Direct broadcasting by satellite (DBS) systems use frequency modulation (FM), rather than amplitude modulation (AM) as used in most terrestrial systems, because of the limited power of the satellite transmitter. Some DBS services use conventional PAL, or NTSC, for coding for the colour television signal but services introduced recently employ the MAC (multiplexed analogue components) system [17].
With conventionally coded signals the teletext data can be carried in the FBI, as for terrestrial transmissions. The received decoding margin is somewhat lower in the case of DBS due to the non-linearity of the FM demodulator in the receiver, but the decoding margin of the demodulated signal is adequate for normal teletext decoders. Furthermore, the satellite aerial system is highly directive and therefore the signal is not normally contaminated by reflections, which are most often the main cause of distortion of digital data signals.
The MAC transmission system uses time multiplexing for the luminance and colour-difference signals rather than frequency multiplexing as used for conventionally coded colour signals. Digital techniques are used to time compress the luminance signal in time by a factor of 3/2, and the lower bandwidth colour-difference signals by a factor of 3. The compressed luminance signal and one colour-difference signal are transmitted in one line period. The other colour-difference signal is on alternate lines. At the receiver the video signals are expanded and the RGB drive signals reconstituted from the luminance and colour difference signals in the conventional manner. If a digital teletext signal were included in the FBI periods of the original video signal it would be time compressed with the luminance signal. At the receiver, the data signal is expanded but unfortunately the compression and expansion of the data signal results in a very significant reduction in the decoding margin, typically some 40%, which renders the data signai unusable by present teletext decoders. The teletext data signal in the FBI period could bypass the compression circuit in the coder, since there are no colour-difference signals present, and use the full line period. A delay circuit would then be necessary in the receiver to compensate for the delay in the video processing circuits.
Although digital compression techniques are used for the video signals, all the other signals such as stereo sound, line and field synchronizing, blanking and the data are all sent as data packets. The format of the transmitted MAC signal is shown in Figure 8.1. The specifications for the MAC transmission standards detail the location and format of the different data packets together with the priority required for different purposes [18].
64 us | |||||
19.5 µs | 35 µs | ||||
625 lines |
sync/ sound/ data |
spare | field 1 | ||
colour difference | luminance field | ||||
sync/ sound/ data |
spare | field 2 | |||
colour difference | luminance field |
Figure 8.1 Transmitted frame for MAC signal
A number of variants of the MAC standard have been specified which have differing data capacities. Nevertheless all of them provide adequate space for a teletext service. The teletext data is conveyed in MAC packets which contain the address header followed by the useful data. Two possible levels of error protection are provided, the lower level corresponding to that used in normal terrestrial systems. The higher level of protection, with its larger overheads, halves the useful data that can be carried by each packet.
It is likely that many satellite receivers will be adaptors for use in conjunction with existing television receivers. If the television signal is transcoded to PAL format the teletext data can also be transcoded to a conventional teletext signal format in the FBI for use with a standard television receiver. Receivers that employ an integral MAC decoder would not require the data transcoding facility as teletext decoding would be part of the MAC decoder function.
When the MAC transmission standard was being established a number of different data coding techniques were considered. A proposal, known as C-MAC, employed a form of coding with a 2 - 4 phase shift key arrangement (symmetrical PSK). This method is technically superior to the other proposed formats, providing very good error performance in the presence of noise and making full use of the satellite channel bandwidth. Nevertheless, after further consideration, a duo-binary coding system [19] was found to be a highly acceptable compromise as it provided an economy of baseband bandwidth and commonality with other proposals made to the EBU. A duo-binary (three-level) coding technique permits a data rate of 20.25 Mbits/second, which conesponds to the video sampling frequency in the 8.5MHz baseband channel. This method is used in the D-MAC system. The D2-MAC system provides half this data rate. B-MAC, which is used for the network distribution of television and data signals by satellite, uses quartenary (four-level) coding and provides a data rate of 14 Mbits/second.
The essential feature of a duo-binary coding system is that it is a three-level data format, as illustrated in Figure 8.2. Within a given bandwidth, this format provides a data rate which is twice that of a two-level system, but the noise threshold is halved. A binary-coded signal can be translated into duo-binary code by passing it through a relatively simple delay and differential coding circuit, as shown in Figure 8.2. The input data is inverted and applied to a differential coding circuit. The output (waveform 3) is 1 only when there is a difference between the two inputs to the exclusive OR gate and one input is a 1. This output is delayed by 1-bit period and added to it to produce the three-level duo-binary data. The 1 in the input data always corresponds to a +1 or -1 in the output, the input 0 is always at the output 0 level. (Had the input data not been inverted then the input 0 would correspond to the output 1 and the input 1 to the output 0 level.) If an odd number of 0 has occurred between the current and last 1 in the output, then the output has a different polarity.
