The new digital radio system DAB (Digital Audio Broadcasting, nowadays often called Digital Radio) is a very innovative and universal multimedia broadcast system which will replace the existing AM and FM audio broadcast services in many parts of the world in the future. It was developed in the 1990s by the Eureka 147/DAB project. DAB is very well suited for mobile reception and provides very high robustness against multipath reception. It allows use of single frequency networks (SFNs) for high frequency efficiency. Besides high-quality digital audio services (mono, two-channel or multichannel stereophonic), DAB is able to transmit programme-associated data and a multiplex of other data services (e.g. travel and traffic information, still and moving pictures, etc.). A dynamic multiplex management on the network side opens up possibilities for flexible programming. In several countries in Europe and overseas broadcast organisations, network providers and receiver manufacturers are going to implement digital broadcasting services using the DAB system in pilot projects and public services. DAB works very differently from conventional broadcasting systems. Most of the system components such as perceptual audio coding, channel coding and modulation, multiplex management or data transmission protocols are new solutions and typically not so familiar to the expert in existing analogue or digital broadcast systems.
The level of standardisation of the DAB system is rather advanced and the various recent international standards and related documents are introduced and referred to for easy access for the reader seeking technical details.
Radio in the Digital Age
Radio broadcasting is one of the most widespread electronic mass media comprising hundreds of programme providers, thousands of HF transmitters and billions of receivers world-wide. Since the beginning of broadcasting in the early 1920s, the market has been widely covered by the AM and FM audio broadcasting services. Today we live in a world of digital communication systems and services. Essential parts of the production processes in radio houses were changed to digital ones in recent times, beginning with the change from conventional analogue audio tape to digital recording on magnetic tape or hard disk, digital signal processing in mixing desks and digital transmission links in distribution processes. In addition, there are also other digital distribution or storage media in a growing music market such as several digital tape or disc formats (CD, MiniDisk or DVD), or streaming and download formats for distribution via the Internet (see also section 1.6.4). Consequently, broadcast transmission systems now tend to change from conventional analogue transmission to digital. The first steps in the introduction of digital broadcasting services were taken by the systems NICAM 728 (Near Instantaneously Companded Audio Multiplex, developed by the BBC for stereo television sound in the VHF/UHF bands), DSR (Digital Satellite Radio, which was already finished), or ADR (Astra Digital Radio), see section 1.6.1, but none were suited to replace the existing conventional services completely, especially for mobile reception. For that reason, the universal digital multimedia broadcasting system Eureka 147 DAB was developed and is now being introduced world-wide. In parallel, other digital broadcasting systems such as DRM (Digital Radio Mondiale, see section 1.6.3) or DVB-T (Digital Video Broadcasting, see section 1.6.2) are under consideration to complement digital radio and television. Normally, it takes a period of a human generation (or at least a period in the life of a receiver type generation, i.e. approximately 10 years) to replace an existing broadcasting system by a new one. Therefore, strong reasons and very convincing advantages are required to justify the introduction of such a new system.
Benefits of the Eureka 147 DAB System
However, there will always be some problems, or additional effort will be needed, when replacing an existing technology by a new one, such as
- lack of transmission frequencies
- costs for development and investment
- looking for providers for new non-conventional services (e.g. data services)
- solving the chicken and egg problem (who will be first – the service provider or the receiver manufacturer?).
Nevertheless, the Eureka 147 DAB system provides a wealth of advantages over conventional audio broadcast systems such as analogue VHF/FM or AM radio, and also partly over other existing digital broadcast systems such as DSR, ADR, etc. The following list will only highlight some key advantages as an overview; many more details will be explained in the corresponding sections of the book.
Quality of service
DAB uses all the possibilities of modern digital communication technologies and can thus provide a much higher level of quality of service, such as
- Superior sound quality: DAB users can enjoy pure undistorted sound close to CD quality. New features such as Dynamic Range Control (DRC) or Music/Speech Control can be used individually by customers to match the audio quality to their needs.
- Usability: Rather than searching wavebands, users can select all available stations or preferred formats from a simple text menu.
- Perfect reception conditions: With just a simple, non-directional whip antenna, DAB eliminates interference and the problem of multipath while in a car. It covers wide geographical areas with an even, uninterrupted signal. Once full services are up and running, a driver will be able to cross an entire country and stay tuned to the same station with no signal fade and without altering frequency.
Wide range of value-added services
DAB is quite unique in that both music and data services can be received using the same receiver. One receiver does it all, such as
- Typical audio broadcasting (main service): Music, drama, news, information, etc., can be received in monophonic or stereophonic form as is well known from conventional radio programmes; there is also the potential to transmit multichannel (5.1 format) audio programmes as well.
- Programme-associated data (PAD): DAB broadcast receivers can display text information in far greater detail than RDS, such as programme background facts, a menu of future broadcasts and complementary advertising information. Receivers attached to a small screen will display visual information such as weather maps or CD cover images.
- Information services: Services from sources other than the broadcasting station are included within the same channel for the user to access at will. These include news headlines, detailed weather information or even the latest stock marked prices.
-Targeted music or data services: Because digital technology can carry a massive amount of information, specific user groups can be targeted with great accuracy because each receiver can be addressable.
- Still or moving pictures: Data can also appear as still or moving photographic pictures, accompanied by an audio service or as separate information.
Universal system layout
The DAB system has a fairly universal and well-standardised system layout which allows applications for all known transmission media and receiving situations.
- Standardisation: The level of international standardisation of all basic principles and transmission tools for the new DAB system is very high (more than 50 international standards cover all the necessary details).
- Unique system design: DAB services will be available mainly on terrestrial, but are also suited for cable and satellite networks, and the same receiver could be used to provide radio programmes and/or data services for national, regional, local and international coverage.
- Wide choice of receivers: It is possible to access DAB services on a wide range of receiving equipment including fixed (stationary), mobile and portable radio receivers, optionally completed with displays or screens, and even personal computers.
Flexibility of multiplex configuration
DAB services are transmitted in a flexible multiplex configuration, which can be easily changed instantaneously to the actual needs of the content providers.
- Multiplex configuration: The arrangement of services in a DAB multiplex may be changed instantaneously to match the needs of the providers of programmes or data services, without interrupting ongoing services.
- Bit rate flexibility: The programme provider can choose an appropriate bit rate for a certain audio programme according to its quality, for instance less than 100 kbit/s for a pure speech programme, 128 kbit/s for monophonic or 256 kbit/s for stereophonic music; also half sampling frequency can be used for lower quality services. So the available bit rate can be split optimally between different services of a DAB ensemble.
Transmission efficiency
Compared to conventional broadcast systems much less economic effort in investment and operation is needed for a DAB transmission system.
- Lower transmission costs for broadcasters: DAB allows broadcasters to provide a wide range of programme material simultaneously on the same frequency. This not only makes room for a vastly increased number of programmes to increase user choice, but also has important broadcast cost-cutting implications.
- Lower transmission costs for transmitter network providers: For digital transmission a DAB transmitter needs only a fraction of the electrical energy compared to a conventional AM or FM transmitter.
- Frequency efficiency: DAB transmitter networks can be designed as Single Frequency Network (SFNs), which saves a lot of transmission frequencies and thus transmission capacity on air.
These advantages of DAB (and there are more if we look further into the details) justify the introduction of DAB into the media world in order to replace the existing conventional radio systems step by step over a longer period.
