All-digital planning and digital switch-over

All-digital planning and digital switch-over Chris Nokes, Nigel Laflin, Dave Darlington 10th September 2000 1 This presentation gives the results of some of the work that is being done by BBC R&D to investigate the requirements for digital terrestrial television planning in the future. Whilst this work has concentrated on the idealised case of planning when analogue has been switched off, it will also look at some of the requirements for the process of switching off the analogue transmitters. This process has been called digital switch-over, but perhaps a more appropriate phrase is digital migration, since this correctly reflects the fact that the switch-over is unlikely to be abrupt, but much more likely to be a phased transition, possibly over several years. 1

All-digital planning studies Introduction All-digital networks Analogue conversions Portable coverage Digital migration strategy European context Conclusions 2 To give you a brief overview of this presentation, after a short introduction I will present the results of our investigations into all-digital networks. I will then deal with the topic of analogue conversions, and show the potential benefits these can bring. We have also started to study the topic of portable coverage - something which appears to be gaining in importance and interest. Finally I will say a little about the requirements for a digital migration strategy, and how these fit into the European context, where many countries are now starting to consider digital television services and the migration from analogue to digital. 2

Introduction PAL started in UK 1967 5 services now using over 1300 stations 1 of these launched 1997, and has limited coverage DTT launched in UK 1998 over 30 services; 6 muxes; 80 stations on air multiplex equalisation UK Government expects switchover 2006-2010 coverage and access targets 3 By way of introduction, I thought it would be interesting to remind you of a little of the history of television in the UK. The colour PAL service was launched in 1967. Although originally planned for 4 services, more recently a 5th service has been planned and launched, although the coverage is necessarily more limited than for the other 4. Altogether these 5 services are now transmitted from over 1300 stations. By contrast, the DTT service was launched in the UK in 1998, but has over 30 services contained within 6 multiplexes, but currently operating from only 80 stations. For a variety of reasons, the initial DTT frequency plan left the coverages of the 6 multiplexes rather uneven, and so there is now a process being undertaken referred to as multiplex equalisation, which aims to make the 6 multiplexes serve as closely as possible the same areas. The UK Government has announced that intends to switch over from analogue to digital broadcasting, possibly starting as early as 2006, and intending the process to be complete by 2010. It has set targets of 99.4 % for the coverage of digital services and 95% for the proportion of people that must have access to digital equipment before switch-over can proceed. However, the precise details of what is meant by these targets have still to be clarified. 3

All-digital network studies Conditions of network studied: Network of 128 transmitters using existing main stations and high power relays Radiated power is 10 db below the analogue There are NO interfering sources from neighbouring countries or other UK stations Our all-digital planning work has made use of a slightly idealised network of 128 stations. However, it is not strictly a clean sheet plan, since the 128 stations consist of the all existing main stations and high power relay stations, and the power assumed for each station has been taken as the power of the analogue station reduced by 10dB, on the assumption that by and large the same basic network will be the most practical for the conceivable future. There are no other interfering sources - all other stations are ignored. 4

Network configurations 4 configurations studied: National single frequency network (SFN) single 8 MHz UHF channel for whole country Regional SFN all stations in a region are on 1 UHF channel Sub-regional SFN main station and its relays are in SFN Multi-frequency network (MFN) like conventional analogue planning 5 For this network of 128 stations, we have considered 4 different configurations: Firstly, a national single frequency network (SFN), where all the stations use the same 8MHz UHF channel. Secondly a regional SFN, where all the stations in a region use the same UHF channel, but neighbouring regions use different channels, so that 4 channels are used in total to cover the country with a regional service comprising 13 regions. Thirdly, a sub-regional configuration, where the main station and its relays use the same channel, but the neighbouring main station within the same region uses a different channel. This again uses 4 channels to cover the country. Finally we have considered the 128 stations as a multi-frequency network (MFN), as is used for conventional analogue planning. For this every station uses a different channel from its immediate neighbours, but we limited the total number of channels used to either 4 or 6. For all of the results based on this network, planning is based on 95% locations for fixed reception, and 70% locations for portable reception - these are the figures that the Chester 97 coordination agreement recommends for good and acceptable coverage respectively. However it should be noted that the planning work within the UK has more normally planned for only 90% locations. 5

