>>TV TRANSMISSION >> SINGLE FREQUENCY NETWORKS FOR DIGITAL VIDEO BROADCASTING

by JM Fernndez, J Capdevila, R Garca, S. Cabanillas, S Mata, A Mansilla and JM Fernndez Engineering R&D - RETEVISION S.A., Spain

ABSTRACT
This paper introduces the Terrestrial Digital Video Broadcasting DVB-T, stating its innovative aspects and its major advantages for data broadcasting, particularly TV broadcasting. It also presents the experimental DVB-T network built up by Retevisin in the framework of the Spanish VIDITER project (Terrestrial Digital Video) and the European ACTS VALIDATE (Verification and launch of Integrated Digital Advanced Television in Europe) and ACTS MOTIVATE projects (Mobil Television and Innovative Receivers). The experience and some of the results of the different tes carried out by Retevisin are afterwards discussed.

WHY DIGITAL TV BROADCASTING?

Generally, the digital technology presents some major advantages in baseband efficiency, flexibi and RF performance that make its use very attractive to broadcasters. Effectively, when addressing to TV broadcasting, the digital signals are more robust and the spectrum use is more efficient; more than one program may be broadcasted using the same bandwidth and having even better picture quality. Moreover those digital signals are easier to process and more computer friendly. The only issue is how long will it take to completely change the technology considering the large number of analogue TV receivers worldwide. Some transition period will probably be started in which both technologies coexist (broadcasting the same TV contents in analogue and digital, ofte called SIMULCASTING, and some additional only-digital contents).

WHAT IS DVB?
The Digital Video Broadcasting Project (DVB) is a marked-led initiative to standardise digital broadcasting worldwide. It is formed of 240 members from more than 35 countries, in which the are representatives of broadcasters, manufacturers, network operators and regulatory bodies. The DVB was formed in September 1993 and along these years has been producing several syste specifications that have become standard in organisms as the ETSI (European Telecommunicatio Standard Institute) or CENELEC (European Committee for Electrotechnical Standardisation). The DVB family of standards The DVB has been producing different system specifications including satellite: DVB-S [EN 30042 cable: DVB-C [EN 300429], terrestrial: DVB-T [EN 300744], microwave: DVB-MVDS [EN 30048] and DVB-MMDS [EN 300749], community antenna: DVB-SMATV [EN 300743] and others. The key word of the DVB standards is interoperability , all of them are part of a family of system that make use of maximum commonality in order to enable the design of synergetic hard- and software. The DVB transmission systems offer a pipe for data containers. They are transparent for SDTV (Standard Definition TV), EDTV (Enhanced Definition TV) and HDTV (High Definition TV), for aud at all quality levels and for all kinds of general data (multimedia data broadcasting).

All the specifications are based on the selection of the MPEG-2 (Moving Pictures Experts Group) f coding audio and video and for the system level.

DVB-T TERRESTRIAL DIGITAL VIDEO BROADCASTING

The DVB-T system for terrestrial broadcasting is probably the most complex DVB delivery system The key feature of the system is the use of COFDM (Coded Orthogonal Frequency Division Multiplexing). This is a very flexible wide-band multicarrier modulation system that uses differen levels of forward error correction, time and frequency interleaving and two level hierarchical channel coding. Basically, the information to be transmitted is split into a given number (2k 1705 or 8k 6817 modulated carriers with individual low bit rate, so that the corresponding symbol time becomes larger than the delay spread of the channel. A guard interval (1/4, 1/8, 1/16, 1/32 of the symbo time) is inserted between successive symbols to avoid intersymbol interference and to protect against echoes. Depending on the channel characteristics, different parameters (sub carrier modulation QPSK, 16QAM, 64QAM , number of carriers 2k, 8k , code rate of inner protection, guard interval an modulation parameter - a) can be selected obtaining different operation modes . Every mode off a trade-off between net bit rate and protection of the signal (against fading, echoes, etc.). Depending on the selected operation mode, 60 different net bit rates could be obtained ranging from 5 to 32 Mbps. The selection of the COFDM modulation system presents two major advantages that make its us very interesting to terrestrial digital video broadcasting:

COFDM improves the ruggedness of the system in the presence of artificial (long distance transmitters) or natural (multiple propagation) echoes. Actually, the echoes may benefit instead of interfere the signal if they fall inside the guard interval. On the one hand, COFDM provides a considerable degree of immunity to narrow-band interferers as maybe considered the analogue TV signals; and on the other hand it is seen those analogue signals as white noise, therefore not interfering or having little effect upon them.

All these characteristics enable a more efficient use of the spectrum (possible use of the so-calle taboo channels, which usually are the only available ones to start new DVB transmissions); and t introduction of Single Frequency Networks (SFN). Moreover, portable and mobile reception of DVB-T signals is possible. One efficient way to achiev that is by using hierarchical transmissions, in which one of the modulated streams (so-called HP High Priority stream), having higher protection against errors but reducing its net bit rate, is use for portable and mobile reception; while the other one (so-called LP Low Priority stream), havin lower protection and higher bit rate, is used for fixed reception. The ACTS MOTIVATE project, in which Retevisin participates, is currently addressing such issue demonstrating and assessing the mobile and portable reception of DVB-T and developing algorithms and models for new enhanced receivers optimised for such reception conditions.

MULTIMEDIA AND INTERACTIVITY

Nowadays the importance for the broadcasters of offering new added value services, especially multimedia and interactivity, is out of question. The number of applications is continuously growing and evolving. Among them: pay per view, NVoD, video on demand, home shopping, home banking, Internet access, etc.

Most of those interactive services are asymmetric; the user expects a great amount of informatio (several Kbps or even Mbps) but request this information through a low speed return channel (fe Kbps). DVB provides network independent protocols together with a full set of network dependen return channels (e.g. PSTN ISDN, DECT, GSM, etc.). The advantage of DVB transmissions is that they do not distinguish between data, video or audio may even be used to broadcast data which itself incorporates audio and video as some Internet pages do). Besides, DVB-T provides the extra advantage of joining portable and mobile reception the previously mentioned characteristics. DVB data profiles DVB foresees four ways of data broadcasting depending on the necessities:

Data piping: asynchronous, non-synchronised, end to end data delivery. Data streaming : streaming oriented, end to end delivery of asynchronous, synchronous or synchronised data. Multiprotocol encapsulation : data services that require the transmissions of datagrams (as the ones of TCP-IP). Data carousels: data services that require periodic transmissions of data modules.

SINGLE FREQUENCY NETWORKS

Traditionally, the analogue TV broadcasting had to face the problem of co-channel interferences prohibiting the re-use of the same channel over considerable distances. This results in an extrem inefficient use of the spectrum. As shown in Figure 1, in conventional 9-frequency layouts, each channel is prohibited over approximately 89% of the land area.

Fig. 1. MFN frequency planing for Conventional Analogue TV An alternative to those Multi Frequency Networks (MFN) is to use a set of transmitters spread throughout a given territory (a city, a region or even a country) temporally synchronised and transmitting at the same frequency. Such configuration is called Single Frequency Network (SFN The advantages are enormous in terms of spectrum efficiency. Whereas in analogue MFN a singl SDTV program was transmitted over 9 RF channels, now more than one program could be broadcasted using a single RF channel. 9 Times less spectrum is used than in MFN! Or, 45 times more programs (assuming 5 SDTV programs per channel) can be broadcasted using the same spectrum! Moreover, taking advantage of the beneficial effect of the echoes inside the guard interval, less power would be needed in locations on the verge of the coverage area of two neighbouring transmitters, signals coming from both would contribute to improve the overall carrier to noise ratio. Shadowed areas can also be served by direct reamplification using a co-channel retransmitter of called gap filler. A drawback of SFNs is that the flexibility of dynamically replacing the contents of a program is lo Effectively, all the transmitters of a SFN must broadcast the same contents in the same moment time.

