5 Teletext decoders and 'in-vision' systems

Basic functions

The teletext decoder, when incorporated in a television receiver, consists of additional integrated circuits associated with the video processing section of the receiver. The decoder input circuit is normally fed with a video signal of some two volts amplitude derived from circuits associated with the video detector. The decoder output provides RGB and blanking signals of a similar amplitude which are fed into the low level RGB amplifiers of the receiver. Teletext decoders are incorporated only into receivers which have remote control. The control signals for the decoder are therefore derived from the remote control circuits [4e].

The major differences between the various teletext systems are seen in the transmission time, which is related to the amount of data required to be transmitted for a given page, and in the complexity of the receiver decoder. The differences in the decoders are principally concerned with the processing that needs to be carried out on the data stream before the information is in a suitable form to be displayed. However, the input and output requirements for the interface between the decoder and the receiver video circuits, and the functions of the input and output circuits of the decoder, are similar in all systems.

Input circuits

The composite video signal is applied to the data acquisition circuit, which is used to process and select the required teletext data for a particular page so that it can be written into memory. Teletext data is recognized by the framing and page address codes irrespective of the actual line in the FBI on which it occurs. The acquisition circuit also performs data slicing, and this function:is probably the most critical as it determines the reliability of data reception in poor reception conditions. For optimum performance the slicing level should normally be at 50% of the data pulse height so that the effect of pulse distortion has minimum effect. Video distortion or co-channel interference may cause variations in black level during the line period, so the optimum slice level will vary during the data line. The acquisition circuit therefore must contain circuits to clamp the video signal and to set the data amplitude at the optimum slicing level, to produce the cleanest possible data signal. The timing functions and data clock, which is locked to the incoming data stream, are also generated by the acquisition circuit. A crystal oscillator normally forms part of this circuit function, the crystal operating at a frequency corresponding to the data rate or a multiple of the data rate. Error-correction circuits are also incorporated into the acquisition circuit. The Hamming-protected bytes are checked and those having only a single bit error are corrected.

Alphanumeric characters forming the page content are transmitted in 8-bit data groups formed from 7-bit character codes plus one (odd) parity bit. The parity bit check for individual characters normally ensures that should an error be received the character will not be displayed. If two errors are received in one byte the decoder will display an incorrect character. Since the teletext magazine of pages is transmitted cyclically, when the requested page is next received characters that were received with an error on the first acquisition will almost certainly be received correctly the second time round and therefore displayed correctly. Should a single error (or odd number of errors) be received for a character that is already displayed correctly, then that character will not be changed.

The viewer selects pages using the remote control handset and the data derived from the remote control system is applied to the data acquisition circuit. This control data will specify the page and magazine number of the page that is required to be displayed. A comparator system is incorporated into the circuit so that only requested data is processed by the error correction circuits. The corrected data is fed, to the memory, as parallel data.

Output circuits

The decoder output circuit contains a character generator ROM for converting the 7-bit character code into a dot matrix pattern form which, in its simplest form, is a 7 x 5 matrix. The ROM contains the complete set of symbols which may be displayed. A typical device would contain at least 96 symbols, which can be selected by means of the 7-bit ASCII code (American Standard Code for Information Interchange) applied to the input circuit. The code table for a basic 96-character ROM is shown in Figure 5.1 and it can be seen, for example, that the code for A is 1000001.

Figure 5.1 Code for 96-character ROM systems

Each character is formed from appropriate dots in the 7 x 5 dot matrix. The ROM is operated on a row scan system, which means that all the dot information in a horizontal row of a character is available simultaneously at the output. A second set of inputs, known as row address input, provides the vertical element of the character by determining which of the seven rows is supplied to the output. The dot information is then fed out in a serial form at a rate determined by the write clock, typically about 6 MHz. This arrangement is illustrated in Figure 5.2. The output signal forms the video output, the character colour being determined by the colour control circuit which selects the appropriate combination of RGB output signals.

