KR19980065627A - Method and apparatus for decoding subtitle data - Google Patents

Abstract

A method of decoding caption data for audio data in a composite video signal, comprising the steps of: extracting a data packet of caption data; retrieving control codes and character data by a plurality of bits of data; Mapping character pointers by code and tail code. You can get a pointer to the completed character directly from the incoming data packet.

Description

Method and apparatus for decoding subtitle data

1 is a block diagram showing a general subtitle processing apparatus.

2 is a waveform diagram of a data transmission signal for one field of line 21 for transmitting caption processing data in a television signal.

Figure 3 is a standard Latin character code table that includes special and alphabetic characters.

4 is a bit map for decoding or encoding a 4-byte language data packet into a character component pointer to designate character data stored in a decoder read-only character memory in a caption processing apparatus according to the prior art.

5A is a bit map for decoding or encoding a three-byte language data packet into a character pointer to designate character data stored in a decoder read-only character memory according to an embodiment of the present invention.

5B is an exemplary diagram of a completed bit map for decoding or encoding Korean characters according to an embodiment of the present invention.

5C is an illustration of a completed bitmap for decoding or encoding an alphabetic character in accordance with an embodiment of the present invention. 6 is a portion of Hangul character code table illustrated to decode or encode Hangul characters according to an embodiment of the present invention.

Explanation of symbols for the main parts of the drawings

1: subtitle data extraction unit 2: subtitle timing unit

3: subtitle generation unit 4: character ROM unit

5: attribute register portion 6: subtitle overlapping portion

The present invention relates to a method and apparatus for decoding caption data, and more particularly, to a method and apparatus for decoding caption data for decoding caption data on a character basis.

In general, subtitle processing refers to processing of data for the audio portion of a television or a home video cassette recorder in a specific line of the vertical blanking period of a standard television signal to be displayed as characters on a television screen. In this case, the specific line to which subtitles are transmitted is generally selected by line 21 of the vertical blanking period of the NTSC standard television signal format, which may be any of lines 10 to 21. In one European standard television broadcast signal, line 16 or line 329 is selected. One known decoder for receiving subtitle data transmitted for an audio portion is designed to transmit and process special characters including musical notes, roman characters, punctuation, attributes, control codes, low accented alphabet characters, and stored character sets. It includes a memory to examine.

In general, a broadcast signal received through a television tuner is demodulated from an audio and video carrier and a selected channel frequency so that the audio signal is adjusted by a volume level control circuit and output to a channel 3/4 modulator, and at the same time video containing subtitle data. The signal is superimposed on the subtitles generated by the caption processing apparatus and output to the channel 3/4 modulator. The combined video and audio signals from the 3/4 modulator preset to either channel 3 or 4 are modulated to the selected frequency and then provided to the television set. When the television set is tuned to the preset channel 3 or 4, the subtitles are displayed on the screen. Will be displayed.

1 shows a schematic block diagram of a general subtitle processing apparatus.

First, with reference to FIG. 1, a signal processing procedure will be described until the caption data transmitted from a specific line of the vertical blanking period of the standard television signal is displayed on the screen as characters.

The received baseband video signal is input to the caption data extraction section 1, the caption timing section 2, and the caption superimposition section 6. In order to obtain the reversal force value required for data slicing, a clock that is operated during the burst is used, and the subtitle data extractor 1 clocked by the subtitle timing unit 2 extracts a serial data stream composed of two bytes to generate a subtitle generator ( 3) to provide.

The subtitle timing unit 2 operating at the subline frequency identifies the position of the vertical blanking period and the received field to determine the starting point of the selected line, for example, the line 21, and the subtitle data extraction unit 1 to determine the subtitle. When the clock is supplied to the caption generating section 3 which receives the data, and the caption superimposition section 6 is superimposed on the incoming video signal, the output of the character ROM section 4, i.e., the attribute data for the caption starts to be stored. The attribute register section 5 is clocked, and the character ROM section 4 is also clocked.

The caption generating section 3 receives caption data, outputs pointer data to the character ROM section 4, and outputs the attribute data to the attribute register section 5. The character is designated by the output pointer data of the subtitle generating section 3 and is selected in the character ROM section 4, and the attribute is set in accordance with the attribute data input to the character ROM section 4 through the attribute register section 5. Is determined. The attribute for the character may, for example, be related to whether the character blinks or has a particular color.

The attribute register section 5 also provides attribute data to the caption superimposing section 6 to determine the attribute of the video signal containing the caption. In this way, a complete subtitle may occur that may be rolled up, popped on, or painted on.