SFN Constraints Following the main constraints that the SFN operation introduces into the network will be briefly assessed: Synchronisation constraints All the signals broadcasted by the transmitters of a SFN must be synchronised in terms of frequency, time and bit. The frequency synchronisation requires that a common reference oscillator shall drive all cascade oscillators within each transmitter. The time synchronisation requires that each transmitter shall broadcast the nth symbol at Tn 1 (where Tn denotes the ideal instant for the nth symbol to be transmitted). The bit synchronisation requires that the same symbol shall be transmitted at the same time. Therefore all carriers shall be identically modulated. Hence the same bits should modulate the sa kth carrier. In order to fulfil those requirements, DVB-T provides the MIP specification [TR101191]. By mean of a SFN adapter, located at the output of the Transport Stream (TS) generation process, Megaframe Identification Packets (MIP) are inserted periodically into the TS. Modulators use thos additional packets for the time and bit synchronisation. The synchronisation mechanisms are based on the existence of two global external references: a frequency reference of 10 MHz and a time reference of 1 pps (pulse per second); and example o global system providing such references is the GPS (Global Positioning System). Transmitter requirements A complete set of specific requirements for transmitters have been identified within the framewo of the ACTS VALIDATE project.

Frequency stability . Each carrier should be transmitted in a frequency within the interval f (Df/100). The transmitter needs the external frequency reference for synchronisation in SF (10 MHz). Oscillators phase noise . One of the most limiting factors found during the tests of transmitters and transposers was the phase noise of the oscillators. In some cases transmitters which are perfectly suitable for PAL transmissions appear to be of no use in th case of DVB-T because of phase noise. Output back-off. The maximum power that can be obtained by a given transmitter is limite by the nonlinearity effect of the amplifiers, consequently affecting to the quality of the reception. For each kind of transmitter used for DVB-T broadcasting there is a fixed output power value that maximises the coverage area. Transmitting below this value, the capabili of the equipment are not fully used, and transmitting above, we obtain added system implementation losses, which are greater than the expected coverage gain. Typical back-o values are in the order of 6 dB or even more.

Professional Gap fillers requirements Similar kinds of requirements stated for the transmitters are also applicable to professional gapfillers or transposers. Though, in this case, the economic point of view should be taken into account. A transposer should be much cheaper than a transmitter, and consequently the overall requirements should not be so restrictive as for transmitters. An additional requirement for the gap fillers is the maximum allowable gain in a given site. This gain is limited by the feedback loop gain, which is basically determined by the input/output ante isolation. It has been proven in field tests that isolations above 100 dB, although difficult, can be feasible. The maximum gain of the transposer is then limited to the isolation value minus a secu margin needed to avoid instability problems that cause strong additional implementation losses a can even produce the unavailability of the system in the area covered by the gap-filler. A typical

security margin value can be around 20 dB. Primary distribution network aspects The primary distribution network addresses the transport of the TV signal in whatever format to transmitter sites for its broadcasting. The TV signal may be transported in digital format (i.e. using the MPEG-2 Transport Stream) or i analogue format (i.e. modulated according the DVB-T specification). Decentralised generation of the DVB-T signal Addresses the transportation of the digital MPEG-2 Transport Stream through the primary distribution network and then modulating the signal inside each transmitter site. The primary network may use fixed terrestrial (e.g. optical fibre, radio links) or satellite links. Various technologies or combination of them can be applied for such purpose (e.g. ATM, PDH, SDH, DVB etc.). Two major advantages arise from the use of this method:

Flexibility: further levels of MPEG-2 multiplexing may be included, for example to provide regional programme variations, although in SFN this feature is not applicable. Signal quality: after the primary distribution the carrier to noise ratio is roughly preserved.