Figure 5.2 Operation of basic character generation ROM

‘Character rounding’ can be used to improve the resolution of the characters and makes use of the fact that all the letters, numbers and symbols in normal use are made up of a series of lines. The character rounding circuit modifies the video signal from the character generator when a diagonal line is being generated so that a smoother display is obtained. The effect is illustrated in Figure 5.3. No increase in the bandwidth of the video amplifer in the receiver is required for this improvement. In effect, the dot resolution of each character is doubled in both the horizontal and vertical directions. A teletext character generated using a ROM containing 7 x 5 dot matrix characters is displayed with a 14 x 10 dot resolution.

Display of alphanumeric data on an interlaced television raster gives rise to ‘interline flicker’, which can be objectionable. This effect is overcome by using a non-interlace (288-line) display for teletext pages. Whilst this effectively removes the interline flicker effects it also prevents the use of character rounding. Recently developed decoder circuits enable an interlaced or non-interlaced display to be switched as a design option. At the same time the dot matrix in the character generator ROM has been made 9 x 10 to improve the character resolution. Interlace must be used for subtitles and news flashes, which are inserted into the video picture.

Figure 5.3 Typical character and location of half dots for character rounding

Certain control functions are also performed in the decoder output circuit. These include the selection of graphics or alphanumerics and flashing of words or symbols. A blanking signal for use in the video circuits when a teletext page, or part of a page, has to be inserted into the video signal is also produced by the output circuit. This is required for news flashes and subtitles as these are normally displayed as boxes within the television picture. Timing signals are also fed to the output circuit to ensure that the RGB video and blanking signals are correctly timed with respect to the receiver display.

The ROM (or ROMs) contained in the output circuit must contain the complete character fonts required for various languages that are required to be displayed. Recently developed output circuits also contain an area of programmable ROM. Special character shapes may then be down loaded from the editing system to enable the receiver to display high-resolution graphics or the special characters required for certain languages such as Chinese. Such displays require more information than that contained in a normal teletext page of 960 characters. Larger page memories are therefore required in the decoder and the access time when such pages are being used becomes correspondingly longer (Chapter 2, page 20).

Free-format systems

Decoders for use on free-data format require two areas of memory, together with a processor and the appropriate resident software. The functional diagram of a basic decoder is shown in Figure 5.4. The data from the de-multiplexing circuit is fed into the acquisition memory. The processor reformats this data in accordance with the control and positional information to meet the display requirement and loads this processed data into the display memory. The parallel output of the display memory feeds the character generator ROM and output circuits.

Figure 5.4 Functional diagram of free-format decoder

Fixed-format system

Decoders for the World Standard Teletext (WST) fixed-data format systems require only a single memory and no processor as the data from the acquisition circuit is already in page format form and the memory therefore feeds directly to the character generator and output circuits. The functional diagram of a basic decoder is shown in Figure 5.5. This simplification of the decoder function means that the fixed-format systems can display information virtually instantaneously, as it is received, and the absence of the processor means that the decoder circuits can be very simple. The higher levels of WST teletext which employ data packets require additional memory together with a processor and the necessary resident software.

Figure 5.5 Functional diagram of a fixed-format decoder

Decoder performance

The performance of a teletext decoder, or of a complete teletext receiver, is normally judged by the errors in the displayed picture. The teletext data signal is digitally coded alphanumeric data and provided the ‘0’ and ‘1’ levels of the data stream are well separated the decoder is able to decode the data without errors.

Degradation of the received data will gradually reduce the separation between ‘0’ and ‘1’ levels until a point is reached when the decoder cannot make correct decisions and errors will start to be produced. The address information, that is the magazine and page number and, in the case of a free-format system, the necessary positional information, is protected by Hamming code. Hamming protection of magazine and page number ensures that the decoder will process only that data which corresponds to the requested page. When the page is first received any initial errors will. show up as an incorrect character on the page of text. When the page is received for the second time the decoder will usually correct these display errors. Thus initial errors only show up when the page is first received. The teletext service can be considered at its limit when (typically) three or four errors are received when a page is first received. As the signal impairment increases the rate of errors increases rapidly, making the page unusable. The edge of the service area for teletext, unlike that for normal colour television reception, is relatively abrupt, the transition from good performance to unusable occurring quite rapidly, as is common for all digital systems.