In the caption processing as described above, in addition to the caption data, display command data for synchronizing the spoken voice with the caption data is required. Accordingly, the decoder for caption processing further includes a processor that additionally decodes the display command data to display the caption in the lower portion of the television image. In addition, video information appearing on the horizontal line of the caption image is replaced with text data output from the caption processing apparatus. The size of the characters varies in proportion to the size of the television screen. For example, a 19-inch subtitle is displayed at a height of 0.5 inches, and white text is displayed against a dark background so that users can easily view the subtitles without interrupting viewing. Appears.

Next, an apparatus and method for decoding or encoding a language data packet into a character component pointer to designate character data stored in a decoder read-only character memory of the caption processing apparatus shown in FIG. 1 will be described.

One subtitle encoder that is commercialized according to NTSC and Federal Communications Commission standards is EN 230 smart encoder from E.E.G Enterprise. This encoder inserts subtitle data into fields 1 and 2 in line 21 of the retrace blanking period. Smart decoders with built-in decoders can recognize gaps in subtitle data and fill these gaps with text service data such as additional subtitles or emergency alerts.

According to the encoding apparatus, two bytes of data are required, one byte per field or two bytes per field, which is understood to specify an ASCII character code which is usually used for encoding a character set for a certain character. According to the NTSC broadcasting method, a transmission rate of 30 frames per second is defined, so that the voice data pronounced as a complete sentence and the display command code at the same time can be as high as 60 characters per second.

However, the caption processing apparatus of this type does not provide caption processing for characters of all countries that follow the NTSC method, but only subtitle data of a language consisting of one character such as English, French, German, and Spanish based on the Roman alphabet. I had a receivable problem. In other words, among the countries that follow the NTSC system, especially Korean characters, that is, Hangul, are syllable characters that can be completed by combining two or more phonemes, that is, at least one of 17 consonants and 11 vowels. Had the problem of becoming difficult.

Accordingly, according to US Pat. No. 813,021 provided as a method for subtitle processing of syllable characters such as Hangul, four-byte syllables in a particular line of an alternating field of NTSC, PAL, SECAM, MUSE or other television or communication signals Language data is sent.

4 shows a bit map for decoding or encoding a 4-byte language data packet into a character component pointer to designate character data stored in a decoder read-only character memory in a conventional caption processing apparatus.

As shown in FIG. 4, these data bytes are formed by encoding syllable blocks of language data in 4 bytes of 2 byte increments in accordance with existing standards, where 4 bytes of data are received in 3 byte points. It is decoded and stored in a character memory containing Roman and numeric and special characters as well as Korean syllable characters.

In other words, four 7-bit data packets composed of one head code and three tail codes are decoded into one 7-bit pointer and two 8-bit pointers using a bitmap method. In the Hangul characters, the above three pointers indicate a letter component for each syllable of one Hangul character. Therefore, one Hangul character is displayed on the screen by sequentially superimposing the character components designated by the character component pointers for each syllable.

However, according to this method, since each pointer byte for a character component of a Hangul character is designated, more data than necessary is required, at least two memory accesses are required, and a plurality of Hangul characters are displayed to represent one Hangul character. This is a hassle to nest text elements. In particular, there is a problem in that unnecessary data is transmitted by using a data packet composed of 4 bytes with respect to a standard Latin character that can be processed as 1 byte or 2 bytes.

Accordingly, an object of the present invention is to provide a caption data decoding method and apparatus capable of decoding caption data of a television signal character by character.

A caption data decoding method according to the present invention for achieving the above object comprises the steps of: extracting a data packet of film data; retrieving control code and character data by a plurality of bits of data; and head code of character data; And mapping the character pointer by tail code.

A caption data decoding apparatus according to the present invention for achieving the above object comprises: a caption data extractor for extracting and outputting caption data from a specific horizontal line of a vertical blanking period; and the caption data input from the caption data extractor A character generator for determining whether to include printable character data and outputting a plurality of character pointers, and a character memory unit for storing a plurality of characters and outputting character data selected by the plurality of character pointers of the character generator; It is characterized by including.

The method and apparatus for decoding caption data according to the present invention are characterized by facilitating encoding and decoding of data by generating and processing caption data for an audio portion represented by characters on a television screen in units of characters.

Hereinafter, a method and apparatus for decoding caption data according to the present invention will be described in detail with reference to the accompanying drawings.

