Chinese characters or English words are used instead, mixed in with the Hangul characters. However, for most purposes Hangul alone is sufficient, representing as it does the sound of the words as they would be spoken.

The Korean Broadcasting System (KBS) had proposed a complex software based teletext system, but the decoder complexity which is required raises doubts as to its commercial viability in the consumer environment. The system also suffers the usual variable format penalty of lack of ruggedness together with rather inefficient transmission coding.

It is an object of the present invention to provide a data display apparatus which can respond to a simplified transmission coding produce a display in Hangul in a more efficacious manner.

According to the present invention, in data display apparatus of the type set forth above a larger character dot matrix has a plurality of different shape formats one of which is selected for the display of a (Hangul) character, the selection being made in dependence on the form of the vowel of the character.

In carrying out the invention a character matrix of 24 x 20 dots is suggested for displaying a Hangul character, equivalent to a group of 4 conventional (12x 10 dot) characters.

This gives sufficient resolution for the most complex shapes and allows a display format of 20 charactersx 10 rows in the 525 line Korean television system. More information is conveyed on a page than might be expected since fewer characters are required per word compared with the English language.

To provide simple display algorithms for the performance of the invention, the character matrix can be split in half vertically or horizontally, as appropriate, to allow common shapes to be used for the consonants in combination with double height and double width character expansion techniques.

In further considering the nature of the invention reference will now be made by way of example to the accompanying drawings of which: Figures la to 7n show, as aforesaid, the elements making up Hangul characters, together with certain characters; Figure 2 shows diagrammatically six display shape formats for displaying Hangul characters in accordance with the invention; Figure 3 shows the effect of halving the height of a character shape; Figure 4 shows an example of displayed Hangul text using Hangul characters constructed using the display shape formats of Fig. 2; Figure 5 illustrates diagrammtically a form of transmission coding for Hangul characters; Figure 6 shows diagrammatically a display algorithm for the performance of the invention; Figures 7 and 8 show a logic diagram for a Hangul character generator;; Figure 9 shows an enable matrix logic for the Hangul character generator of Figs. 7 and 8; and Figure 10 is a block diagram of a teletext receiver in which the invention can be embodied.

Referring to Fig. 2, the shape formats S1 and S2 use simple halving of the character matrix for vertical and horizontal vowels respectively. The shape format S1 has a vertical initial consonant IC and a vertical vowel VV.

The shape format S2 has a horizontal initial consonant IC and a horizontal vowel HV.

Shape format S3 is reserved for the more complex composite vowels CV (diphthongs and triphthongs) having both horizontal and vertical elements and an initial consonant IC.

The initial consonant shape IC would be used in normal size (12x 10) for shape format S3; double height (12x20) for shape format S1 and double width (24x 10) for shape format S2. These shape formats are used when no final consonant is present.

When a final consonant FC is present, shape formats S4, S5 and S6 are used instead. These use the same techniques as the shape formats S1, S2 and S3, respectively, but the height of the initial consonant IC and vowel VV or HV is halved and the final consonant FC is placed underneath. This is a simple algorithm to implement but it does lead to some asymmetry. For example, in shape format S5 the initial and final consonants IC and FC might be the same, but the initial consonant IC will be half the height of the final consonant FC; e.g., as shown in Fig. 1n, the character (nun) will have a half size upper consonant. This compromise does, however, give an acceptable quality of display. Fig. 3 shows two possible basic versions for the consonant r. Each shape reduces to the same shape when displayed half height by displaying only alternate dot lines. However, when both the basic versions and their respective half sizes are displayed together, it is seen that the lower version gives the more visually pleasing combination. Fig. 3 also shows the full and half size versions of the consonant r and the vowel u formed into the character (rur), but without double width.

It may be feasible to deal with particularly difficult combinations of elements as special cases, but unless they are restricted the number of stored shapes required will multiply rapidly. increasing the cost of the decoder. Alternatively, a different division of the matrix for the shape formats S4, S5 and S6 could be used, but almost certainly this would require two different versions of each element to be stored, increasing the cost again.