Coverage of national SFN (224 µs guard interval) 100% 90% 80% Fixed roof-top reception Ground floor portable reception Percentage population 70% 60% 50% 40% 30% 20% 10% 0% QPSK rate 1/2 64QAM rate 1/2 64QAM rate 2/3 QPSK rate 1/2 64QAM rate 1/2 64QAM rate 2/3 This slide shows the coverage that is obtained for the various national SFN s that we have investigated, using the longest possible guard interval of 224µs. Whilst this would require the use of the 8K mode of the DVB-T specification, we are not proposing that the UK adopts this mode - we are simply investigating the limits of achievable coverage. The different modes have varying degrees of ruggedness, and corresponding bitrates from 5-20 Mbit/s. The blue bars on the left show the coverage for fixed roof-top reception, and as you can see, exceptional coverage is obtained for the most rugged modes, with reduced coverage for the higher bit-rate modes. The pink bars on the right show the coverage for ground floor portable reception, and whilst again the coverage is excellent for the most rugged mode, it is interesting to note the more rapid reduction for the higher capacity modes. An interesting possibility emerges if the use of hierarchical modes is considered. A hierarchical transmission could provide a service with a rugged layer giving, for example, one programme with 5 Mbit/s for good portable coverage, and at the same time, a less rugged layer offering additional programmes with an extra capacity of 13Mbit/s. Generally a fixed roof-top antenna would be required to receive both layers. Although mentioned here in the context of SFN s, this principle applies equally to MFN s, where slightly increased capacities can be achieved. 6

Comparison of network types (64-QAM rate 2/3) Percentage population 100% 90% 80% 70% 60% 50% 40% 30% 7 µs (2k only) 28 µs (2k or 8k) 56 µs (2k or 8k) 112 µs (8k only) 224 µs (8k only) 20% 10% 0% National SFN 4 channel regional SFN 4 channel subregional SFN 4 channel MFN 6 channel MFN This slide shows the comparison of coverage for the various types of network configuration. This time, all of the systems are using 64-QAM code rate 1/2, with a payload of between 20-24 Mbit/s depending on the guard interval. As you can see, although reasonable coverage is obtained with the national SFN, better coverage is obtained with the regional and sub-regional versions, and better still with the MFNs. However, it must be remembered that all these other configurations are using 4 or 6 channels to obtain this increased coverage. Another way to look at this is that the national SFN considered so far is not yet dense enough to prevent self-interference from limiting its coverage. If a significantly more dense network were considered, this problem could be eliminated. Whilst the increase in coverage obtained with the 6 channel MFN does not appear all that significant compared to the 4 channel MFN, there is a hidden problem. With the 4 channel MFN, as the network is built up, it becomes increasingly difficult to allocate channels to new stations without causing interference to existing stations in the network. Therefore it might prove very difficult to increase the coverage much beyond that shown. With the 6 channel MFN the position is much easier, and there is plenty of scope for expansion of the network. Whilst this work has concentrated on the use of a single type of network to achieve consistent coverage, it may also be appropriate to consider a mixture of network types. 7

Analogue conversions Re-use of high power analogue assignments for digital services Chester 97 agreement allows for this with 7dB reduction Simplifies migration process Looks an attractive way forward BUT: Complex network planning required analogue network uses many rebroadcast links Difficult for viewers could use short digital test transmissions even that may be confusing! Further study required 8 We now turn to the subject of analogue conversions. An analogue conversion means the re-use of an analogue channel assignment at a station for digital transmissions. This process was agreed and allowed for within the Chester 97 multilateral coordination agreement, provided that a conversion ratio of at least 7dB is used. Since it implies the use of the same channel over the same broad area after migration, the effects of doing this are very limited, and so from a planners perspective it appears very attractive. However, this attraction hides some complexities - the current analogue network is heavily interdependent, in that most of the relay stations take their signal feeds off-air from parent relay or main stations. Therefore any change in the network will have consequences for the rest of the network, and a wholesale conversion would be difficult to plan for. Furthermore, conversions will be difficult for viewers - one day they will need to be watching analogue, and the next they will need to watch digital. It would seem likely that such an overnight switch would require a period of test transmissions in advance, to allow viewers to check the readiness of their digital equipment, and to minimise the level of problems that will need to be dealt with when the final switch-over is implemented. However, such test transmissions could still prove confusing and awkward for viewers. It is worth noting that even such an overnight switch, would probably only take place for one transmitter at a time, so the complete switch-over for the whole country would still be phased. Clearly this is an area where more study and pilot testing will be required. 8