Subtitle data for the audio signal is transmitted as a composite data signal encoded in line 21 of the standard NTSC video signal. This signal is composed of a clock execution signal, a start bit, and two bytes of data, as shown in FIG. Each 8-bit of the 2-byte data consists of 1-bit parity and 7-bit ASCII or special characters and command code data, as shown in FIG. Since 8 bits are transmitted per field, a data transmission rate of 480 bits, that is, 60 bytes, is obtained for 1 second. This data stream contains encoded instruction information that provides instructions for formatting other characters in addition to the characters to be displayed. According to the present invention, not only ASCII code data but also Korean character data are transmitted.

2 also shows in detail the waveform diagram for the subtitle data signal of line 21 in the vertical blanking period. The caption data signal is composed of a horizontal blanking pulse, a color burst, a clock execution signal, two data bits of logic level 0, one start bit of logic level 1, and 16 bits of data. The clock execution signal, or data decoder clock, has a frequency of 0.5034965 MHz, which is 32 times the horizontal line frequency of 15.732 KHz, with seven sine wave cycle bursts phase locked to a two-byte subtitle data execution signal having a frequency twice that value. Is done. The last two cycles of this clock execution signal and two bits of logic level 0 and one start bit of logic level 1 constitute an 8-bit frame code indicating the start of data. The data following this 8-bit frame code is transmitted in NRZ format, followed by 7 bits for each character data followed by odd parity bits for error detection.

Prior to each subtitle transmission, a preamble code is preceded to form a row address and a display color code, wherein the printed and non-printed character codes of the preamble address form a control code. The first byte of this control code is a non-printing ASCII character except a code between 0000000 and 0010000 within the ASCII character code 0011111, and the second byte is a printing character code between 0100000 and 1111110. In addition, this control code is transmitted two consecutive times in the odd and even fields of line 21 for accuracy of information, including for example an alert or roll-up or subtitle termination. On the other hand, a subtitle intermediate control code may be used to block between two words to change the display condition, such as a color, underline, or italics used for emphasis.

All data received after the control code as described above is interpreted as a character and loaded into the memory as a character to be printed.

Typically, 15 rows are assigned to display subtitles on a television screen, with a maximum of 32 characters per row. Up to four rows are used in a normal subtitle processing operation in order to synchronize the subtitles with a typical activity scene such as a movie of a television image and so that the user is not greatly disturbed for viewing. And these four character rows can be located anywhere in the 15 assigned rows but generally appear in the bottom row so as not to interfere with the activity of the television image. The first character row starts on line 43 of the odd or even field of the NTSC standard broadcast signal. Each character row occupies 13 lines of odd or even fields, that is, 26 lines of two-field frame 525 horizontal lines.

According to the data transmission signal format of line 21, one data byte of a signal to be transmitted is composed of a word of 7 bits and a parity of 1 bit. Therefore, at least three 7-bit words are required to transfer characters that can be displayed using 16-bit data, such as Korean characters. Therefore, up to 20 Korean characters per second can be processed according to a transfer rate of up to 60 bytes per second. Can be. On the other hand, since 8-bit data is used to display standard line characters such as the English alphabet, at least two 7-bit words are required, but in consideration of transmission of characters requiring three 7-bit characters, two standard Latin characters are used. Allocates 16 bits for a character. This also allows for processing up to 40 standard Latin characters per second.

5A shows a bitmap for decoding a three-byte language data packet into a character pointer or encoding a character data packet according to the character pointer to designate character data stored in a decoder read-only character memory according to the present invention. It is.

The principle of this bitmap scheme may be modified to implement a bitmap scheme that differs without departing from the scope of the present invention.

As shown, a 3-byte data packet consists of a 1-byte head code and a 2-byte tail code, where bit 7 of each code can be a parity bit to match the data transfer method of line 21 and to P. Display. Bits 2 through 6 of the head code in the data packet are set to 11001 to indicate that it is a head code, and bits 0 and 1 are mapped to bit 0 of character pointers 0 and 1, respectively. Meanwhile, bits 0 through 6 of tail codes 0 and 1 in the data packet correspond to bits 1 through 7 of character pointers 0 and 1, respectively. At this time, the bit 7 of each character pointer can distinguish between the character displayed in the character code table including the Hangul character code table KSC5601 illustrated in the present invention and the ASCII standard Latin character.