The main signal degradation which causes errors to teletext reception is that due to reflections. These distort the pulse shape, causing confusion between ‘1’ and ‘0’ levels in the data stream. The teletext data signal is also degraded by noise and co-channel interference. Impulsive interference does not normally cause any problems unless it occurs during the period when the teletext data signal is being acquired by the decoder, that is during the FBI.

The accepted design target for receivers is to operate with an input signal having a decoding margin (Chapter 9, page 97) of about 25%, with only two or three errors occurring when the page is first acquired. Allowing for degradation in the tuner and IF circuits, the decoder would need to operate with an input signal having a decoding margin of about 20%.

Signal path distortion

To optimize teletext performance the distortion that can occur in the IF amplifiers and video demodulator must be minimized. The use of surface wave filters (SWAF) significantly improves the performance of IF amplifiers, compared to LC block filters.

The video demodulator is the main source of non-linear distortion. Simple diode demodulators, which respond to the envelope of the RF signal, introduce a high level of quadrature distortion. This seriously affects the data eye characteristics, causing a loss of eye-height and symmetry (Chapter 9, page 93). The vestigial side band transmission requires a detector which responds only to the modulation component in phase with the carrier. Fully synchronous demodulators give the best eye-height performance, but cost prevents their widespread use in domestic television receivers. However, ‘quasi-synchronous’ demodulators perform well with both television signals and teletext data.

Receiver de-tuning has more detrimental effect on teletext reception than that on the reception of the television video signal. A tuning accuracy of +50kHz is desirable and thus a high performance AFC system is necessary if digital tuning is not used.

To measure and assess the performance of teletext decoders accurately requires a source of signals with controlled levels of distortion, and special signal generators are produced for such tests (Chapter 9, page 99).

In domestic locations teletext performance can easily be marred by poor aerial installations and reflected signals. Experience has shown that when an aerial installation is adjusted for good teletext performance the resulting colour television picture is usually significantly improved. In general, good teletext reception is more difficult to achieve with transmissions in Band I because receiver aerials are much less directive than those used for the higher frequency bands, and at the same time electrical interference from motor cars can be more troublesome.

Multi-page decoders

Decoders capable of storing several teletext pages are now starting to be used in WST receivers. A processor is also incorporated in such decoders to control the acquisition and storage of required pages. Various strategies are being employed to reduce the access time and also to make the decoder more ‘user-friendly’.

An arrangement that does not involve any editorial function is to incorporate in the decoder a memory with an eight-page capacity and to arrange for the capture of the seven pages subsequent to that requested by the viewer. This allows the viewer to step through the magazine with a virtually instantaneous display of the requested page.

An alternative arrangement that is now specified as a WST option [13] is to use additional information provided by the editor. The editor adds this information to the page in a form of a data packet, that is a row of data that is not displayed by the decoder but is used to instruct the decoder which additional pages it should capture. This instruction is based on the editor’s anticipation of the viewer’s requirements. The decoder is also made more ‘user-friendly’ by arranging that a command from a single key press on the viewer’s handset selects both magazine and page number. Special additional information is displayed on the bottom row of the page to provide simple instructions for the viewer’s use. For example, the additional instruction row might contain four topics such as sport, news, financial and travel information. Each of these topics would have a different colour background, and buttons on the handset would have corresponding colours. When the appropriate coloured button is pressed the decoder will immediately capture the pages corresponding to this topic, e.g. sport. The names on the bottom row instruction set would then change to, for example, football, cricket, tennis and swimming. Again, when the appropriate coloured button is pressed the pages associated with that particular sport are captured ready for display.

This technique has been developed as part of World System Teletext (WST) and is now called Fastext. The ‘tree’ structure of the Fastext page is illustrated in Figure 5.6. The editor can provide an escape from the branches by making one of the options (usually no. 4) lead back to the main magazine, or the viewer can use the page numbers.