Steps of Development
In the early 1980s the first digital sound broadcasting systems providing CD like audio quality were developed for satellite delivery. These systems made use of the broadcasting bands in the 10 to 12 GHz region, employed very little sound data compression and were not aimed at mobile reception. Thus, it was not possible to serve a great majority of listeners, such as those travelling in cars. Also, another feature of the well-established FM radio could not be provided by satellite delivery, namely ‘‘local services’’. Consequently terrestrial digital sound broadcasting was considered as an essential delivery method for reaching all listeners. At first investigations were initiated by radio research institutes looking into the feasibility of applying digital modulation schemes in the FM bands. However, the straightforward use of pulse code modulation (PCM) in the upper portions of the FM band generated intolerable interference in most existing FM receivers and was spectrally very inefficient. Mobile reception was never tried and would not have succeeded. A much more sophisticated approach was definitely necessary. In Germany the Federal Ministry for Research and Technology (BMFT, now BMBF) launched a research initiative to assess the feasibility of terrestrial digital sound broadcasting comprising more effective methods of sound data compression and efficient use of the radio spectrum. A study completed in 1984 indicated that promising results could be expected from highly demanding research activities. As a new digital sound broadcasting system could only be implemented successfully by wide international agreement, BMFT set the task for its Project Management Agency at DLR (German Aerospace Centre) to form a European consortium of industry, broadcasters, network providers, research centres and academia for the development of a new digital audio broadcasting system. Towards the end of 1986 a consortium of 19 organisations from France, Germany, The Netherlands and the United Kingdom had signed a co-operation agreement and applied for notification as a Eureka project. At the meeting in December 1986 of the High Level Representatives of the Eureka partner states in Stockholm the project, now called ‘‘Digital Audio Broadcasting, DAB’’, was notified as the Eureka 147 project. National research grants were awarded to that project in France, Germany and The Netherlands. However, owing to granting procedures official work on the project could not start before the beginning of 1988 and was supposed to run for four years. Credit must also be given to the European Broadcasting Union (EBU), which had launched work on the satellite delivery of digital sound broadcasting to mobiles in the frequency range between 1 and 3 GHz, by awarding a research contract to the Centre Commun d’Etudes de Te´le´diffusion et Te´le´communications (CCETT) in Rennes, France, prior to the forming of the DAB consortium. As the CCETT also joined the DAB project, the work already begun for the EBU became part of the DAB activities and the EBU a close ally and active promoter for DAB. Later, this proved very important and helpful in relations with the International Telecommunications Union (ITU-R) and the standardisation process with the European Telecommunications Standards Institute (ETSI). From the beginning the goals set for the project were very demanding and difficult to achieve. Perfect mobile reception was the overall aim. In detail the list of requirements to be met included the following items: . audio quality comparable to that of the CD . unimpaired mobile reception in a car, even at high speeds . efficient frequency spectrum utilisation . transmission capacity for ancillary data . low transmitting power . terrestrial, cable and satellite delivery options . easy-to-operate receivers . European or better world-wide standardisation. The first system approach considered at least 16 stereo programmes of CD audio quality plus ancillary data to be transmitted in the 7 MHz bandwidth of a television channel. This definitely cannot be achieved by simply transmitting the combined net bit rates of 16 CD-like programme channels, which are around 1.4 Mbit/s each, over the TV channel. So a high degree of audio data compression without any perceptible loss of audio quality was mandatory. Data rates below 200 kbit/s per stereo channel had to be achieved. Unimpaired mobile reception was also required to overcome the adverse effects of multipath signal propagation with the associated frequency selective fading. Audio data compression and the transmission method became the central efforts of the research project. Both tasks were addressed in a broad and comprehensive manner. For audio coding four different approaches were investigated: two sub-band coding systems competed with two transform coding systems. Similarly, for the transmission method four different schemes were proposed: . one narrow-band system . one single carrier spread-spectrum system . one multicarrier OFDM system . and one frequency-hopping system. All approaches were developed to an extent where – either through experimental evidence or at least by thorough simulation – a fair and valid comparison of the performance of the proposed solutions became possible. The period of selection of and decision for the best suited audio coding system and the most appropriate transmission scheme was a crucial moment in the history of the Eureka 147 consortium. For audio coding the greatest part of the selection process happened external to the consortium. All four coding schemes previously had been within the activities of the ISO/IEC Moving Pictures Experts Group (MPEG), which worked on standardisation for data compressed video and audio coding. There the solutions offered by Eureka 147 competed against 10 other entries from other countries, world-wide. The MPEG Audio Group set up a very elaborate and qualified audio quality assessment campaign that was strongly supported by Swedish Radio, the British Broadcasting Corporation and the Communications Research Centre, Canada, among others. The very thorough subjective audio quality tests revealed that the audio coding systems submitted by Eureka 147 showed superior performance and consequently were standardised by ISO/IEC as MPEG Audio Layers I, II and III. Within the Eureka 147 consortium, after long consideration Layer II, also known as MUSICAM, was selected for the DAB specification. The process of choosing the most appropriate transmission method took place within the Eureka 147 consortium alone. In simulations performed according to rules worked out by the members the four approaches were put to the test. This showed that the broadband solutions performed better than the narrow-band proposal. Among the broadband versions the spread-spectrum approach had a slight advantage over the OFDM approach, while the frequency-hopping solution was considered too demanding with respect to network organisation. However, the OFDM system was the only one that was already available in hardware with field-test experience – in the form of the coded Orthogonal Frequency Division Multiplex (COFDM) system, while the spread-spectrum proposal by then was not developed as hardware at all and was estimated to be very complex. So, the choice fell on COFDM, which has since proven to be an excellent performer. A further important decision had to be made relating to the bandwidth of the DAB system. From a network and service area planning point of view as well as obtainable frequency spectrum perspectives, an ensemble of 16 programmes on one transmitter with the 7 MHz bandwidth of a TV channel proved to be much too inflexible, although in experiments it had provided very good performance in a multipath environment. Therefore, a considerable but reasonable reduction in transmission bandwidth was necessary. In Canada experiments with the COFDM system revealed that substantial performance degradation begins around 1.3 MHz and lower. So, a reasonable bandwidth for a DAB channel or ‘‘DAB block’’ was defined as 1.5 MHz. This allows several possibilities, as follows. A 7 MHz TV channel can be divided into four DAB blocks, each carrying ensembles of five to seven programmes. With four blocks fitting into 7 MHz service area planning is possible with only one TV channel, without having adjacent areas using the same DAB block. Furthermore, 1.5 MHz bandwidth is sufficient to transport one MPEG coded audio/video bit stream. After the above-mentioned important decisions had been made the members of the consortium all turned their efforts from their individual approaches to the commonly defined system architecture and with rapid progress developed the details of the complete basic DAB specification to be submitted to the international standardisation bodies. By that time, another European research project, the JESSI flagship project AE-14, was eagerly awaiting the DAB specification to begin the development of chip-sets for DAB. Also, the standardisation bodies like ETSI were well aware and waiting for the submission of the specification since Eureka 147 members had been very active in testing and presenting the results of their research and development on many important occasions, together with or organised by the EBU. The first official presentation of DAB took place at the World Administrative Radio Conference 1988 (WARC’88) in Geneva for the delegates of this conference. As the issue of a frequency allocation for digital sound broadcasting in the L-band around 1.5 GHz was up for decision, a demonstration simulating satellite reception was presented. A transmitter on Mont Sale`ve radiated the DAB signal in the UHF TV band, giving in downtown Geneva an