And bit 7 of the character pointer for ASCII standard Latin characters represented by 7 bits is set to 0, and bit 7 of the character pointer for characters represented using two bytes, such as Korean, is 1 Is set to. For example, if bit 7 of the character pointer 1 is 1, it means a character using two bytes. Therefore, the character pointers 1 and 0 are a pair of character pointers for one character. On the other hand, if bit 7 of character pointer 1 is 0, it can be seen that it is an ASCII standard Latin character that can be specified by only one byte. Therefore, bit 7 of character pointer 0 should be checked separately. Bit 7 of character pointer 0 is also a value of 0 for any standard Latin character; if 1, character pointer 0 of this language data packet and character pointer 1 of the next data packet together specify a character. It means a pointer. Here, if only one standard Latin character is represented, the character pointer needs only one byte, so character pointer 0 is replaced by 00000000 so that bit 0 of the headcode in the data packet is specified as 0 and bit 0 to tail code 0 of the tail code 0. 6 is also specified as 0000000.

Although not shown, the encoding device in the present invention includes a standard personal computer having an input / output device suitable for the present application. For example, it is also possible to simultaneously broadcast recorded audio subtitles and news using speech recognition and encoding circuitry applied to data input.

In order to understand the encoding method of the present invention, an encoding method for the Hangul character “angle” will be described in detail with reference to FIGS. 5B and 6.

The Hangul character “angle” corresponds to the coordinates (B0, A2) in the Hangul character set of FIG. 6, which exemplifies a part of the Hangul character code table KSC5601. If each coordinate value is mapped to character pointers 0 and 1, respectively, and the most significant bit is given priority in binary format, bits 0 through 7 of character pointer 1 are 00001101, and bits 0 through 7 of character pointer 0 are shown in FIG. 01000101. The contents of these two-character pointers are bit mapped into the contents of one byte of head code and two bytes of tail code for data transfer.

First, bits 0 and 1 of the head code are mapped to bits 0 and 1 of character pointers 0 and 1, respectively, and are set to 00, bits 2 to 6 are always 11001 as head code signals, and bit 7 is set to 0 as odd parity. do. Next, bits 0 through 6 of tail codes 0 and 1 correspond to bits 1 through 7 of character pointers 0 and 1, respectively, and are set to 1000101 and 0001101, and bit 7 is set to 0 as odd parity. As a result, it can be seen that the head code of the data packet that can be obtained by the character pointers B0 and A2 for the Hangul character “angle” is 00110010 and the tail codes 0 and 1 are 10001010 and 00011010.

Referring now to Figures 5C and 3, the encoding method for the word "Hi" consisting of the letters "H" and "I" will be described.

As described above, the alphabet can be represented by only 7 bits, so the standard Latin character set of FIG. 3 is used. The letter “H” in code 3 corresponds to code 48 and “I” corresponds to code 69. If you specify a character pointer of one byte for each alphabet, in Figure 5C, bits 0 through 7 of character pointer 1 are the values 00010010 for "H", and bits 0 through 7 of character pointer 0 for "i". Value 10010110. By bit mapping the contents of these two-character pointers to the contents of the data packet for data transmission, the head codes of 10110011 and tail codes 0 and 1 of 00101100 and 00100101 can be obtained.

As described above, the present invention has the effect of simplifying the encoding and decoding process of the data by processing the subtitle data for the audio portion represented by the character on the television screen in units of characters.

Claims (10)

CLAIMS 1. A method for decoding subtitle data received on a particular horizontal line of a vertical blanking period of a composite video signal, the method comprising: a plurality of data bytes, indicated by a plurality of bits appearing in one of the plurality of data bytes. Decoding a data packet, storing a plurality of characters in a memory, retrieving whether the data packet contains nonprinting characters, and a plurality of residual bits for the plurality of bits representing the data packet. Mapping to a plurality of character pointer bytes designating the in-memory character for the printable character. The method of claim 1, wherein the number of bytes of each data packet includes three. The method of claim 1, wherein each data packet includes one head code and two tail codes. 2. The method of claim 1, wherein the number of bytes of the character pointer includes two. The method of claim 1, wherein the number of bits appearing in one byte of each data packet includes five. The method of claim 4, further comprising: searching for a character that can be designated as one byte and a character that can be designated as two bytes of the character pointer. 7. The method of claim 6, wherein the data packet composed of characters that can be designated by one byte of the character pointer includes two characters that can be designated by one byte. An apparatus for decoding caption data received at a specific horizontal line of a vertical blanking period of a composite video signal, comprising: a caption data extracting unit for extracting and outputting caption data at a specific horizontal line of a vertical blanking period; A character generator for determining whether the caption data input by the extractor includes printable character data and outputting a plurality of character pointers, and storing the plurality of characters to be selected by the plurality of character pointers of the character generator; And a character memory unit for outputting character data. 9. The apparatus of claim 8, wherein a data packet is identified by a plurality of bits of the caption data. 9. The apparatus of claim 8, wherein the caption data includes one head code and two tail codes.