Figure 5.6 Fastext tree page structure

An alternative system developed by the Institut fur Rundfunk-technik (IRT) [15] in Germany, that does not require data packets is called TOPS. In this system the pages and magazines are arranged on a topics basis. A special control page, or ‘table of pages’ (TOPS) page, which defines all the other pages, is also transmitted. Each 8-bit character position in this TOPS page defines the category of each of the other transmitted pages. The TOPS page is captured and held in the non-display decoder memory. Its function is to instruct the decoder processor as to which pages should be acquired ready for display in conjunction with the viewer’s instructions received via the remote control system.

The magazines are organized using three basic page types: major topic pages, sub-topic pages and information pages. The handset has three corresponding buttons, together with one for reverting back to the previous page. When the receiver is first switched on, the decoder is programmed to display the first major topic page. The TOPS buttons on the remote control handset cause the decoder to display the next information page, or the next sub-topic page, or the next major topic page. To guide the viewer, the titles of the next sub-topic and major topic page are displayed on row 24. These titles are transmitted in additional information tables, which are contained in the nine information pages added to the transmission. The location of these information pages is contained in the page linking-table which forms part of the special TOPS control page that is used by the processor. The structure of the ‘TOP’ database is illustrated in Figure 5.7. Page and magazine numbers are also used so that the teletext service can be used with receivers that are fitted only with conventional decoders.

Figure 5.7 ‘TOP’ data base structure

The ideal receiver arrangement would be for the decoder to store the complete magazine of pages so that any page would then be instantly available, or, if storage is limited, for the processor to be programmed to automatically store the pages of the viewer’s choice.

Teletext adaptors

Teletext adaptors are designed so that they may be added to a conventional television receiver to provide a teletext facility. The teletext adaptor normally has its own tuner, IF amplifier and demodulator driving a teletext decoder together with a remote control system. In fact it incorporates virtually all the circuits contained in a television receiver with the exception of the CRT and the associated time-bases. The RGB outputs from the teletext decoder are re-coded into a standard colour television signal (PAL, NTSC or SECAM). This signal is then modulated on to an RF carrier so that it can be fed into the aerial socket of a conventional television receiver. Such adaptors are not ‘user-friendly’ in that when the television channel is changed the decoder channel selector also has to be changed in sympathy, and a second remote controller is usually required for the adaptor. The requirement for two remote controllers can be eliminated by remodulating the video and sound signals onto the output carrier of the adaptor so that the television receiver then only needs to be tuned to the adaptor’s output signal. Alternatively, if the receiver has a suitable video input it can be driven with RGB signals from the adaptor. In this way the adaptor provides the remote control of the television receiver but the video signal suffers in the re-modulation process when an RF feed is used to drive the receiver.

In general, the main limitation of teletext adaptors is the need to re-code the RGB signals from the teletext decoder into a standard colour coded television signal when no RGB input facility is available. Text signals are usually coloured and have fast edges. When such signals are converted into a coded colour television signal the bandwidth of both the colour and luminance compo-nents are limited, which causes a very significant loss of legibility in the final displayed text. Teletext adaptors have not proved popular due to the relatively high cost, and limited performance due to the re-modulation processes.

Control of video recorders by teletext

A standard video recorder contains a tuner with an AFC system, an IF amplifier and demodulator together with a remote control system. The output to the receiver is either an RGB signal or a remodulated RF composite video signal. The incorporation of a teletext decoder into the recorder therefore allows the unit to be used as a teletext adaptor with very little additional cost.

When the recorder is set up to record a programme at a specific time, it is assumed that the programme schedule will run according to the advertised times. In practice published schedules are often indicated only in 5-minute increments and variations may occur at the time of transmission, which means that the programme as recorded may be incomplete. This problem can be overcome if the broadcaster provides a control signal to activate the recorder in step with the broadcast programme. A signal for this purpose must be programme related in real-time that its presence or absence will activate the recorder, and it must also contain a code to relate it to the particular programme.