As drawbacks several DVB-T modulators are needed (one in each transmitter site) which increas the overall cost of the network and imply the need to synchronise them (see SFN Constraints). Another problem could be the jitter introduced by the multiplexing and remultiplexing processes Centralised generation of the DVB-T signal Addresses the modulation of the TV signal according the DVB-T specification at a central point an transporting through the primary distribution network the analogue COFDM signal. The primary network may use fixed terrestrial (e.g. radio links) or satellite links. The major advantage of using this method is that the number of DVB-T modulators in the netwo is reduced, being even possible one single modulator for the entire network. Most of the synchronisation problems of the SFN disappear, especially when using a satellite link which not even static delays (as the ones introduced by terrestrial radio links) are introduced by network. Notwithstanding this method looses flexibility with respect to dynamically replace programs at th remultiplexer sites. Anyway this is not affecting in the case of SFN where no remultiplexing is allowed. Another very important problem is that generally the final signal looses quality (i.e. the C/N is degraded in the primary distribution and this degradation can not be recovered) which translates into less coverage area. Secondary distribution network aspects The secondary distribution network addresses the broadcasting of the TV signal from the transmitter sites to the final user receiver. As previously mentioned, the DVB-T provides several operation modes to adapt the signal to the radio channel characteristics. Every operation mode offers a trade-off between net bit rate that m be allocated to the TV programmes and protection of the signal against echoes, noise, fadings, e

Table 6. Maximum transmitter sites separation depending on the operation mode for SFN (8 M channel) An important issue for SFN is the use of 8k modes (i.e. 6817 sub carriers), especially for wide ar SFN (see Table 1), since this mode allow longer echoes (they have longer guard intervals) than t 2k modes.

SPANISH REGULATION
Following the Spanish Technical Regulation (issued the 16th of October 1998) on Digital Terrestr Television will be briefly described: Channel allocation The following frequency bands are reserved for the DVB-T service: a) 470 to 758 MHz (ch.: 21 to 56) b) 758 to 830 MHz (ch.: 57 to 65) Completely available from 31 October 1999. c) 830 to 862 MHz (ch.: 66 to 69) Completely available from 30 June 1999. Service planning Four SFN national coverage channels, carrying at least four services in each channel, will be set in the 830-862 MHz band. One channel with national coverage and regional re-multiplexing (regional SFN), carrying at leas four services, will be set up in the 758-830 MHz band. Simulcast with the analogue TV. One SFN regional coverage channel, carrying at least four services, will be set up in the 758-830 MHz band. Simulcast with the regional analogue TV. N channels with local coverage will be set up in the 758-830 MHz band. The 470-758 MHz band will be used for analogue TV transmissions, MFN and local broadcasting until the analogue switch off, the 1st of January 2013.

Fig. 2. Phases of the establishment of Terrestrial Digital TV in Spain Time scale

The Figure 2 shows the four phases of the coverage plan. The coverage is thought in terms of population, not territory.

Phase I: SFN national channels, 50% coverage, 12 months duration from the 30th of June 1999. Phase II: National with regional re-multiplexing channel, 50%, 8 months duration from the 31st of October 1999. Phase III: all channels, 80% coverage, 18 months duration from the 30th of June 2000. Phase IV: all channels, 95% coverage, 10 years duration from the 31st December 2001.

Operational modes of DVB-T The technical specifications of the Digital TV transmitters will follow the 8 MHz, 8k mode of the European Telecommunication standard EN 300 744.