An example of Hangul text using the invention is shown in Fig. 4. The character shapes may be improved in consultation with Korean language experts. The dot patterns for the various elements need careful consideration to give an acceptable display for normal, double height, double width and half height versions in combination.

Assuming that these techniques can give an acceptable quality of display, the amount of storage required for dot matrix patterns is drammatically reduced. The invention needs storage for 14 vowels in a 12 x 10 matrix, 7 composite vowels in a 24 x 20 matrix, 14 consonants in a 12x 10 matrix and 16 double consonants or combinations in a 24x 10 matrix, i.e.

(14x 12x 10)+(7x24x20)+(14x 12x 10) +(1 6 x 24 x 10)10,560 bits. This compares with 1500x24x20--720,000 bits to store 1500 likely characters, or 12,369x24x20=5,937,120 bits for all theoretically possible element combinations stored individually.

The 10,560 bit storage may have to be increased somewhat to cope with special cases. Complexity in the decoder for the selection of the correct element size and composition of the Hangul character will also be necessary.

A suitable transmission coding for the purposes of the invention will now be considered. It will be recalled that a Hangul character is made up of an initial consonant (1 of 19), vowel (1 of 21) and possible final consonant (1 of 30). This may be simpiy coded as follows: Initial consonant 19 5 bits Vowel 21 5 bits Final consonant 31 5 bits 15 bits The final consonant is coded as 31 versions, 30 real final consonants and a 'null' code meaning there is no final consonant present.

Also a code for 'space' is required; this might perhaps be done best using another vowel code combination (22 instead of 21).

Unfortunately, 1 5 bits is just too many to fit neatly into an existing teletext specification which uses groups of 7 data bits plus a parity bit. However, this difficulty is removed if the coding for the initial consonant and vowel (which are always present) is combined. If 20 possible initial consonant codes IC are combined with 24 possible vowel codes VC, 480 code combinations are necessary which can be coded in 9 bits. Together with 32 possible final consonants coded as 5 bits, only 14 bits in total are necessary.

Effectively then, 14 data bits provide the transmission coding for each Hangul character, from which 15 bits are reconstituted in the decoder; i.e. 5 bits each for initial consonant, vowel and final consonant. The 14 transmitted bits occupy two conventional bytes as in the normal rows of an existing teletext specification. Since a Hangul character is twice the width of a conventional character the same number of bytes (40) is needed per row. As there are only 10 rows of text on a page less data needs to be transmitted than usual. As with other 525 line systems, data would be sent in 'geared' form with 32 bytes per data line to permit a 5.72Mb/s data rate. A total of 13 data packets would be required per page (10 part rows and 3 'fill-in' packets).

The decoding scheme used is shown in Fig.

5. For each character, two transmitted bytes 2BY each of which has 7 data bits 7B and a parity bit P, are used. This yields 14 data bits 14DB. Five of these bits 5B, having 32 valid combinations 32C, provide the codes for the final consonants (FC=30), a null (NU=1) meaning no final consonant, or a code extension (cue=1). The last code combination allows the meaning of the other bits to be redefined for spacing attributes, English etc.

The remaining 9 bits 9B are used to reconstitute the two 5-bit codes 5B' and 5B" for initial consonant IC and vowel VC, respectively. In each case two bits 2B' and 2B" are transferred directly, the remaining 3 bits coming from the decoder DEC. The decoder DEC operates to convert the remaining 5 bits 5B with 30 valid combinations to two groups of 3 bits 3B' and 3B" having 5 and 6 valid combinations respectively. A simple encoding algorithm is used i.e.

Z=SY+X where Z is the transmitted 5-bit code, Y represents 3 bits used for the vowel code and X represents 3 bits of the initial consonant code. In decimal terms, X can take the values 0 to 4 and Y the values 0 to 5. So, for example, if Z is 13, then Y must be 2 and X is 3. These outputs from the decoder DEC are combined with the 2 directly transferred bits to produce one of 20 possible initial consonant codes IC and one of 24 possible vowel codes VC.