angle of signal incidence similar to that from a geostationary satellite. Mobile reception was perfect and the delegates to the conference were highly impressed. In consequence the conference assigned spectrum in the L-band for satellite sound broadcasting with terrestrial augmentation permitted. One year later DAB was presented at the ITU-COM exhibition in Geneva. In 1990 tests with mobile demonstrations were run by Canadian broadcasters in Toronto, Ottawa, Montreal and Vancouver. DAB demonstrations and exhibits have been shown at all International Radio Shows (IFA) in Berlin and the UK Radio Festivals in Birmingham since 1991. Four International DAB Symposia have been held up to now: 1992 in Montreux, 1994 in Toronto, 1996 again in Montreux and 1999 in Singapore. In 1994 a mobile DAB demonstration was presented at the Arab States Broadcasting Union Conference in Tunis. From all these activities world-wide recognition and appreciation was gained for DAB. The efforts and results of finding acceptance for DAB in the United States deserve an extra paragraph. As early as 1990 – by invitation of the National Association of Broadcasters (NAB) – the consortium presented DAB at low key at the NAB Convention in Atlanta, Georgia. This led to a very elaborate mobile demonstration and exhibition at the next NAB Convention in 1991 in Las Vegas, Nevada. Several of the NAB officials by that time were very interested in reaching a co-operation and licence agreement with the Eureka 147 consortium. However, strong opposition against that system – requiring new spectrum and the bundling of several programmes onto one transmitter – also arose. The US radio industry – that is, the broadcasters – feared new competition from the licensing of new spectrum. They preferred the idea of a system approach named ‘‘In-Band-On-Channel (IBOC)’’, where the digital presentation of their analogue programmes is transmitted together and within the spectrum mask of their licensed FM channel (see also section 1.6). This of course would avoid the need of new licensing for digital broadcasting and thus keep new competition away. However appealing and spectrum efficient this concept may be, the realisation might prove to be a very formidable technical task. No feasible development was available at that time, but fast development of an IBOC system was promised. Eureka 147 DAB performed flawlessly at Las Vegas, but those opposing the system claimed that the topography around Las Vegas was much too favourable to put DAB to a real test. So it was requested that DAB should next come to the 1991 NAB Radio Show in San Francisco to be tested in a very difficult propagation environment. One main transmitter and one gap filler were set up and mobile reception in downtown San Francisco was impressively demonstrated. An announced demonstration of IBOC broadcasting did not take place as the equipment was not ready. In spite of the good results the opposition to DAB gained momentum and the NAB officially announced its preference for an IBOC system. This was not in line with the intentions of the Consumer Electronics Manufacturers Association (CEMA) of the Electronics Industry Association (EIA) which came to an agreement with the NAB to run very strictly monitored laboratory tests in Cleveland and field trials in San Francisco comparing the Eureka 147 DAB with several US IBOC and IBAC (InBand-Adjacent Channel) systems. The results were to be presented to the Federal Communications Commission (FCC) for rule-making relating to digital sound broadcasting. While Eureka 147 soon had equipment ready for the tests, the US proponents for a long time could not provide equipment and delayed the tests until 1995. In the laboratory tests DAB outperformed all competing systems by far. Claiming to have been unfairly treated in the evaluation of the laboratory tests, all but one of the US proponents of terrestrial systems withdrew from the field tests in San Francisco. For Eureka 147 the Canadian partner Digital Radio Research Inc. (DRRI) installed on contract a single frequency network of one main transmitter and two gap fillers to provide coverage for the area designated for mobile testing. Again DAB provided excellent performance as was documented in the final reports of the CEMA. Nevertheless, the United States still pursued the concept of IBOC although several generations of IBOC equipment and redesigns have only produced marginal performance. Even though numerous Americans now admit that DAB is a superior system they claim that it is not suited for US broadcasters. Finally, in October 2002, the Federal Communications Commission FCC approved In-Band On-Channel (IBOC) systems for the AM and FM band developed by the company iBiquity. Stations may implement digital transmissions immediately; however, AM stations may send the IBOC signal during the day only. Standardisation in Europe and world-wide occurred at a better pace. The first DAB-related standard was achieved for audio coding in 1993, when the International Organisation for Standardisation / International Electrical Commission (ISO/IEC) released the International Standard IS 11172-3 comprising MPEG/Audio Layers I, II and III [IS 11172]. Also in 1993 ETSI adopted the basic DAB standard ETS 300 401, with several additional standards following later (now replaced by [EN 300401]). The ITU-R in 1994 issued Recommendations [BS.1114] and [BO.1130] relating to satellite and terrestrial digital audio broadcasting, recommending the use of Eureka 147 DAB mentioned as ‘‘Digital System A’’. Consumer equipment manufacturers have also achieved several standards concerning the basic requirements for DAB receivers issued by the Comite´ Europe´en de Normalisation Electrotechnique (CENELEC) (for more detailed information see section 1.5). Even though the technology, the norms and standards had been developed, the most critical issue was still not resolved: provision of frequency spectrum for DAB. WARC’88 had allocated 40 MHz of spectrum in the L-band to satellite sound broadcasting, allowing also terrestrial augmentation. Through the intense intervention of several national delegations WARC’92 conceded primary terrestrial use of a portion of that allocated spectrum, which for several countries is the only frequency resource available. The L-band – very well suited for satellite delivery of DAB – on the other hand becomes very costly for terrestrial network implementation. VHF and UHF are much more cost efficient than the terrestrial L-band. There was no hope of acquiring any additional spectrum below 1 GHz outside of the bands already allocated to broadcasting.
So, in 1995 the Conference Europe´enne des Administrations des Postes et des Te´le´communications (CEPT) convened a spectrum planning conference for terrestrial DAB in Wiesbaden, Germany, that worked out an allotment plan for DAB in VHF Band III (mostly former TV channel 12) and in the L-band from 1452 to 1467 MHz, allowing for all CEPT member states two coverages of DAB, one in the VHF range, the other in the L-band. This decision made possible the installation of experimental DAB pilot services in many countries of Europe and the beginning of regular services starting in 1997 with Sweden and the United Kingdom. More and more countries – also outside of Europe – are following suit. As the available spectrum was not sufficient to move all existing FM and future programmes to DAB, new frequency bands have been opened for DAB in a later step.
Organisations and Platforms
A few of the organisations and bodies were or still are busy promoting and supporting the development and introduction of DAB system world-wide:
Eureka 147 consortium
As mentioned above the Eureka 147 consortium was the driving force in the development of DAB. It formed a Programme Board, dealing with strategic planning, contractual and legal affairs, membership and promotion, a Steering Committee, planning the tasks of the working groups and making all technical decisions, and four working groups of varying task assignments. DLR in Cologne, Germany, was chosen to act as the managing agency and as the project office for Eureka affairs. While the original consortium did not accept new members at the beginning, this policy was changed when the basic specification of DAB [EN 300401] was ready to be released to the public for standardisation in 1992. The Eureka 147 Programme Board established rules for the entry of new partners into the consortium, requiring, for example, an entrance fee of DM 150,000 from industrial companies while entry for broadcasters and network operators was free of charge. The money earned from the entrance fees was mainly reserved for system promotion and to a smaller extent for organisational expenses. In 1993 the first new members were admitted and soon the consortium grew to 54 members from 14 countries. The research work of the original partners had led to a number of basic patents for DAB, individually held by several members. Consequently the consortium came to the conclusion to offer licences as a package to all necessary intellectual property rights (IPR) through authorised agents, one for receiver matters and another for transmitter and measuring equipment. In 1997 the idea of merging Eureka 147 with the promoting organisation WorldDAB (see below) was discussed with the result that a gradual merger was adopted and the final merger completed by the beginning of the year 2000. The activities of the EU-147 Project Office were transferred to the WorldDAB Project Office by the end of 1998. With the complete merger at the end of 1999 Eureka 147 ceased to exist. The members of the consortium can now co-operate in WorldDAB Module A, where technical issues of DAB are handled.