THE RETEVISIN EXPERIMENTAL NETWORK

The experimental DVB-T network of Retevisin was built in the framework of the Spanish VIDITE project and the European ACTS VALIDATE project. It consists of two transmitters; one is located Torrespaa (Madrid) and the other in Navacerrada (separated around 50 Km). The DVB-T emitte power is 900 W and 200 W respectively. The network also includes a professional gap filler (emitting 10 W) located 5 Km away from the Torrespaa transmitter. Preliminary assessment of the network was carried out from February 1996 until November of th same year. Afterwards the network, configured as a Multi Frequency Network (MFN), was tested until March 1997. The objective was to gather data to establish comparisons with future SFN measurements. In parallel, several laboratory tests have been performed to verify the main parameters of the DVB-T specification. The current SFN configuration, in channel 26, was set up in March 1998. Field tests are being do since then obtaining very encouraging results for the near future establishment of terrestrial digi broadcasting services in Spain. The main DVB-T characteristics for SFN, always applied to the Spanish case (in terms of legislati environment and reuse of existing broadcasting sites), were assessed in urban, suburban and ru areas. In the framework of ACTS MOTIVATE project, the follow up of ACTS VALIDATE, more tests are foreseen during 1999 to assess portable and mobile reception.

Fig. 3.Retevisin Experimental Digital Terrestrial TV network

NETWORK TOPOLOGY
Retevisin has installed a complete digital terrestrial TV chain, compliant with DVB-T, and made of four parts:

Production TV studio and master control room. Source coding, data insertion and programme multiplex. Primary distribution network (so-called gathering/transport network). Secondary distribution network (so-called broadcast network).

The production TV studio and the master control room are both located at the Retevisins Laboratory premises, in Pozuelo de Alarcn (Madrid), and their role is to feed the experimental network with a programme bouquet (four programmes) embedded in a Transport Stream (TS). A additional data channel is also inserted into the TS in order to test data broadcasting services, lik Internet access. Primary distribution network The primary distribution network has been designed to transport the signal from the Retevisins Lab in Pozuelo up to the two transmitter sites: Torrespaa and Navacerrada. The transport network was carrying the MPEG-2 TS from the Retevisions Lab to Torrespaa via a optical fibre link. From this point to the second transmitter site (Navacerrada), a digital radio link was used. An important issue of the trials was related to the primary distribution of the signal, for that purpose, besides the previously mentioned methods, SDH and PDH transport of the MPEG-2 TS were both assessed. Moreover, analogue primary distribution of the signal (decentralised generation of the DVB-T signal) was also addressed during the trials; a transponder of the Hispasat satellite was used to provide the transmitter sites with the DVB-T signal already OFDM modulated. Although the resul were satisfactory this option presented a worse performance in terms of carrier to noise ratio tha the digital distribution one. Secondary distribution network The secondary distribution network has been designed to work either as a MFN (dual frequency) as a SFN in channel 26. As previously mentioned, the network is made out of two transmitters a a professional gap filler (or transposer). With the current configuration, the generation of the DVB-T signal is not centralised, therefore at each transmitter site there is a modulator equipped with a GPS for the frequency, time and bit synchronisation. The DVB-T modulators are able to reconfigure themselves using specific data sent within the MPE 2 Transport Stream, following what has been established in the MIP specification for SFN (TR 101191).

SUPPORTING LABORATORY TESTS

Hereafter some selected results of the most relevant laboratory tests carried out up to date are presented. Reception in AWGN Figure 4 represents the theoretical and measured implementation losses with the presence of Additive White Gaussian Noise (AWGN) for 8K, 64 QAM modes. It can be observed that the actua values are always lower than 3 dB respecting the theoretical ones.

Nevertheless the measured noise factor value for the professional equipment is greater than the expected value, this value (9 dB) has been taken as reference for the specific coverage studies, recommended by the results obtained in the ACTS VALIDATE project.

