No. 293 (January 2003)
Backwards compatibility forever?
One of the fundamental principles of broadcasting is that of "backwards compatibility", which implies that enhancements of broadcast services should not significantly degrade reception on existing receiving equipment. For example, the transmission system for FM stereo was designed to provide good performance on stereo receivers whilst hardly affecting mono reception - and, later, RDS was added to FM stereo. Similarly, analogue TV broadcasts were enhanced by the successive addition of new features, such as colour, teletext and stereo sound. In each case, access to the new features was possible only with new receivers, whilst earlier receivers continued to receive the basic services.
Backwards compatibility ensures that audiences are not disenfranchised by changes in technology. This policy is obviously sensible because the public's investment in broadcasting receivers is much greater than the broadcasters' investments in transmission networks. Nevertheless, compatibility cannot be maintained across major technological discontinuities, such as the transition from AM to FM or from analogue to digital. Nobody suggested that the audio CDs should be designed so that they could be played on gramophones designed for LP records - or even that CD players should play 12-inch (30 cm) records!
One of the recurring themes in these editorials is that broadcasters should learn from the experiences in parallel industries, such as information technology and telecommunications. New file formats, such as JPEG-2000, can be used on computers built several years before the file format was defined - because new software can run on old computers. Similarly, in the 1970s, new software for error correction techniques was transmitted from Earth to the Pioneer spacecraft, thus upgrading the performance of the on-board telecommunications systems - even though the spacecraft had left the Earth several years earlier. Broadcasting engineers have long cherished the notion that, one day, it will be possible to upgrade broadcasting receivers that were sold many years previously. But, at the beginning of the 21st century, this still seems a distant hope!
This edition of the EBU Technical Review contains articles describing the latest developments in compression schemes, such as H.264/AVC and Windows Media 9 Series. Obviously, new compression systems offer significant longer-term benefits to broadcasters. For a broadcaster introducing digital TV in 5 or 10 years from now, today's MPEG-2 standard will not be the optimal system in terms of spectrum efficiency. It is, however, possible that MPEG-2 might be selected if the broadcaster wanted to deliver services to a large base of existing digital TV receivers. Broadcasters introducing digital TV today have little choice: MPEG-2 may not be the "best" system, but MPEG-2 receivers are available at low prices and will become much cheaper in the next few years.
Is the concept of backwards compatibility still applicable in the digital era? Will it be possible to broadcast hybrid services that can be received on TV sets based on MPEG-2 as well as on more modern systems? One obvious solution would be to transmit two parallel streams of data: one for each compression scheme. Although this would work, it would be pointless because it would require a total data rate in excess of that needed for MPEG-2 alone. The benefits of new compression systems are achieved only if transmissions based on older compression schemes are withdrawn.
How can broadcasters upgrade from one digital compression scheme to another? Unfortunately, the answer seems to be "with great difficulty" - unless it proves possible to upgrade existing receivers by downloading new software for the latest compression schemes (remembering that such receivers may need to be capable of handling several compression schemes). Is this a realistic prospect in the near future? Probably not, because all digital TV receivers use integrated circuits designed for a specific compression scheme - rather than a much more expensive general-purpose processor that is capable of running software for multiple compression schemes.
Although dedicated hardware remains significantly cheaper than software solutions for the reception of digital TV services, the situation is not so bleak for digital radio - where the bit-rates are significantly lower and, consequently, less demanding in terms of processing power. One pioneering example was the Psion Wavefinder DAB radio, introduced in the year 2000: this was essentially a software radio used in conjunction with a personal computer. This amply demonstrated that the processing power of modern computers is more than sufficient for the demodulation of the COFDM signals, and de-multiplexing and decoding of the audio and data signals. Other examples of software radios are highlighted in this issue of EBU Technical Review in the article about Digital Radio Mondiale (DRM).
These first steps towards software-defined radios for broadcasting are important, but it will be many years before such radios become the norm. It will be even longer before software-defined TV sets become widely available. The benefits seem obvious: software-based receivers will ensure that consumers can buy hardware in the knowledge that it will be future proof. Of course, the degree of future proofing may be limited by hardware (e.g. the processing power and memory capability). It is unlikely that consumers will buy equipment that is "over-powered" by a factor of 10 just because that power might be needed in 20 years. In addition, the move away from hardware to software solutions brings its own problems: as I have often said before, "You may have never seen an error message on your TV set, but don't worry - you soon will!". Users of PCs rightly fear the appearance of the "blue screen of death". If you regularly use interactive services on digital TV, you are probably familiar with frozen displays that can be rectified only by disconnecting the power supply. We can only hope that the reliability of software will dramatically improve before we enter the era of software-defined receivers.
In summary, the traditional policy of ensuring backward compatibility may have to be abandoned in the digital era. Software-defined receivers offer an elegant solution to compatibility, but are unlikely to be widely adopted in the near future.
Philip Laven
Director
EBU Technical Department
| European Broadcasting Union Case postale 45 Ancienne Route 17A CH-1218 Grand-Saconnex Geneva Switzerland techreview@ebu.ch |
 |