National DAB platforms
The first national DAB platform DAB Plattform e.V. was initiated by the German Ministry for Education and Research in 1991. Founded as a national platform it soon accepted members from Austria and Switzerland and consequently dropped the word ‘‘national’’ from its name. It soon reached a membership of 52 organisations. The main objective of this platform was the promotion of the DAB system in German-speaking countries. The organisation of and activities for DAB presentations at the IFA events in Berlin were highlights in the promotion programme of the German platform. The members decided to end the existence of DAB Plattform e.V. by autumn 1998 after considering it to have achieved its objectives. Likewise, national platforms were established in several other European countries. France launched the Club DAB France, The Netherlands the Dutch DAB Foundation and the United Kingdom the UK DAB Forum. Promotional activities for the public and information to governmental agencies were again the objectives.
EuroDAB/WorldDAB
Forum Strongly stimulated and supported by the EBU, an organised co-operation of national platforms and other interested bodies lead to the founding of the EuroDAB Forum in 1995. A EuroDAB Project Office was set up at EBU Headquarters in Geneva. Membership rose quickly, bringing together interested organisations not only from Europe but from many parts of the world. Accordingly the General Assembly of EuroDAB decided to change its name to the WorldDAB Forum in 1997. WorldDAB has around 100 members from 25 countries world-wide. Promotion of DAB is again the main objective. Formerly EuroDAB and now WorldDAB issue a quarterly DAB Newsletter. The Forum has organised all but the first International DAB Symposia mentioned above. Technical, legal and regulatory as well as promotional issues are dealt with in several modules of the Forum. Since January 2000 the Eureka 147 consortium has fully merged with the WorldDAB Forum. The WorldDAB Project Office moved from Geneva to London in 1998 and has also acted for Eureka 147 since 1999. An extensive web-site of WorldDAB and actual DAB information is maintained at this office [www.WorldDAB]
AsiaDAB Committee
The Asian DAB Committee of the WorldDAB Forum was formally established in June 2000 in Singapore. The main focus of the AsiaDAB Committee is to raise awareness of DAB in Asia and to speed up the roll-out of DAB in this important part of the world. The committee encourages co-operation and sharing of experiences on DAB matters in close co-operation with the WorldDAB Forum. It is a valuable platform for industry players and regulators in the Asian marketplace. For more details see [www.AsiaDAB].
Milestones of Introduction
At the time of writing (early 2003) DAB remains – after more than 10 years’ development and standardisation – still in the phase of early introduction. However, in some regions programme and service providers have already started regular DAB services. Although DAB was primarily developed in the Eureka 147 project as a European broadcast system, it can be shown already in this phase that most of the industrial developed regions world-wide (i.e. Asia, Australia, and parts of America such as Canada and South America) are now interested in introducing this highly sophisticated universal broadcast system, too. The world-wide penetration of DAB is mainly supported by the WorldDAB Forum and – especially in the Asian marketplace – by the AsiaDAB Committee (see above). The actual status of introduction can be found in Appendix 2 and at the web-sites also given above. Although the DAB system is still in a status of introduction, at the time of writing over 280 million people around the world can receive more than 400 different DAB services. Only in the important market places of United States and Japan are proprietary digital radio systems being developed,
International Standardisation
The new DAB system shows a very high level of standardisation in its principles and applications, which is rather unusual for a new broadcasting or multimedia system. (There are more than 50 corresponding international standards and related documents.) After the initial development by some broadcasters and related research institutes, the first international standards were passed by the ITU-R (International Telecommunications Union, Radiocommunications Sector). In the period following, the system has been developed within the Eureka 147 project, supported by a wide representation of telecommunication network providers, broadcasters, receiver manufacturers and related research institutes, in close co-operation with the EBU. In the following some basic standards of the DAB system are listed. A more complete listing of the corresponding standards and related documents can be found in the Bibliography, which is referred to within various parts of the text.
Basic requirements and system standards
The basic ITU-R Recommendation [BS.774] shortly specifies as ‘‘Digital system A’’ the main requirements to the new broadcasting system DAB. Other ITU-R Recommendations [BS.789] and [BO.1130] regulate conditions needed for additional frequency ranges for emission. Several ITU World Administrative Conferences (WARC’79, WARC’85, WARC’88, WARC’92) dealt with digital radio and allocated frequency bands for satellite and terrestrial digital sound broadcasting. More details were decided in [CEPT, 1995]. As a result of developments within the Eureka 147 project, the Main DAB Standard or DAB Specification [EN 300401] was approved by the ETSI European Telecommunications Standards Institute, which defines the characteristics of the DAB transmission signal, including audio coding, data services, signal and service multiplexing, channel coding and modulation. (This standard was formerly often called ‘‘the ETS’’; now it is a European Standard EN.)
Audio coding
DAB represents one of the most important applications of the generic ISO/ MPEG-1 Layer II audio coding scheme [IS 11172]. The use of this ISO/IEC coding standard is recommended by the ITU-R in [BS.1115], and certainly in the DAB Specification [EN 300401]. DAB is also designed to transmit MPEG-2 Layer II audio [IS 13818], for instance for lower quality half-sample-rate transmission or multichannel audio programmes, see also [ES 201755]. Other standards specify procedures and test bit-streams for DAB audio conformance testing [TS 101757], or audio interfaces for transmission within the studio region [IEC 60958], [IEC 61937].
Data services
Owing to the special needs for data services in DAB a new standard for Multimedia Object Transfer (MOT) was created defining the specific transport encoding of data types not specified in [EN 300401] and ensuring interoperability between different data services and application types or equipment of different providers or manufacturers. Additional guidelines are given in the ‘‘MOT Rules of operation’’ [TS 101497], ‘‘Broadcast Web-site application’’ [TS 101498] or ‘‘Slide show application’’ [TS 101499]. An important bridge between the well-known radio data service RDS for FM [EN 50067] and the data services in DAB is provided by the European Norm [EN 301700] which defines service referencing from FM-RDS and the use of RDSODA (open data applications). A transparent data channel for DAB transmission is described in [TS 101759]. Late standards were given for Internet Protocol (IP) datagram tunnelling [TS 101735] and for the DAB Electronic Programme Guide (EPG) [TS 102818].
Network and transmission standards
Based on the DAB main standard [EN 300401] additional standards are given to define the DAB multiplex signal formats for distribution (networking) and transmission. This is the so-called Service Transport Interface (STI) [EN 300797] and [TS 101860] for contribution networks between service providers and broadcast studios, the Ensemble Transport Interface (ETI) [EN 300799], and the Digital baseband In-phase and Quadrature (DIQ) interface [EN 300798] for DAB channel coding using OFDM modulation. To provide interactive services, transport mechanisms for IP Datagram Tunnelling [ES 201735], Network Independent Protocols [ES 201736] and Interaction Channel Through GSM/PSTN/ISDN/DECT [ES 201737] are defined.
Receiver requirements
Based on the DAB specification [EN 300401] additional standards are given to define the implementation of DAB receivers. [EN 50248] describes DAB receiver characteristics for consumer equipment for terrestrial, cable and satellite reception. EMC parameters for receivers are identified in EACEM TR-004. [EN 50255] specifies the Radio Data Interface (RDI) between DAB receivers and peripheral data equipment. A special command set to control receivers is described in [EN 50320]. The ETSI Technical Report [TR 101758] lists general field strength considerations for a DAB system. More general requirements to radio receivers concerning EMC are given in [EN 55013] and [EN 55020].