Fig. 4. C/Nmin in AWGN Protection ratios The DVB-T signal presents ruggedness in front of high power PAL signals in both cochannel and adjacent channels. The measured co-channel protection ratios (PR) against PAL signal for the mode 8K, sub-carrier modulation 64 QAM, code rate 2/3, guard interval 1/4 is about -2 dB. This means that the DVB-T signal could cope in the limit with co-channel PAL interferences if the peak sync power of the PAL signal is not more than 2 dB above the power of the DVB-T signal. The PR measured values for adjacent channels (in the order of -24 dB) are more than 10 dB wor than the foreseen for commercial equipment. With regard to interference from other DVB-T signals, the co-channel PR values are approximate the same ones as the C/Nmin measured values for AWGN channel. Multipath propagation The performance of DVB-T against echoes has been broadly assessed and reported previously by different sources in the ACTS VALIDATE and MOTIVATE projects. The implementation losses added (DC/N) for 0 dB echoes within the guard interval were not grea than 8 dB in the operation mode previously mentioned (i.e. 8k, 64QAM, 2/3, 1/4). The feasibility of receiving DVB-T signals in typical Rayleigh channels has been also tested. This issue is especially important in the case of portable and mobile reception. Non-linearity effects The behaviour of the system implementation losses due to non-linearity effects in transmitters a professional gap fillers has been measured in different equipment and DVB-T operation modes. Figure 5 rates (8k, 64QAM, 1/4 of guard interval).

Fig. 5. Non-linearity effects in DVB-T Feedback in SFN When using professional gap fillers in a SFN (transposers) to cover shadowed areas, it is up to th network designer to select the gain of the different transposers in order to increase the coverage but avoiding the negative effect of a high gain due to the feedback limitations. Figure 6 shows the behaviour of the implementation losses due to positive feedback measured in typical transposer.

Fig. 6. Use of transposers in SFN Effect of Phase Noise in Local Oscillators Figure 7 shows the phase noise measured in the transposer local oscillator. It can be observed th the curve (the upper one in the figure) follows what was foreseen (the lower curve in the figure t has been issued by a signal generator). Therefore the spectral mask proposed by the ACTS VALIDATE project in ref. 9 has been verified.

Fig. 7. Phase noise in Local Oscillators (LO) Other measurements Other laboratory measurements have been performed to assess the DVB-T specification in a real environment, some interesting aspects treated were: feasibility of distributing DVB-T signals through MATV installations (Master Antenna Television), technical feasibility of using domestic ga fillers in SFN networks and an evaluation of the behaviour of the demodulators in presence of impulsive noise. Laboratory tests of hierarchical modulation for portable and mobile reception have been recently started and will continue during 1999.

FIELD TESTS
Hereafter some selected results of the most relevant field tests carried out up to date are presented. The field trials were performed in the Retevisin DVB-T network settle in the area of Madrid. The network was configured as MFN and as SFN transmitting in channel 26 (514 MHz).

A mobile unit to obtain measurements in different areas (urban, suburban and rural) was equipp with the following elements:

A telescopic directional antenna (10 m) An omnidirectional antenna (1,5 m) A GPS receiver DVB-T demodulator MPEG-2 decoder TV set BER meter Field strength meter Spectrum analyser Laptop PC

The following operation modes were mainly considered during the trials:

8k FFT, 64 QAM, 2/3 FEC, 1/4 GI 8k FFT, 64 QAM, 3/4 FEC, 1/4 GI

Received spectrum The frequency spectrum of the received DVB-T signal was measured using an ESVB and scanning with a resolution bandwidth of 120 KHz and a step of 50 KHz. The standard deviation (s) of the sampled values of the spectrum within the nominal bandwidth gives an indication of the type of transmission channel. Table 2 shows the assumed classification channels according to its s.

Figures 8 and 9 show the spectrum received in Ricean (typical in rural and sub-urban environments) and Rayleigh channels (typical in urban areas).

Fig. 8. DVB-T Frequency spectrum in Ricean channel (s =2.04 dB)

Fig. 9. DVB-T Frequency spectrum in Rayleigh channel (s=3.48 dB) Some areas of Madrid were presenting a strong PAL co-channel interference, however in some cases the DVB-T equipment was still able to decode and present the transmitted TV images. The received spectrum in those points looks as shown in figure 10.