Such a control signal can be carried as teletext data within Packet 8/30, and very little additional circuitry is required in a video recorder that contains a teletext decoder to make use of this facility. Two data formats for such control signals are incorporated into the WST Specification [13]. The control code for either format is contained in Packet 8/30 and the data is maintained for the duration of the programme. In Format 1, the control code is a special 16-bit binary code number assigned to the programme by the broadcaster for each programme. This number is entered into the recorder by the user and the recorder is then left in the stand-by mode, i.e. the RF signal circuits and decoder are left switched on and tuned to the required channel. When the valid progamme identity number is received, the recorder switches into the record mode whilst the number continues to be received. This mode of operation requires the broadcaster to make available the special programme related code for each programme.

The Format 2 method has been specified by the EBU [16] to be compatible with the Video Programming System (VPS) which is already in use in a number of countries in Europe. The programme identification label is a 20-bit code which directly relates to the time of transmission as published. The user enters into the recorder this programme time in the usual way, and the recorder converts this automatically to the value which will match the identification label broadcast in Packet 8/30 for the duration of the programme. The programme can, in fact, be broadcast at any time as the recorder will only automatically record the programme when the appropriate Packet 8/30 is received. It is envisaged that where fully integrated or linked television receivers and recorders are available the special programme code will be generated from the teletext television programme guide pages. The user will have a cursor control facility on the keypad and an ‘enter’ button. When the cursor is moved to the required programme time, pressing the ‘enter’ button will cause the corresponding programme code to be generated, which together with information concerning the television channel etc. is entered automatically into the recorder. The recorder control logic can assist the user by drawing attention to potential time conflicts between recordings and also the duration of the tape required. At the broadcasters’ premises, the necessary programme information for generating the Packet 8/30 code is automatically provided to the teletext system from the programme scheduling computer.

In-vision systems

A teletext system can provide a source of input pages for an information service on either a broadcast or a cable television network. This arrangement is generally referred to as an ‘in-vision’ facility. The teletext page is decoded to RGB signals and re-coded into a PAL (or SECAM, or NTSC) colour signal so that it can be received by normal television receivers. The different pages are transmitted cyclically, each one being transmitted for a display time of about 10 to 15 seconds depending on the content. A series of pages can also be transmitted very rapidly so as to provide an animated display sequence in the programme. The pages can be specially created or the service can use any normally available teletext pages. The teletext decoder is equipped with a processor and page memory of adequate capacity for the service, typically 30 to 100 pages. The processor is programmed to select the required pages from the teletext input and hold them in memory ready for transmission. Each page is then ‘enabled’ for transmission for the required display time. Simple in-vision decoders usually use the same display time for each page, but more complex decoders allow the display time for each page to be varied and this is set when the page sequence is chosen for the service.

The wideband RGB signals are matrixed by the PAL coder into a wideband luminance signal and two narrowband colour-difference signals. Decoded teletext signals have fast edges, as in a receiver they are fed directly from the decoder to the video output stages. If the RGB output signals from a teletext decoder were coded into a composite video signal, transmitted over a broadcast or cable system, received, demodulated and decoded into RGB signals they would suffer considerable degradation. This degradation of the signal is acceptable for normal television pictures but it makes a text display very difficult to read. The vertical elements of a character consists of very narrow pulses which are reduced in amplitude and have very poor coloration; whereas the horizontal elements, which are represented by much wider pulses, are reproduced at full amplitude with good coloration. Furthermore, the reproduced saturation of the page background colour may not be optimum for the text colours.

The legibility of the text page can be greatly improved by text enhancement before the PAL coder. The width of the narrow vertical pulses, which form the vertical elements of the characters, are stretched to reduce the bandwidth, and the saturation levels of the background colours are reduced. The character shapes available from more recently produced decoder circuits have wider vertical elements, therefore this effect is less pronounced. The background colours can be reduced in amplitude by the processor software. When adjustments are being made to an in-vision decoder to improve the legibility of the final pages, it is important that the pages are viewed on a television receiver using an RF input, so that compensation can take into account all the potential distortions that occur in the complete system.

Next