Guidelines for implementation and operation
Additional to the given standards, more detailed guidelines and rules for implementation and operation are compiled in [TR 101296] as a main guideline document for service providers and manufacturers. A guide to standards, guidelines and bibliography is given in the ETSI Technical Report [TR 101495]. A broadcaster’s introduction to the implementation of some key DAB system features [BPN 007] is provided by the EBU.
Relations to Other Digital Broadcasting
Systems In addition to the European DAB system which is mainly covered by this book, several other digital sound broadcasting services exist which have been or are being developed and (partly) introduced. These systems differ in many aspects (specifications, service complexity, parameters) from the DAB concept. Some are focused more strongly on a single application or service (e.g. stationary reception) or provide lower audio quality levels (such as WorldSpace or DRM). Except for ISDB-T, which uses technologies very similar to DAB, all other systems are not expected to be able to comply with the audio quality and quality of service in mobile reception provided by DAB. Nevertheless the basic concepts and limitations of some of these systems will be introduced briefly. Also, there are a few cable-based digital radio services, which are mainly derived from existing satellite or terrestrial radio systems (for instance, ADR, DVB, etc.) so it may not be necessary to describe them separately. In general, those systems use QAM schemes in order to achieve high data rates in the limited bandwidth of the cable. This is possible because of the high signal-to-noise ratio available. However, these services have only local importance, depending of the extension of the broadband cable distribution network used.
Satellite-based Digital Radio Systems
Although the DAB system can also be used for satellite transmissions, a number of different proprietary systems have been designed by companies providing direct broadcasting satellite services. Of course, many radio programmes are transmitted via telecommunication satellites to feed relay transmitters or cable head stations. These systems are beyond the scope of this section. Generally, the system layout in the direct broadcasting systems focuses either on stationary reception with highly directive antennas (e.g. ‘‘dishes’’) in the 11 GHz range (e.g. ADR) or on portable and possibly mobile reception in the L-band (1.5 GHz) and S-band (2.3/2.5/2.6 GHz) allocations (e.g. WorldSpace, XM and Sirius). Historically, the first system for the direct satellite broadcasting of digital radio services was Digital Satellite Radio (DSR). In contrast to later systems, no sound compression was used but only a slight reduction in data rate as compared to CD by using a scale factor. This service has recently been closed down because of the end of the lifetime of the satellite used (Kopernikus). Therefore, it is not described here. More details can be found in [Schambeck, 1987]. The basic building blocks of modern satellite systems are audio coding, some kind of multiplexing, and modulation. While there are things in common with DAB with respect to the first two aspects, these systems do not use OFDM modulation because of the non-linear travelling wave tube amplifiers on satellites and the low spectral efficiency. Instead, often relatively simple PSK schemes are used, which are considered sufficient because, when using directional receiving antennas, multipath propagation does not occur in satellite channels. Satellite systems face severe problems when the service is to be extended to mobile reception. Because only line-of-sight operation is possible owing to the limited power available in satellite transmissions, at higher latitudes, where geostationary satellites have low angles of elevation, either terrestrial retransmission in cities and mountainous regions is necessary, or the service has to be provided by several satellites at different azimuths in parallel to fill shaded areas. Another approach is to use satellites in highly inclined elliptical orbits which always show high elevation angles. However, this requires several satellites in the same orbit to provide continuous service and difficult switching procedures have to be performed at hand-over from the descending to the ascending satellite.
Astra Digital Radio
The Astra Digital Radio (ADR) system was designed by Socie´te´ Europe´enne des Satellites (SES), Luxembourg, to provide digital radio on its geostationary direct TV broadcasting satellites. This co-positioned family of satellites labelled ‘‘ASTRA’’ covers Central Europe operating in the 11 GHz range. Brief system overviews are given by [Hofmeir, 1995] or [Kleine, 1995]. A detailed description of the system is available at the ASTRA web-site [www.ASTRA]. The system design is compatible with the sub-carrier scheme for analogue sound transmission within TV transponders. It uses MPEG Layer II sound coding (as does DAB) at a fixed bit rate of 192 kbit/s for a stereo signal. Forward error correction is applied using a punctured convolutional code with code rate 3/4. The data are differentially encoded to provide easy synchronisation. QPSK modulation is used for each of up to 12 ADR sub-carriers which are combined with the analogue TV signal and are FM modulated. The baseband bandwidth of the digital signal is 130 kHz which is the same as for an analogue sound sub-carrier. This allows existing analogue radio services to be replaced by the digital ADR service one by one. It is also possible to use a whole transponder for ADR only. In this case, 48 channels can be accommodated in one transponder. Additional data can be sent together with the audio signal. ADR uses the auxiliary data field of the MPEG audio frame to do this, but the coding is different from the DAB Programme Associated Data (PAD; see Chapter 2). ADR uses 252 bits per audio frame for this purpose, which results in a net data rate of about 6 kbit/s, because a (7,4) block code is used for error correction. This capacity is flexibly split into ‘‘RDS data’’, ‘‘Ancillary Data’’ and ‘‘Control Data’’. RDS data are coded in the format of the RDS [EN 50067] which is used on a digital sub-carrier on FM sound broadcasts and provides service-related information such as labels and Programme Type (see also section 2.5). The reason for using this format on ADR is that several broadcasters use ADR to feed their terrestrial FM networks. Hence these data can be extracted from the ADR signal and fed to the RDS encoder of the FM transmitter. Ancillary data are used by the broadcaster for in-house purposes. The control data field contains a list of all ADR channels to provide easy tuning and allows transmission of parameters for conditional access if applied. With respect to DAB, ADR provides the same quality of audio, because the same coding mechanism is used. The number of channels available, however, is much larger. While DAB currently only provides six to seven services in each network, ADR can provide 12 digital radio channels on each transponder in addition to a TV channel. Therefore, several hundred channels are available and in use by public and private broadcasters from all over Europe. ADR decoders are integrated in high-end satellite TV receivers and are therefore widespread. The most important drawback of ADR is that it can only be received with stationary receivers using a dish antenna. The flexibility of the system with respect to data transmission and future multimedia extensions is considerably lower than that of DAB.