Of course, the 5-bit input to the decoder DEC has 32 possible combinations rather than the 30 required. The remaining two combina tions could be used for code extension if necesary, either in simple or decoded form. The decoder function itself could be a divider circuit or a look-up table in ROM, a simple function compared to the overall teletext decoder complexity.

Having received a transmission code for a Hangul character and processed the data to yield three 5-bit codes, the teletext decoder has to generate the character for display. The algorithm used is shown in Fig. 6.

When the algorithm is entered into (ST) the first task as represented by the box VC? is to examine the code for the vowel, as the vowel shape determines the subsequent composition of the character. Assuming it is not a 'space' code, one of the three vowel types is possible. Any one of 9 'vertical' vowels 9VV implies the S1 basic shape format with the initial consonant to the left, whereas any one of 5 'horizontal' vowels 5HV places the initial consonant above using the S2 basic shape format. Any one of 7 'composite' vowels 7CV, however, uses most of the matrix itself with the S3 basic shape format, leaving the initial consonant to be added at the top left.

These rules apply if a 'null' code is received, as represented by the boxes NU, indicating that no final consonant is present. The boxes CC? represent the interrogation for the final consonant. If a final consonant is present, as signified by the boxes 30 FC, shape formats S4, S5 and S6 are used instead of S1, S2 and S3, respectively, and the final consonant is placed underneath the vowel. The box SS' represents the resultant shapes composed from 1 of 19 initial consonants IC and 1 of 21 vowels V, and the box SS" represents the resultant shapes composed from 1 of 19 initial consonants IC, 1 of 21 vowels V and 1 of 30 final consonants FC.

The storage necessary for each element can be optimised according to the type of element, with appropriate use of double width, double height and half height techniques as necessary. Referring again to Fig. 2, a vertical vowel VV can be stored as a 12x 10 matrix and used in this form in shape format S4, with double height for shape format S1. The initial consonant IC would be stored in a 12x 10 matrix and used this size in shape formats S3 and S4; double height in S1; double width in S2; half height in S6 and double width/half height in S5. These same initial consonant forms are also valid final consonants FC and can be used double width for S4, S5 and S6. Similar techniques are used for the horizontal and composite vowels.

The half height matrix size is used in shape formats S5 and S6 for initial consonants IC, horizontal vowels HV and composite vowels CV. With careful choice of character shape it is proposed to produce acceptable displays in half height using alternative lines of the basic 12x 10 matrix. This allows horizontal lines to be the same thickness whether they are normal or half height. This assumes a non-interlace display. As aforesaid, Fig. 3 shows an example of a character of the same thickness normal and half height size.

It can be seen that the character generator has to assemble up to three components to produce a Hangul character. However, the access time of the character memory is the same as normal, because when scanning any given t.v. line only two elements are required, in twice the width of a conventional letter.

Similarly, the page memory access time is the same as normal, since 14 bits are required for 20 characters per row. The difference for generating Hangul is the logic necessary to assemble the character, together with the appropriate matrix sizes for the elements. Access time to a character may be longer because of the processing involved, but this can be compensated for in the timing without difficulty.

Figs. 7 and 8 when joined together with Fig. 7 above Fig. 8 show a logic diagram of a Hangul character generator.

Three latches L1, L2 and L3 receive the character input data for an initial consonant IC, a final consonant FC and a vowel V, respectively. These three latches are driven by a clock CC, which is for example a 0.5 MHz clock. The outputs from the latches L1 and L2 are appied via a multiplexer MX as data for addressing consonant and final consonant read-only memories ROM1 and ROM2. The output from the latch L3 is applied directly as data for addressing three further read-only memories ROM3, ROM4 and ROM5 which contain the composite vowels CV, the horizontal vowels HV and the vertical vowels VV, respectively. The outputs from the two latches L2 and L3 are also applied to two decoders DE1 and DE2.The decoder DE1 controls the multiplexer MX and also a matrix MA1 which determines whether vowels and consonants are to be single or double height, as determined by the final consonant present in characters. The decoder DE2 controls an enable matrix EM which supplies enable signals A to G. The decoder DE2 identifies to the enable matrix EM the type of vowel, vertical, horizontal or composite, which is represented by the latched vowel data V. A combined lines-perrow counter CC and decoder DEC which are driven by a clock CL at the line rate, provide further outputs to the matrices MA1 and EM.