This map tries to show why analogue conversions appear so attractive. The red area shows the approximate coverage of the analogue service in the London area from the Crystal Palace transmitter. The blue area shows the corresponding coverage for the current digital transmissions, which are operating at 20dB less power than for the analogue services to prevent interference to analogue. As you can see, there are significant areas where analogue reception is possible but digital is not. However, if the analogue transmission were converted to digital with a conversion ratio of 10dB, the digital transmissions would match the coverage of the analogue transmissions almost exactly. Analogue conversions are probably the only way that such a coverage match can realistically be achieved. 9

Portable coverage Need to know analogue portable coverage Need to define portable Ground floor or first floor Good or just watchable pictures range of scenarios could be considered 10 We have also been studying the issue of portable coverage (reception with a set-top antenna), which seems to be gaining in importance and interest. Before considering digital portable coverage, we thought we should know a little more about analogue portable coverage. However, before we can do that, we need to define what we mean by portable - do we mean reception on the ground floor or on the first floor, and are we referring to good pictures being received, or will we accept a lower standard of picture for portable, as it seems likely that many people watch under such conditions. Ultimately a range of scenarios must be considered. 10

Analogue field strength distributions (i) 100% Percentage of total population for transmitter 90% 80% 70% 60% 50% 40% 30% 20% 10% Crystal Palace Bilsdale 0% 50 60 70 80 90 100 110 Field exceeded (dbµv/m) To simplify the process of assessing a range of scenarios, we thought it might be instructive to draw a curve of the predicted field strength distributions from some typical transmitters. Two of these are shown in this graph. The graph shows the percentage of the population covered by a transmitter receiving a field strength greater than a certain value. For example, for the Crystal Palace transmitter which serves the London area, nearly 30% of people are predicted to have field strengths in excess of 100 dbµv/m. This is not surprising since Crystal Palace is very close to the population it serves, being situated in a densely populated region of South London. On the other hand, Bilsdale is situated in the middle of the North York Moors, and serves a very scattered population. Consequently, almost none of the target audience receive a field strength in excess of 100 dbµv/m. The conclusion from this exercise is that there is no such thing as a typical transmitter. Only a graph representing the entire population will suffice!... 11

Analogue field strength distributions (ii) 100 90 80 UK population Percentage of UK population 70 60 50 40 30 20 10 0 50 60 70 80 90 100 110 Field exceeded (dbµv/m).and here is such a graph! This is just the same as the preceding graph, but for the whole UK population. This can now be used to estimate percentages of population for various portable scenarios. 12

Estimated portable coverage - range of scenarios Type Location Estimated noise limited coverage Analogue (Good picture) Analogue (Noisy picture) Digital (after conversion of whole analogue network with 10dB conversion ratio) Digital (current DTT network) Ground floor 39% First floor 55% Ground floor 64% First floor 76% Ground floor 45% First floor 59% Ground floor 12% First floor 26% Coverage from current DTT network is much less as low transmission powers are used 13 The table shows the population percentages estimated from the previous graph. These are of course only noise limited estimates, and take no account of interference, or ghosting in the case of analogue, both of which can sometimes be a significant obstacle to set-top reception. The table shows that much higher values are obtained for the first floor case, and for the case when noisy pictures are assumed. The digital portable coverage, assuming that the entire analogue network is converted with a conversion ratio of 10dB, is better than the analogue coverage for good portable reception, but some way short of the coverage for noisy portable reception. This suggests that today s portable reception can largely be matched by DTT after switch-over, if such an analogue conversion is used, except in the case of viewers used to very noisy analogue pictures today. The current DTT network, however, provides a much lower coverage. This is mainly due to the fact that the transmission powers are in general about 20 db less than the analogue stations, and in some cases are further restricted by the need to protect analogue transmissions with directional transmitting antennas. 13