WorldSpace
The WorldSpace digital radio system was designed by the US company WorldSpace Inc., to provide digital radio to developing countries on its two geostationary satellites ‘‘AfriStar’’ covering Africa, Southern Europe and the Near and Middle East, and ‘‘AsiaStar’’ covering India, China, Japan and South East Asia. A third satellite named ‘‘CaribStar’’ is intended to cover Central and South America but has not yet been launched. The system downlink operates in the L-band. The standard is not public. A description is given by [Sachdev, 1997], for example. The basic building blocks of the systems are ‘‘Prime Rate Channels’’ (PRCs) which each transmit 16 kbit/s of data. Several of these (typically up to eight, resulting in 128 kbit/s) can be combined to provide a stereo sound channel or data channels. For audio coding, MPEG Layer III is used. Each satellite serves three coverage areas (spots). For each spot, two multiplex signals containing 96 PRCs (i.e. 12 stereo channels) each in a time domain multiplex (TDM) are available. Each of these data streams is QPSK modulated. One multiplex is completely assembled in a ground station and linked up to the satellite from which it is retransmitted without further processing. The other multiplex is processed on board the satellites by assembling individual PRC uplinks to the TDM signal. The advantage of this is that there are no additional costs to transport the signal from the broadcaster to the uplink station. Moreover, a signal can be broadcast in several spots with only one uplink. Owing to the lower path loss of the L-band signals as compared to 11 GHz, the WorldSpace system is able to provide significant field strength within the coverage area. Therefore reception is possible with portable receivers outdoors and inside buildings using a low-gain patch or helical antenna at a window where line-of-sight conditions apply. Minor attenuation by foliage or trees may be tolerable. At the time of writing approximately 10 different receivers are available on the market. With respect to DAB, WorldSpace addresses a completely different broadcasting environment. Clearly, in developing countries digital satellite radio will considerably improve the number and quality of broadcast receptions, because in many parts up to now there has only been short-wave coverage. Mobile reception is probably not an urgent need and hence the system is not particularly designed for it. From the point of view of audio quality, by using MPEG Layer III at 128 kbit/s the quality will be comparable to that of DAB at 160–192 kbit/s. At this bit rate, however, the number of channels per spot is restricted to 24. Owing to the large coverage areas this number may not be sufficient to serve the multilingual and multiethnic audience. Hence, the audio bit rate will have to be reduced in order to increase the number of channels available. Nevertheless the sound quality will in many cases still be comparable at least to mono FM, which is still a major improvement on short-wave AM. Utilising the concept of PRCs the system is very flexible and can be extended to data transmission and multimedia contents. This is important because the lifetime of the satellites is approximately 15 years.
Satellite Systems in the United States
In the United States there is no allocation for digital radio in the L-band, but in the S-band. Currently, two companies have established satellite services targeted at mobile reception and covering the whole of the continental United States (i.e. except Alaska and Hawaii). Sirius Satellite Radio uses three satellites on inclined orbits to achieve high elevation angles and terrestrial repeaters in metropolitan areas, see [www.Siriusradio]. The other company, XM Satellite Radio, uses two high-powered geostationary satellites at 858 and 1158 which both cover the entire continental United States and thus provide spatial diversity to overcome shading, see [www.XMRadio]. Nevertheless terrestrial retransmitters are also used. The systems provide approximately 100 audio channels in a quality comparable to CD. The services are pay-radio based on a monthly subscription fee. With respect to sound quality these systems should be comparable to DAB, and with respect to reliability of service these systems should be comparable to WorldSpace; however, because of the measures taken they should show better performance in mobile and portable reception.
Terrestrial Digital Broadcasting Systems
Digital Video Broadcasting – Terrestrial (DVB-T)
The DVB system was developed a few years after the DAB system. On the audio and video coding level as well as on the multiplexing level, it is completely based on the MPEG-2 standard. This is in contrast to the DAB standard, where the audio coding Introduction 17 is MPEG, but the multiplex control is independent and specially adapted for the requirements. There are three DVB standards that differ very much in the transmission scheme: DVB-S for satellite [EN 300421], DVB-C for cable [EN 300429], and DVB-T for terrestrial broadcasting [EN 300744]. Here we consider only the last one. For a detailed description of all three standards, see [Reimers, 1995]. The DVB-T standard has been designed for stationary and portable terrestrial reception with multipath fading. In contrast to DAB, mobile reception was not required. On the coding and modulation level, many methods have been adopted from the DAB system, where they had already been implemented successfully. Like DAB, DVB-T uses OFDM with a guard interval. There are two transmission modes. The first is called 8K mode (because it uses an 8192-point FFT) and has a symbol length of the same order as in DAB transmission mode I. It is suited for single frequency networks. The second is called 2K mode (because it uses a 2048-point FFT) and has a symbol length of the same order as in DAB transmission mode II. It is not suited for single frequency networks. The guard interval can be chosen to be 25% of the total symbol length, as for DAB, but also shorter guard intervals are possible. The total bandwidth of about 7.6 MHz is suited for terrestrial 8 MHz TV channels. The system parameters can be scaled for 7 MHz TV channels. A major difference from the DAB system is that DVB-T uses coherent modulation. Different signal constellations between QPSK and 64-QAM are possible. At the receiving side, the channel amplitude and phase have to be estimated. For channel estimation, nearly 10% of the total bandwidth is needed for pilot symbols. The coherent modulation scheme with channel estimation by a Wiener filter is more advanced than the differential demodulation scheme of DAB. It is even more robust against fast fading in mobile reception situations [Hoeher, 1991a], [Schulze, 1998], [Schulze, 1999]. The OFDM transmission scheme for DVB-T includes a frequency interleaver that consists of a simple pseudo-random permutation of the sub-carrier indices. It is very similar to the DAB frequency interleaver. In contrast to DAB, the DVB-T system has no time interleaving. Time interleaving only makes sense for mobile reception, which was not a requirement for the design of the system. The channel coding is based on the same convolutional codes like the DAB system. Code rates between 1/2 and 7/8 are possible. Unequal error protection (UEP) is not possible for the DVB system, since the data and transmission level are completely separated. As a further consequence, in contrast to DAB, the whole multiplex has to be coded with the same code rate. To reach the very low bit error rates that are required for the video codec, an outer Reed–Solomon (RS) code is applied: blocks of 188 bytes of useful data are coded into blocks of 204 bytes. Between the outer RS code and the inner convolutional code, a (relatively) short byte interleaver has been inserted to break up the burst errors that are produced by the Viterbi decoder. The data rates that can be carried by the DVB-T multiplex vary from about 5 Mbit/s to about 30 Mbit/s, depending on the modulation (QPSK, 16-QAM or 64- QAM), the code rate (1/2, 2/3, 3/4, 5/6 or 7/8) and the guard interval. Mobile reception of DVB-T, even though not a specific design feature, has been widely discussed and gave rise to many theoretical investigations and practical experiments. Furthermore, there seems to be a controversy between DAB and DVB adherents that concentrates very much on this question. Unfortunately, even though both systems are very close together, the chance has been missed to design a universal common system. From the technical point of view, we will list some comments on the question of mobile reception:
1. Unlike as one might expect, the differential modulation of DAB is not more robust than the coherent modulation for DVB even for severely fast-fading mobile reception. The coherent scheme is more robust and spectrally more efficient, see [Schulze, 1999].
2. For mobile reception in a frequency flat fading situation, the frequency interleaving even over a relatively wide band of 8 MHz is not sufficient. As a consequence, the lack of time interleaving may severely degrade the system in some mobile reception situations, see [Schulze, 1998].
3. Synchronisation losses may sometimes occur for mobile receivers. In contrast to DVB-T, this was taken into account in the design of the DAB system: the null symbol allows a rough frame and symbol synchronisation every 24 ms or a maximum of 96 ms. The FIC (see 2.2.2.) can then immediately be decoded and provides the information about the multiplex structure. And, in the (typical) stream mode, the data stream does not need separate synchronisation because it is synchronised to the physical frame structure.
4. Bit errors in the scale factors of the MPEG audio data (birdies) are very annoying for the human ear. For a subjective impression this is much worse than, say, block error in a picture. The DAB system uses an extra CRC (that is not part of the MPEG bit-stream) to detect and to conceal these errors. This is not possible for DVB-T.
As a consequence, the DVB-T system allows mobile reception in many typical environments, but it may fail in some situations where more care has been taken in the DAB system.