The counter CC is reset each field by a reset pulse RS. The signals from the decoder DEC to the enable matrix EM identify "top half row" and "top quarter row" of character rows. AND-gates GA1 and GA2 and an inverter INV gate the various signals into the multiplexer MX and the matrix My 1. An OR-gate G01 supplies a control signal X to a second single/double height matrix MA2 which addresses the memory ROM 1 to signify "halve" height. A third, double height matrix MA3 ad dresses the memory ROM 5.

Enable circuits EN receive the enable signals A to G as indicated. The text output TO is produced from a 24-bit shift register SR which is filled by the outputs from the enable circuits EN. This register SR is driven by a clock CK at 12 MHz and is loaded every 2,us in response to a load pulse 2,us.

The enable matrix logic is shown in the table T of Fig. 9. The first column EN shows the enable signals A to G, and the other columns show the logic states (0 or 1) or "don't care" (x) for a vertical vowel VV, a horizontal vowel HV, a composite vowel CV, a final consonant FC, the top quarter row TQR and the top half row THR, respectively. These logic states have none, one or more alternatives as indicated and occur when the outputs of each memory ROM(.) are enabled onto the text video output. The third alternative for the enable signal F is for a normal final consonant, and the enable signal G output is for a final consonant combination.

The television teletext receiver shown in Fig.

10, in which the invention can be embodied, has its front end 4 connected to receive an incoming television video signal VS. For normal picture display by the television receiver, the demodulated video signal VS' is applied to a colour decoder which produces the R,G and B component signals for the picture display.

Time base circuits for a display tube (not shown) receive the usual line and field synchronising pulses from a sync. pulse separator circuit which extracts these synchronising pulses from the video signal VS'. The element 5 represents the colour decoder and these other circuit elements which are provided for conventional picture display.

The demodulated video signal VS' is also applied to a teletext decoder section of the television teletext receiver which deals with the receipt and display of the alpha-numeric text and other teletext information that is received in digitally coded form. This section comprises a video processor circuit 6 which performs inter alia data slicing for retrieving teletext data pulses D from the video signal VS'. The video processor circuit 6 also produces input data clock pulses C from the data pulses D. The data pulses D are fed together with the clock pulses C to a data acquisition circuit 7 which is operable to feed selected groups D/G of the teletext data pulses to a memory 8 as address and display or other information. The memory 8 has a capacity for storing at least one page of information, comprising a plurality of data rows.

A logic processor 9 is operable in accordance with select signals S applied to it from a remote control arrangement 10 to control which groups of teletext data pulses are acquired by the data acquisition circuit 7. The arrangement 10 has a receiver part 10a and a remote transmitter part comprising a transmitter 10b and a keypad 10c. The processor 9 is further operable to read out from the memory 8 display information for the selected page, for application to a character generator 11 which is responsive to this display information to provide R, G, B component signals for the display. A timing circuit 12 provides timing signals on connections tl to t3 for the circuit elements 7, 8 and 11. These circuit elements and the timing circuit 12 are accessed by the processor 9 via an interface circuit 13.The operation of the timing circuit 12 is synchronised with the received video signal V by a composite pulse signal VCS which contains the line and field synchronising pulses which are separated from the demodulated video signal VS' in the video processor 6.

In the television teletext receiver shown in Fig. 10, only single line connections have been shown for the interconnections between the various circuit elements for the sake of simplicity. However, it with be apparent to a person skilled in the art that in practice most of these interconnections would be multi-line.

