 | Issue: | Asia-Pacific I 2007 |
| Article no.: | 12 | |
| Topic: | Bringing TV to mobile | |
| Author: | Hubert Zimmermann | |
| Title: | CEO | |
| Organisation: | UDcast | |
| PDF size: | 304KB |
About author
Hubert Zimmermann is the Chief Executive Officer of UDcast. Previously, Mr Zimmermann was director of high availability software engineering at Sun Microsystems. He joined Sun upon the acquisition of Chorus Systems, a global system software company of which he was CEO. Before founding Chorus Systems, Mr Zimmermann led a number of R&D projects at INRIA, France Telecom R&D and the French Ministry of Defence. His other credentials include leading the development and the international standardization of the seven layer OSI reference model at ISO and CCITT, sitting on France Telecom Scientific Advisory Board (1995-2000) and receiving several awards (IEEE, ACM, French Academy of Sciences) for his work in the field of computer communications. Mr Zimmermann earned engineering degrees from Ecole Polytechnique (Paris) and Ecole Nationale Supérieure des Telecom (Paris).
Article abstract
Portable TV, television on the move, has never been popular. The sets were too big and heavy, the antenna was awkward and the battery life was bad. New technology, the cell phone, 3G mobile networks and handsets, and a number of new standards for mobile TV including DVB-H, have finally made mobile TV a viable proposition. Earning a reasonable return on investment, however, requires a relatively low cost, high-performance, infrastructure solution such as that provided by a flexible direct-to-transmitter architecture.
Full Article
The TV was one of the most successful products developed during the second half of the 20th Century. Yet, unlike its simpler predecessor, radio, it has been a struggle to turn portable, mobile television into a successful mass-market product. Numerous problems with the products offered in the past – such as short battery life and large, vulnerable antennas – resulted in low take-up despite all the millions spent in development and advertising. Recently, however, the DVB-H, Digital Video Broadcast – Handheld, standard – explicitly defined for broadcasting to mobile devices – is bringing mobile TV to a new level in terms of the image quality and low distribution costs. With recent technical developments, such as the advent of comparatively inexpensive and low power LCD screens and improved battery technology, mobile TV is now looking far more attractive than at any time before. In order to maximize the commercial success of emerging Mobile TV networks, a good balance must be struck between providing more advanced and attractive services to increase revenue, and the cost of providing it. One of the most attractive ways to increase revenue is to provide localized content; therefore, Mobile TV systems need to be optimized to support this content while keeping their bandwidth requirements and equipment costs low. Technical challenges Because DVB-H is a broadcast technology operating in licensed radio frequencies, at least part of its challenge stems from the need to balance its spectrum requirements with those of other broadcasting services. DVB-H is largely adapted for the UHF (470-862MHz) spectrum, but it is also possible to deploy it in the L (1.67GHz) and S (2.2GHz) bands. The UHF bands offer excellent penetration of buildings and hence the largest cell size, but these frequencies are currently not available everywhere in the world. The optimum coverage of dense urban areas will in any case require significant investment in base stations in order to achieve the required level of signal strength and quality. DVB-H, by reusing much of the previously defined DVB-T, Digital Video Broadcast – Terrestrial, standard, is also able to reuse part of the equipment and network-planning set-up of DVB-T broadcasters. However, because DVB-T is not intended for mobile use, issues such as handover between broadcast zones and reception quality of the built-in antenna receivers do not apply. The distribution network supplying content to the towers must be carefully engineered. Commercial challenges The two main sources of revenues from DVB-H services are subscription fees and advertising revenues. First and foremost, subscribers are looking for a similar programmeming experience to that of traditional TV. This includes the availability of local TV stations, regional news and coverage of regional sports events. The subscription value can be increased by providing locally adapted services; for example, weather and traffic news that is specific to the viewer’s area. In addition to subscription income, advertising revenues are likely to be a critically important means of preserving an adequate revenue level. The latest multimedia technologies, when combined with various ‘canned’ advertising services, are now making the cost of creating video advertising extremely attractive. Local video advertising, significantly less costly than a national campaign, massively increases the potential advertiser base, enabling much-improved advertising sales pitch results. Hence, the potential for localization – knowing exactly where the viewer is located in relation, for example, to an advertiser’s store – greatly increases the commercial viability of the DVB-H service. However, localization has a significant impact on the sort of network architecture to be employed. Mobile TV network architectures Distributed The traditional distributed DVB-H architecture has a distributed set of IP encapsulators (one per region if not per SFN, single frequency network cell), each of which receives content via the contribution IP network. The distribution network is then used to distribute the MPEG-2 encoded transport stream to the modulators within the local SFN cells. Centralized The centralized architecture has its IP Encapsulators located all together in a single location. There are no external contribution networks, but the distribution network is, accordingly, considerably larger and more complex. The centralized architecture does not directly rule out localization of the content, but it is clearly optimised for centralized content. This architecture is similar to national broadcasting networks today. Direct-to-transmitter Direct-to-transmitter is the logical extension of the centralized architecture. Rather than using a terrestrial distribution network, the encoded signals are transmitted via satellite to all transmission towers. Unfortunately, a localization system is difficult to introduce using this architecture without sending multiple separate high-bandwidth streams via the satellite. Flexible direct-to-transmitter architecture The ideal solution enabling both local content and use of low-cost satellite distribution is to use the direct-to-transmitter approach, while adding an inexpensive local content splicing component. The new component sitting on the transmission path at each tower enables the content that is intended for the local area to be selected from the IP encapsulated MPEG2 transport stream. It then regenerates the stream including this selected subset of contents. The creation of a device for this sort of modification of the transmission stream is far from straightforward, requiring intimate knowledge of both satellite transmission and DVB-H. One solution was the development of the patented DVB-H ‘isochronous splicer’ adapter, which allows for the isochronous, completely deterministic regeneration of the encapsulated stream. The result combines the low capital and operational expenditure offered by satellite transmission with not only the content flexibility of the distributed model, but also the synchronization and handover advantages of the centralized model. In this ‘isochronous splicer’, a single transport stream is created by the DVB-H IP encapsulator, which contains all the local content for each region as well as the common national data. The stream is completely standard except for some additional in-bound control, added to the PSI/SI, Programme Specific Information/Service Information, tables, which identifies which cell’s particular local content it is destined for. However, it has to be noted that since the transport stream includes all the content for all areas, it consumes more encapsulation bandwidth than would be the case if the encapsulation included only the content that can be broadcast by a single DVB-H transmitter. The isochronous splicer’s filtering algorithm uses the modified PSI/SI information to identify which parts of the stream are to be broadcasted locally, and reconstructs the stream so that it comprises only relevant content. The filters perform this task in real time in a completely deterministic fashion so that not only are the timing requirements of DVB-H adhered to, but the synchronization between different towers is also guaranteed. Conclusions Network architecture for provision of DVB-H is something that an operator wants to be able to choose only once. The cost of capital investment means that in most cases migrating a network from one type to another is likely to ruin any hope of a respectable, let alone positive, return on investment. The simplicity, synchronization advantages and low capital investment required for the satellite architectures mean that deployment of such direct-to-transmit networks is far more attractive from a commercial perspective. Thanks to the ‘flexible direct’ architecture, made possible by the DVB-H isochronous splicer technology, low-cost satellite distribution can be combined with the flexibility of a terrestrial distributed architecture. This not only minimizes the network’s capital expense and operating expense, it also increases the return on investment by opening a new field of revenue and partnership opportunities for local TV channels based upon geo-addressable advertising.