Migration strategy - requirements Must be developed with Government and industry Must be co-ordinated with all of Europe Likely to start by extending current DTT network where this is cost effective Likely to use analogue conversions at switch-over to improve coverage main stations: large coverage gains (e.g. Crystal Palace) relays: bring DTT to areas not previously served satellite and cable also complement DTT coverage Should maximise availability of services Should minimise costs to broadcasters and viewers Needs rapid penetration of digital ready equipment 14 I will now deal with some of the requirements of a digital migration strategy. Clearly the strategy must be agreed with both the Government and the whole of the broadcasting industry, but it will also need to be co-ordinated with the requirements of our European colleagues. It seems more than likely that the basis for the migration will be the current DTT network, since that already provides quite good coverage to many areas. However it would also be desirable from a planning point of view to extend that network where possible, and where it is cost effective to do so. Given the attractiveness of analogue conversions it seems likely that the migration strategy will want to make use of this technique. This is likely to be for both main stations, for large coverage gains, as I have demonstrated with Crystal Palace, and for some relays, to bring DTT to areas not currently served. Of course, satellite and cable services also complement the DTT coverage. The migration should aim to maximise the availability of all services to all viewers, and in particular the free-to-air services. However, at the same time it should minimise the costs of migration to both broadcasters and viewers. To meet these requirements, a rapid availability and take up of digital ready equipment will be required, and that includes video recorders, as well as televisions and set-top boxes. 14

European context UK switchover plans must be taken within European context Range of requirements from different countries fixed or portable / mobile network structures (SFN / MFN or possibly cellular) DVB-T mode Likely that a mixture of plans will need to be accommodated Proposed that ITU regional planning conference will be needed first session: planning requirements about 2003 second session: the plan about 2005 15 As I mentioned, the migration strategy must be co-ordinated with the requirements of our European colleagues, and we are working together with them to study this problem. It is already clear that there are a range of different requirements from different countries, ranging from fixed services to portable and mobile services, with different network structures, using SFNs, MFN s and in some cases even cellular broadcast networks. These will all require different modes of the DVB-T system to be used. It is therefore likely that a mixture of these plans will need to be accomodated, and to achieve this it is proposed that there should be an ITU regional planning conference. Although there are no definite plans yet, it seems likely that this will consist of a first session to define the planning requirements around 2003, and a final session around 2005 to construct the plan itself. This could ultimately be an extension of the existing Stockholm plan for analogue transmitters, which was drawn up in 1961, or may be a completely new plan. 15

European planning options Three main options can be considered: conversions attractive: transition can be phased not optimum; limited scope for SFN s extension of current DTT networks further extension possible after analogue switched off also not optimum complete re-plan likely to be fairest, and most efficient use of spectrum but transition could be very difficult new transmitting equipment and receiving antennas may be needed 16 From the discussions that have been held so far, it appears that there are three main planning options emerging for digital migration: the use of conversions, as I have already described. These are even more attractive in the European context, because they would allow for a phased transition, with only bilateral agreements required for each stage to proceed. However, it seems likely that the widespread use of conversions will not lead to an optimum plan, and there would be limited scope for SFNs; a second option, again mentioned above, is the extension of the current DTT networks. The possibilities for extension would further increase once the analogue services were switched off. However, this is again not likely to lead to an optimum plan; by far the fairest, and probably most efficient plan would emerge if a complete re-plan were undertaken. However this would make the transition very difficult indeed, since there could potentially be knock-ons in every country from a change in another country. Furthermore, new transmitting equipment and receiving antennas would be required in many cases. 16

Conclusions Network studies Conversions Portable coverage Migration strategy and the European context 17 In conclusion, this presentation has shown the results of some of the alldigital network studies which we have carried out, comparing different types and configurations of network. For the 128-stations studied, the MFN s proved to give better overall coverage than the SFN s, but required more channels to do this. This position might be reversed with a much more dense SFN. An attractive option has been mentioned which uses hierarchical modulation to proved a rugged layer for portable and mobile reception in some areas, and a higher capacity but less rugged layer for fixed reception. We have looked at the attractiveness of analogue conversions, and seen how they could be used to virtually match the current analogue coverage. However, there use is not without its problems, and these need to be studied in much more detail. Some estimates of portable coverage have been given, and we have seen how the use of conversions broadly allows analogue portable coverage to be matched by the digital service. Quite outstanding coverage is obtained if more rugged lower bit-rate modes are used. Finally we have set down some of the requirements for a digital migration strategy, and shown that these must be coordinated with the requirements of our European colleagues. 17