![]() Figure 8.2 Duo-binary coding of data |
A typical decoding arrangement is shown in Figure 8.3. The output of two slicing circuits is applied to an OR gate so that the output data is the sum of the data contained in the two levels of the input signal.
![]() Figure 8.3 Decoding duo-binary coding of data |
The data capacity of the sound and data packet multiplex of the D-MAC or D2-MAC packet systems can be divided in a flexible way between sound channels, the service identification channel and other broadcast data services such as teletext, subtitling and telesoftware, together with a transparent data channel. If a vision signal is not required then the data capacity of a MAC packet channel can be increased by replacing the area of the frame normally occupied by the vision signal (Figure 8.1) with data bursts composed of data packets, in a similar manner to the existing data burst.
Teletext services using the D-MAC system are carried in the data packets. The format of the teletext data is similar to that employed in WST in which 45-byte teletext rows or packets are used. (These should not be confused with MAC sound/data packets.) The first three bytes of the 45-byte block comprise a clock run-in sequence followed by a framing code, which provides synchronization of the teletext data when carried in the FBI of terrestrial television signals. These first three bytes are not required for recovery of the data from MAC packets and are therefore not included in the MAC teletext data block. The remaining 42 bytes consist of a 2-byte magazine and teletext data packet address group (MPAG) followed by 40 bytes of teletext characters. A control byte (CB) is added and then a 2-byte cyclic redundancy check (CRC) which covers the 40 bytes of the teletext characters within the block. The data in both MPAG and CB are Hamming coded for error protection and the structure of the teletext block is illustrated in Figure 8.4.
HAMMING PROTECTED PREFIX |
|||
CB | MPAG 2 BYTES |
40 BYTES OF DATA | CRC |
TELETEXT DATA ROW (OR PACKET ) |
Figure 8.4 Teletext data block
Two types of MAC packet are specified for conveying teletext information in the sound/data multiplex, each having different levels of error protection. The first level of protection is intended for use with conventional teletext data services where any error correction can be achieved from the repeated acquisition of the data. The repetitive information in the control byte, together with the CRC code on each 40-byte data field, allows majority logic or bit variation techniques to be used for correction of errors in the block. These arrangements are particularly useful when 8-bit data is being conveyed, for which there is no parity protection on individual bytes. The first level protection of the MAC packet comprises two teletext data blocks carried in the 90-byte useful data field of the MAC packet, as shown in Figure 8.5(a).
23 | 8 | 720 BITS | ||
a) | PH | PT | TELETEXT DATA BLOCK | TELETEXT DATA BLOCK |
b) | PH | PT | TELETEXT BLOCK GOLAY ERROR PROTECTION CODE |
Figure 8.5 Teletext data blocks
in MAC-packet.
(a) First level protection, two teletext blocks per MAC packet;
(b) second level protection, one teletext block per MAC packet
For the second level of protection only one teletext block is carried in each MAC packet. The entire 45-byte teletext data block is coded using the (24,12) extended Golay code. This provides a high level of forward error protection. The 90-byte useful data capacity of the MAC packet therefore consists of 30 Golay words, each containing 12 bits of the teletext data block, as outlined in Figure 8.5(b). The forward error correction of the second level packets affords a significant increase in ruggedness of the data in the presence of errors, so that repetition of data for error correction purposes may be reduced or eliminated.
As for other packets within the MAC sound data multiplex, the packets carrying teletext data commence with a 23-bit packet header (PH) followed by a packet type (PT) byte. The packet header comprises a 10-bit packet address, a 2-bit continuity index and an 11-bit protection suffix provided by a (23,12) Golay code. The continuity index is incremented on successive packets with the same address. For controlled access to the teletext packets the PT byte indicates the type of scrambling that has been applied to packet data. There are three possible values for the PT byte in this application, corresponding to free access scrambled, controlled access scrambled and unscrambled data.
The existence of teletext services in the MAC sound data multiplex is signalled in the service identification (SI) channel. The packet address and sub-frame location for the teletext services are given by an entry in the list of indices carried in data group 0 of the SI channel. This locating information may be supplemented in a subsidary data group, by the name and language of the service and details of any related service features (for example, conditional access).
A teletext service implemented in conjunction with a MAC packet transmission system would utilize conventional teletext data management facilities, as described in Chapter 4. The location of the teletext data, and the special coding required for the MAC packet system, are functions of the MAC coder, in which these aspects form a background task. As the MAC systems have a great deal of flexibility from the data transmission point of view (conditional access, scrambling etc), the final details can be specified only as part of a complete system specification.