Fig. 10. PAL co-channel interference in the DVB-T spectrum Correction factors for 70 %, 90 %, 95 % and 99 % of locations In order to calculate the minimum field strength for planning purposes (fixed reception), the following field measurements related with the correction factors for locations in a small area (typically 100 m x 100 m) were done. It should be noted that these correction factors include effects that are not considered by propagation models (e.g. multipath). Test procedure Having the Yagi antenna placed at the top of a 10 meters mast, the mobile unit was following (a Km/h uniform speed) a linear 100 m path. Meanwhile the measurement equipment was taking more than 1000 samples (one each 9.5 cm). The DVB-T modes examined in this test are shown Table 3:

Table 3. DVB-T modes tested Results Figure 11 shows a field strength profile corresponding with a real data file. The X-axis represents

the different points tested along the 100 m path and the Y-axis shows the measured voltage in dBV. From this profile, it is possible to compute the median value (V50%) and additional signal levels (V70% and V95%) where Vx% is defined according the following expression:

where f(v) is the probability density function The correction factors (Cx%) are defined as the difference between the signal level at V50% and the signal level at Vx%. The standard deviation is computed for each set of measurements (~1000 samples).

Fig. 11. 50%, 70% and 95% coverage levels In MFN the estimation by means of the theoretical value (see ref. 10), i.e. assuming a log-norma distribution of the field strength for planning purposes, is on the measured values. Figure 12 shows the field strength distribution of the values measured by the mobile unit in suburban areas subtracting the mean value corresponding to each route so as to emphasise the field strength dispersion.

Fig. 12. Field strength distribution in suburban areas (MFN) However in SFN the field strength can not be assumed to follow a log-normal distribution. For low coverage factors the estimation is already quite good but as coverage increases the difference between the measured and computed values enlarges, though, in principle, it depends on the network structure and the considered reception location.

Figure 13 shows the frequency distribution of the measured and computed values concerning to suburban areas in SFN. It should be noticed that the plotted values have been obtained subtract to the field strength the mean value corresponding to the particular route.

Fig. 13. Field strength distribution in suburban areas (SFN) Analogue satellite transmissions This test was carried out to verify the feasibility of using an analogue satellite link as primary distribution network for DVB-T signals. The centralised generation of DVB-T signals and its following analogue distribution via satellite could imply certain advantages for SFN, among them:

Possibility of using a single DVB-T modulation system for the entire network. Minimisation of the requirements associated with the synchronisation of the modulation process. Faster deployment of the broadcasting network. All transmitter sites straightforwardly covered.

But also certain important disadvantages, among them:

Local remultiplexing not allowed. Although this is not required for SFN. Loss of C/N, which translates into a lower coverage area than in the digital case.

Test procedure The transmission of the DVB-T signal was performed by means of the FM modulated system used the analogue television and through the Fixed Satellite Services (FSS) of Hispasat. The DVB-T signal FM modulated by a 70 MHz IF carrier was conveyed through a 36 MHz band-pa filter in order to bind the range frequency to the transponders bandwidth. Next, this signal was shifted to the transmission frequency by means of an up-converter so as to boosted by a travelling wave amplification system (TWT). The signal received from the satellite was then amplified using a low noise amplifier (LNA). Afterwards it was converted to an intermediate frequency (IF = 70 MHz) by a down-converter whose output was connected to a noise generator in order to change the C/N ratio at the FM demodulator input. In this way it was possible to check the noise margin reduction that the link w introducing. To assess the effect of the satellite, two different tests were done; the first one, called satellite loop, was as explained above, the second, called FM loop, consisted in modulating and demodulating in FM the DVB-T signal without the transmission to the satellite.

The DVB-T modes tested were the following ones: 8k, 64 QAM, 1/4 GI and FEC 2/3 and 5/6. Results Figure 14 shows the behaviour of the link (BER vs. C/N) in the cases of FM loop and satellite loop It should be noticed that the satellite loop not only implies a degradation of the C/N ratio but als change in the slope of the behaviour in comparisons with the FM loop. Other characteristics stated were:

The 25 MHz frequency deviation was admitted as the optimum value to achieve the best C/ ratio independently of the operating mode. The reduction of the noise margin due to the link was approximately 3 dB for the R=2/3 mode and 8 dB for the R=5/6 mode.