For instance, whereas the teletext data pulses D retrieved from the video signal VS' would be applied serially to the data acquisition circuit 7 over a single connection, serial-to-parallel conversion would take place within this circuit 7, so that the groups D/G of teletext data pulses would be applied to the memory 8 in parallel over a multi-line connection. Also, the connection between the processor 9 and the interface circuit 13 would be a multi-line bus, for instance, a so-called 12C bus. The processor 9 can be a commercially available microcomputer; e.g. from the MAB 8400 Series (Philips). The circuit elements 7, 11, 12 and 13 can be the integrated circuit EURO CCT type SAA 5240 (Mullard). The circuit element 6 can be the integrated circuit VIP2 type SAA 5230 (Mullard).

Although a composite television receiver for receiving both normal picture information and teletext information is exemplified in Fig. 10, it will be appreciated that the teletext decoder section for data acquisition together with the front 4 may be provided as a separate teletext decoder which is adapted to feed either a CRT display monitor or other display device, or a conventional television receiver.

A teletext television system is described in United Kingdom patent specification 1 370 535. Also, the document "Broadcast Teletext Specification", September 1976, published jointly by the British Broadcasting Corporation, Independent Broadcasting Authority and British Radio Equipment Manufacturers Association, gives details of a specification on which are based the CEEFAX and ORACLE (Registered Trade Marks) teletext television systems operated in the United Kingdom by BBC and IBA respectively.

Claims (12)

1. Data display apparatus -for displaying data represented by digital codes, the displayed data being composed of characters the shapes of which are determined by selected dots of a dot matrix which constitutes a character format for multi-element syllabic characters, wherein the character dot matrix has a plurality of different display shape formats one of which is selected for the display of a character, the selection being made in dependence on the form of the vowel of the character.

2. Data display apparatus as claimed in Claim 1, wherein six different display shape formats are used for the display of Hangul characters comprising a vowel, an initial consonant and may be a final consonant.

3. Data display apparatus as claimed in Claim 1 or Claim 2, having a character memory containing a basic character dot matrix for each of the character elements provided, an element as used in a displayed character being displayed original size, double height, double width, or half height, depending on the display shape format required for displaying the character.

4. Data display apparatus as claimed in Claim 3, wherein character elements are displayed half height by using only alternate lines of their basic character dot matrices.

5. Data display apparatus as claimed in any preceding Claim, comprising processor means responsive to a received transmission code for a Hangul character to determine: (a) what is the vowel code, (b) whether the determined vowel code pertains to (i) one of a plurality of 'vertical' vowels, (ii) one of a plurality of 'horizontal' vowels, or (iii) one of a plurality of 'composite' vowels, and (c) what is the final consonant code, if any, the process means being operable to select the appropriate display shape format for the character in accordance with the above determinations.

6. Data display apparatus as claimed in any preceding Claim, using display shape formats of the form shown in Fig. 2.

7. Data display apparatus substantially as hereinbefore described with reference to the accompanying drawings.

8. A method for displaying data represented by digital codes, the displayed data being composed of characters the shapes of which are determined by selected dots of a dot matrix which constitutes a character format for multi-element syllabic characters, wherein the character dot matrix has a plurality of different display shape formats one of which is selected for the display of a character, the selection being made in dependence on the form of the initial vowel of the character.

9. A method as claimed in Claim 8, wherein six different display shape formats are used for the display of Hangul characters comprising a vowel, an initial consonant and may be a final consonant.

10. A method as claimed in Claim 8 or Claim 9, using a basic character dot matrix for each of the character elements provided, an element as used in a displayed character being displayed original size, double height, double width, or half height, depending on the display shape format required for displaying the character.

11. A method as claimed in Claim 10, wherein character elements are displayed half height by using only alternate lines of their basic character dot matrices.

12. A method as claimed in any preceding Claim, comprising the steps of: (a) receiving a transmission code for a Hangul character, (b) determining the vowel code of the character, (c) deciding whether the determined vowel code pertains to (i) one of a plurality of 'vertical' vowels, (ii) one of a plurality of 'horizontal' vowels, or (iii) one of a plurality of 'composite' vowels, (d) determining the final consonant code, if any, and (e) selecting the appropriate display shape format for the character in accordance with the above determinations.