In-Band-On-Channel Systems
In the United States the broadcasting environment is considerably different from the European one. Technical parameters of the stations (AM or FM, power, antenna height) resulting in different coverage areas and quality of audio service appear to be important economic factors that are vital to the broadcasters and which many of them think have to be preserved in the transition to a digital system. Although the DAB system can provide such features to a certain extent by applying different audio data rates and coding profiles, the idea of multiplexing several stations into one transmitting channel does not seem to be very attractive to some US broadcasters. In the early 1990s, a number of US companies therefore started to work on alternative approaches to digital audio radio (DAR) broadcasting known as IBOC (In-Band-On-Channel) and IBAC (In-Band-Adjacent-Channel). Systems were developed both for AM and FM stations in MW and VHF Band II. The basic principle is to use low-power digital sideband signals within the spectrum mask of the channel allocated to the station.
Later in the 1990s a number of IBOC/IBAC systems and the Eureka 147 DAB system and a satellite-based system (VOA/JPL) were tested in both the laboratory and field in the San Francisco area. The results of these tests were not encouraging for the IBOC/IBAC approaches, see [Culver, 1996], stating explicitly that ‘‘of all systems tested, only the Eureka 147/DAB system offers the audio quality and signal robustness performance that listeners would expect from a new DAR service’’. Since then work on IBOC systems has nevertheless continued. At the time of writing (early 2003), the US National Radio Systems Committee (NRSC) has published evaluation reports and is now in the process of drafting standards both for FM and AM IBOC systems developed by a company called iBiquity Digital Corporation (a merger of the former proponents USA Digital Radio and Lucent Technologies). In October 2002 the systems were approved by the US Federal Communications Commission FCC and implementation is now possible. The system designed for Medium Wave (‘‘AM IBOC’’) has two basic configurations [Johnson, 2003]. The first is designed to be used in parallel to the analogue AM signal. It has OFDM blocks on either side of the AM carrier (even in the frequency band used by the AM side bands). However, the modulation and power level is different for the subcarriers depending on the frequency offset from the carrier. The carriers near to the AM carrier are QPSK modulated, those more than about 4 kHz from the carrier use 16-QAM and those more than 9.5 kHz from the carrier use 64- QAM. These have the highest power and carry most of the data. The total bandwidth of the signal is 29.4 kHz. This means that the digital transmission is not really on channel but uses both adjacent channels for the original transmission. Therefore, the digital signal can only be broadcast during daytime, when there is no sky wave propagation. In a situation where no analogue AM transmission is required, the whole channel can be used for the digital transmission. In this case, the unmodulated AM carrier is still transmitted but the AM channel is filled with the 64-QAM modulated OFDM carriers. The FM IBOC system [Peyla, 2003] is designed in a similar way. There is a hybrid mode where two OFDM blocks are sent on either side of the band used by the analogue FM signal. They extend from +129 kHz to +198 kHz, respectively and carry both the same digital information. The digital part of the signal can be extended by adding further OFDM carriers between + 101 kHz and +129 kHz. If there is no analogue FM signal, the inner part of the channel can also be filled with the digital transmission resulting in an OFDM block consisting of 1092 carriers in a band 396 kHz wide. In all cases, the two digital side bands carry identical data to increase the robustness of the system. Both the FM and AM IBOC systems use PAC as the audio coding system. However, independent of the individual system details there are two principle problems with the FM IBOC approach which are difficult to overcome. Owing to the fact that the FM IBOC signal is limited in bandwidth to the FCC channel spectrum mask, the bandwidth is not sufficient to overcome frequency selective fading in the VHF frequency range. This means that in portable or mobile reception a fade of the whole signal can occur. In this case there is hardly a chance to reconstruct the signal even if strong error correction codes and ‘‘long’’ interleaving are used (think of stopping at a traffic light). Therefore, from the point of view of coverage, IBOC systems are not expected to be superior to FM, because they will fail in the same locations as FM does. In iBiquity’s FM IBOC system it is proposed to delay the analogue signal against the digital one and store it in the receiver. Whenever the digital transmission fades, the receiver would blend to the analogue signal. Every IBOC signal will to a certain extent impair the analogue signal in the same channel and in adjacent channels. The effect of this is additional noise on the analogue audio signal and this may affect different analogue receiver circuits in different ways. Therefore the broadcaster has no control of what the digital signal will do to all the different analogue receivers being used. To keep this noise sufficiently low, the digital signal level must be kept far below the level of the analogue signal, e.g. 23 dB below the power of the FM carrier in iBiquity’s FB IBOC system. Although the digital system requires less signal-to-noise ratio than FM, this means that the coverage of IBOC will be very limited.
Integrated Services Digital Broadcasting (Japan)
In Japan, the NHK Science and Technical Research Laboratories have proposed a concept called Integrated Services Digital Broadcasting (ISDB). From this approach a system was created which can be configured for terrestrial and satellite broadcasts of radio, TV and multimedia services, see [Kuroda, 1997], [Nakahara, 1996], [ARIB, 1999]. For the terrestrial system (ISDB-T) the most important aims are rugged reception with portable and mobile receivers, use of Single Frequency Networks (SFNs) to achieve frequency efficient coverage and to have a flexible scheme in order to be future-proof. To achieve this, a variant of OFDM (see section 2.2) was developed which is called Band Segmented Transmission (BST)-OFDM and means that each signal consists of a number of basic OFDM building blocks called segments with a bandwidth of 571, 500 or 428 kHz (in TV channels with 8, 7 and 6 MHz bandwidth) which are spread over the band where frequencies are available. For the OFDM parameters of the system there are three sets of parameters that are similar to transmission modes I, II and IV of the DAB system. The system may be configured for channel bandwidth of 6, 7 or 8 MHz resulting in a segment bandwidth. A single segment is sufficient to broadcast audio programmes and data, but it is also possible to combine three segments for this purpose. For TV signals 13 segments are combined to form a 7.4/6.5/5.6 MHz wide signal. In all cases the basic modulation and coding modes of the segments are (D)QPSK, 16QAM and 64QAM, and several codes rates between 1/2 and 7/8 are available. In the 3 and 13 segment modes the individual segments may be coded and modulated in different ways to achieve hierarchical transmission capacity. For the audio and video coding the MPEG-2 standard will be used including AAC for sound transmission. In 1998 draft standards on digital terrestrial television and radio based on the ISDB-T concept were established, see [STRL, 1999], and field tests are being performed. Test transmissions are announced to start in Tokyo and Osaka in late 2003. From the point of view of the relation to DAB, ISDB-T basically uses the same principles and hence provides similar features as do DAB and DVB-T, respectively. By using the MPEG-2 standard for coding, the audio and video quality will be comparable. Due to the use of AAC for audio coding, however, the number of audio services which can be provided in a given bandwidth will be approximately twice that provided by DAB. Owing to the similarity of the OFDM parameters used, and similar bandwidth (at least when three segments are combined for audio), the coverage properties will also be similar to those of DAB. The advantage of ISDBT is that the radio and TV systems are based on the same basic building blocks (although the coding and modulation will be different), which allows for reuse of some circuits in receivers, whereas DAB and DVB-T differ in many details. Another advantage of ISDB-T is the availability of a narrow-band mode (i.e. a single segment) which can deliver one or two radio programmes to small regions or communities, although potentially with some reception problems due to flat fading. With this segmentation ISDB-T can also flexibly use the frequencies not occupied by analogue services in the introductory simulcast phase. Such possibilities are not present in DAB.