Fig. 14. BER vs. C/N Minimum Carrier to Noise for reception The minimum C/N for reception was also assessed during the field tests. For that more than 300 measurements were performed in portable and fixed reception with the help of the mobile unit previously described and using specially developed software. Figure 15 shows the fixed reception main screen, in which it is possible to visualise all the network parameters and the measurement results.

Fig. 15. Fixed reception measurement SW main screen Hereafter the results obtained in a suburban environment, using the 8k, 64 QAM, 1/4 GI, 2/3 FE DVB-T operational mode are represented. All figures represent the obtained C/Nmin in the different test points and compares those values with the theoretical ones (Ricean and Rayleigh channels). Two considerations should be taken in account: 1. 3 dB Implementation losses due to the receiver are included. 2. Theoretical values concerning SFN are under assessment.Therefore, the same estimated value are used for MFN and SFN indistinctly. Figures 16 and 17 presents the results corresponding to the case of MFN and SFN.

Fig. 16. MFN Fixed reception C/Nmin measurements

Fig. 17. SFN Fixed reception C/Nmin measurements In general, the received minimum carrier to noise ratio was as expected, the average value is we positioned between the theoretical Ricean and Rayleigh ones. The standard deviation is similar fo MFN and SFN. Future tests Other field tests will be performed during 1999 to assess the DVB-T specification.

Multimedia and Interactivity: Retevisin plans to start a new test campaign following the fi experiences in which the feasibility of such features were demonstrated in the laboratory a experimental network, particularly with a FTP and Internet access applications. Mobile Reception and Hierarchical Modulation: Mobile TV reception is an important feature the DVB-T specification that Retevisin intends to assess in depth, for that in the framewor of the ACTS MOTIVATE project, a campaign of measurements will be started in short term period. The hierarchical modulation tests are going to follow the laboratory tests carried ou recently to assess the usefulness for portable and mobile reception of such modulation scheme in a real environment.

CONCLUSIONS
The most important issues related to the DVB-T specification and Digital Terrestrial TV service planning have been assessed and demonstrated in a practical case. The Retevision experimental network has been of great importance for evaluating different aspec such as initial coverage studies, multipath channel distortions, robustness against interferences (both of digital and analogue signals), transmitter non-linear distortions, oscillator phase noise, e Moreover, an important feature of DVB-T networks, the SFN configuration, has been successfully implemented, providing very encouraging results for the future Digital Terrestrial TV regular service, foreseen in Spain in the summer of 1999. BIBLIOGRAPHY 1. ETSI, 1997. Digital Video Broadcasting (DVB); Framing structure, channel coding and modulat for digital terrestrial television. EN 300 744 2. DVB, 1997. Specification of a Megaframe for SFN synchronisation. TS 101 191 3. ETSI, 1997. Digital Video Broadcasting (DVB); Measurement guidelines for DVB systems. ETR 290 4. VALIDATE, 1998. Implementation Guidelines to DVB-T. Del. D03 of ACTS/VALIDATE. DVB - TR 101 190 5. VALIDATE, 1998. Project Final Report. 6. Caizares, P. et al, 1996. The first Spanish experience on digital terrestrial television broadcasting. International Broadcasting Convention. IEE No. 428

7. MOTIVATE, 1998. Jos M. Fernndez et al. DVB-T tests in Spain. First results. Retevision/007 8. MOTIVATE, 1998. A. Mansilla et al. Analogue satellite transmission of a DVB-T signal. Retevision/006 9. VALIDATE, 1997. Transmitter Performance Specification. Del D14 of ACTS/VALIDATE 10. CEPT, 1997. The Chester 1997 Multilateral Coordination Agreement relating to Technical Criteria, Coordination Principles and Procedures for the introduction of DVB-T. Proceedings of ECMAST 99 (Springer, Lecture Notes in Computer Science, Vol. 1629)

HOME

GO

browse a category

Copyright 2000, Sandberg Media