Digital Radio in the Broadcasting Bands below 30 MHz
The frequency range below 30 MHz (short, medium, long wave) offers the possibility to provide a very large coverage area for broadcasting with only one high-power transmitter and very simple receivers with no directional antennas. Using this frequency range is the easiest – and often practically the only possible – way to cover large countries with low technical infrastructure. On the other hand, traditional broadcasting in this frequency range uses the most antiquated and simplest possible transmission scheme: double-sideband AM with a carrier. From one point of view this is a waste of power and bandwidth efficiency in such a valuable frequency range at a time when every year more and more efficient communication systems are being designed. From a practical point of view, this old technique provides only very poor audio quality and thus finds acceptance only in regions where no other service is available. The obvious need for a more advanced system working in this frequency range gave rise to a consortium called Digital Radio Mondiale (DRM) to develop a worldwide standard for such a system. The DRM standard [ES 201980] was first published as a European Telecommunications Standard and has now been approved by ITU-R and IEC. An overview of the basic system features is given e.g. by [Stott, 2001], latest information is available from [www.DRM]. The available bandwidth in the frequency range below 30 MHz is typically very small: the usual channel spacing is 5, 9 or 10 kHz. For the new system, it should be possible to bundle two 5 kHz channels to one 10 kHz channel. For such narrow-band channels, much more data reduction is required than for the DAB system. The Advanced Audio Coding (AAC) data reduction scheme of MPEG in conjunction with a newer technique called Spectral Band Replication (SBR) (see also section 3.3.5) allows for an audio quality comparable to FM at a data rate between 20 and 25 kbit/s, see for example [Dietz, 2000]. This is a great improvement compared to conventional AM quality. But bearing in mind that the DAB system typically transmits little more than 1 Mbit/s (e.g. six audio programmes of 192 kbit/s) in a 1.5 MHz bandwidth, even this low data rate is relatively high compared to the extremely small bandwidth of 10 kHz. This means that the DRM transmission scheme must provide more than twice the bandwidth efficiency compared to DAB. The physical channel depends on the type of wave propagation, which is very special in this frequency range. It differs from night to day and it depends on the solar activity. Also interference caused by human noise plays an important role. In situations where only ground-wave propagation is present, there is a simple white Gaussian noise channel that is relatively easy to handle. But typically the wave propagation is dominated by ionospheric scattering (sky wave). This means that the channel is time and frequency variant – just like the mobile DAB channel (see section 2.1). Travel time differences of the wave of the order of several ms may occur which cause severe frequency-selective fading. The motion of the ionosphere causes a time variance in the channel with a Doppler spectrum with a typical bandwidth of the order of 1 Hz. The DRM channel has similar problems in time- and frequency-selective fading as the DAB channel, so the DRM group decided on the same solution: OFDM (see section 2.1). The parameters, however, are totally different from the ones used in DAB because the channel characteristics are quite different. For sky wave propagation there are several modes with the OFDM symbol durations of 26.67 ms, 20 ms and 16.67 ms and guard intervals of 5.33 ms, and 7.33 ms, respectively. With this symbol length, the OFDM signal consists of approximately 200 sub-carriers in a 10 kHz wide channel. The requirement of a very high-bandwidth efficiency leads to a much higher signal constellation for each sub-carrier than used for DAB. 64-QAM or 16-QAM will be used depending on the propagation situation. Because of the very large time variance of the channel, a large amount of the total bandwidth (up to about 1/6) is needed for channel estimation pilots. The overall code rate is between 1/2 and 2/3 for the required data rate. It can be shown [Schulze, 1999] that such a coded 64-QAM OFDM system with a good channel estimator is surprisingly robust in a fading channel even with conventional 64-QAM symbol mapping. For DRM it has been decided to use the multilevel coding approach first proposed by [Imai, 1977]. This approach is very similar to that described by [Wo¨rz, 1993] for a coded 8-PSK system. With multilevel 64-QAM, an additional gain of up to 1 or 2 dB can be reached compared to conventional 64-QAM. DRM is the first system that implements such a coded modulation scheme; see also [Dietz, 2000]. The same convolutional codes as DAB will be used here. For ground mode propagation, another transmission mode is defined, because less channel estimation is needed, the guard interval is only 2.33 ms long, and even weaker codes can be used in a Gaussian channel to allow for a higher data rate. Channel coding must be supported by interleaving to work in a fading channel. Frequency interleaving alone is not sufficient, especially if the echoes are too short or if only a two-path propagation situation occurs. Time interleaving is restricted by the delay that is allowed by the application: after the receiver has been switched on the listener cannot wait too long for a signal. A delay of 2 seconds is used for sky wave modes, and a delay of 0.5s is used in ground wave. Since the time variance is very slow (1 Hz Doppler spread), the interleaving will not be sufficient in many situations.
Web-casting
A completely different way to provide radio services is the use of the Internet as a universal multimedia platform. That means Web-casting is a multimedia extension of the Internet. As audio and video signals can be digitised into a file or stream of data, which can easily be distributed over the existing Internet structures. The term Webcasting means the publishing of audio and video files or streams on web-sites, for live and/or on-demand delivery to the general public. A main function of broadcaster’s web sites is to support the core radio and TV services in order to make them more attractive to the customer. The basic equipment for an Internet radio customer is a PC, complete with a sound adapter, a rather fast modem and Internet access, possibly via ISDN or ADSL. (Conventional Internet connections using bit-rates less than 14.4 kbit/s are not well suited at least for streaming applications due to the limited audio or video quality achieved.) The availability of improved and scaleable audio coding schemes, on the other hand, is creating the opportunity for a better definition of audio quality obtainable via the Web. The Internet’s impact on broadcasters has changed from its initial position as a site for posting information related to the station and programme to an additional medium to reach new customers through streaming audio, live or on-demand, still and moving pictures, and new interactive concepts. The importance of Internet audio services, which are using only a small part of the capacity of the net, is already significantly increasing audio information delivered via IP. Several software companies have produced the necessary tools to distribute audio and video via the Internet with the goal to develop the Web as a new mass medium, similar to the traditional audio and video broadcasting services. With the current streaming techniques and 56 kbit/s modems an audio quality can be achieved surpassing by far that of MW broadcasting with audio bit rates which are reduced by a factor of 50, compared to those used with the CD. For both, either for live streaming or downloading of audio clips, there are several audio players (software codecs) in use which can provide sufficient audio quality in the range mentioned below, such as . Microsoft Windows Media . RealNetworks G2 . Q-Design Music Codec 2 together with Quic ktime streaming . MPEG-2/-4 AAC ‘‘Advanced Audio Coding’’ . MP2, which is MPEG-1/-2 Layer II . MP3 (close to MPEG-1/-2 Layer III) . Yamaha Sound VQ. The advent of such a large number of audio codecs has brought a radically new approach to standardisation: standards have become less important, since decoders (which are normally simple and do not require a lot of processing power) are downloadable to the client machine along with the content.
Therefore, in the Internet environment there is no longer a need for a single coding system as is the case in conventional broadcasting. Service providers decide which coding scheme to use. One of the advantages of this ‘‘deregulated’’ approach is that decoders can be regularly updated as the technology advances, so the user can have the latest version of the decoder all the time. The business model of audio streaming is likely to change due to the advent of multicasting. Today, ISPs charge per audio stream. In multicasting situations, however, a single stream will be delivered to several users. The user will then be charged according to the occupancy of the servers used. Due to the huge competition in the audio decoder market, audio streamers will be increasingly available for free. Recent quality assessment tests show that most of the codecs on the market can satisfy the (rather low) requirements of the Internet community for PC based audio presentation, see [Stoll, 2000]. Nevertheless, web-casting will have its own place in the radio world because of the access to hundreds of radio programmes world-wide. Much more detailed information on web-casting and its audio quality expectations is contained in [BPN 022], [BPN 035] and [Stoll, 2000]. The necessary Internet Transport protocols and delivery technologies are described in [Kozamernik, 2002].
