Several events are causing many broadcasters to wonder whether the ATSC 8-VSB modulation method was the best choice for U.S. digital television terrestrial broadcasting (DTTB). Tests and demonstrations by Sinclair Broadcasting highlighted the limitations in today's implementation of 8-VSB. The modulation method chosen by European broadcasters under the DVB-T standard is COFDM.
Broadcast engineers got a look at the robustness of COFDM at this year's NAB convention in Las Vegas. NDS, Microwave Radio Corp. (now Adaptive Broadband) and others had live demonstrations showing solid microwave COFDM reception from moving vehicles in situations where analog signals would have been unusable.
In June and July, Sinclair conducted demonstrations that compared reception of COFDM and 8-VSB signals. The same channel, transmitting antenna and average power level were used for both signals. All reports I've seen from people witnessing the demonstration confirm what Sinclair's engineers said: Aligning the antenna for COFDM reception was much easier than for 8-VSB reception.
In addition, COFDM worked in several situations where the 8-VSB signal was impossible to receive reliably on consumer indoor TV antennas.
While researching this series, I've talked to receiver and chipset manufacturers, broadcast equipment manufacturers and the engineers at Sinclair Broadcasting. I've also searched the Web and Internet newsgroups for reports from consumers on 8-VSB reception. This month I'll describe the advantages COFDM offers over 8-VSB and the price we must pay for those advantages.
You will see that neither solution is perfect. Which is best depends on broadcasters' priorities, and there will be differences of opinion. Next month, I will discuss broadcasters' varying requirements for a DTV system and look at the improvements coming in the second-generation 8-VSB receivers.
Some stations are having a particular problem with ATSC transmission that affects 8-VSB reception. This may account for some of the intermittent reception problems.
What is COFDM? The initials stand for Coded Orthogonal Frequency Division Multiplexing. If you are interested in learning more about COFDM, an excellent tutorial by J.H. Stott at BBC Research and Development is available on the Internet.
Basically, COFDM spreads a datastream across multiple carriers ÷ either 1,705 carriers or 6,817 carriers in a 7.61 MHz bandwidth (8 MHz channel) under DVB-T. The use of multiple carriers, in an orthogonal set, is, in my opinion, the major difference between COFDM and 8-VSB. The carriers may be modulated with QPSK, 16 QAM or 64 QAM.
This flexibility is one of the advantages COFDM has over the ATSC 8-VSB system. The DVB-T system allows different numbers of carriers, different modulation type for each carrier and different coding and guard intervals, depending on how robust a system is required.
Coding plays a large part in the performance of a digital transmission system. Coding, in its simplest definition, is a method of adding additional data to a one-way datastream to allow errors to be corrected in the receiver.
Because COFDM uses multiple carriers, it is possible to use channel-state information such as the signal-to-noise ratio for each carrier, for example, to determine how much weight to give the data from that carrier. This is a very simplified description of the "soft decision" process, but it should not be too hard to understand in practice.
With carefully designed coding and frequency interleaving, COFDM is able to work with zero dB echoes. Given this amount of echo, entire groups of carriers will be in the noise. The "soft decision" erases the data from the noisy carriers. However, with coding and interleaving, the transmitted datastream can be recovered from the data that survives.
Interference to one carrier is easy to deal with compared with echoes, which affect multiple carriers. Stott's The How and Why of COFDM, mentioned earlier, explains how this works in detail.
There is a lot more to COFDM than I am able to describe here. I haven't touched on orthogonality, guard intervals or pilot information. Refer to the links at the end of the column to learn more.
While not an "advantage," before a system can be seriously considered equipment has to be available for it. After the Charlotte tests were completed, some TV engineers were interested in testing COFDM for DTTB in the U.S. However, at that time the only commercial applications of COFDM were for data and digital audio broadcasting.
Equipment wasn't widely available for COFDM digital TV transmission until last year. However, in spite of a late start, the market for DVB-T COFDM-based equipment is large enough that most major manufacturers have announced COFDM TV transmitters.
Harris has a Web site devoted to its line of DVB-T transmission and monitoring equipment. Itelco will be showing a DVB-T exciter and transmitter system at IBC 1999 in Amsterdam this month. Thomcast's Comark division is expected to have a COFDM module for its MODAP digital TV modulator available soon.
If COFDM is so robust, why don't we just scrap 8-VSB, ignore the years of development and field-testing, and switch to DVB-T?
As you may suspect, the robustness of COFDM comes at a price. That price is lower data rate, worse threshold performance and increased transmitter peak power. The debate is about how high this price is and whether it's worth it.
For the Sinclair demonstrations, the usable COFDM data rate was 18.66 Mbps. While that's a bit less than the 19.39 Mbps data rate used in the 8-VSB system, leading HDTV encoder manufacturers have said they can encode HDTV signals with a rate as low as 14 Mbps. The COFDM parameters were described as 1,705 carriers (the "2k" mode) modulated with 64 QAM, coded with 3/4 forward error correction and a 1/8 guard interval.
As far as the Sinclair demonstrations were concerned, the price paid in lower data rate appears acceptable, given today's encoder technology. At some point, that may change if the channel is used for something other than video programs and each bit becomes more valuable.
The price in threshold performance is more difficult to determine. Several factors affect receiver threshold. There is agreement, however, that 8-VSB can handle a lower carrier-to-noise ratio (C/N) than COFDM. The argument is about how much.
DVB, on its Facts About DVB-T Web page (see sidebar for link), says tests show that the LSI Logic DVB-T chipset is "about 1.4 dB" worse than 8-VSB in C/N. The DVB page noted, "This becomes even more insignificant when traded against the major gains in flexibility and multipath performance of DVB-T."
8-VSB proponents disagree with that assessment. ATSC and DVB use different methods of measuring threshold, leading to different results. Which is correct?
Dr. Yiyan Wu, senior research scientist with the Communications Research Centre in Ottawa, Canada, wrote a Performance Comparison of ATSC 8-VSB and DVB-T COFDM Transmission Systems for Digital Television Terrestrial Broadcasting for presentation to ICCE 1999. It is so evenly written, both DVB-T and 8-VSB proponents could refer to it to support their arguments.
Dr. Wu has studied both COFDM and 8-VSB transmission systems and has written a study cited by the Consumer Electronic Manufacturers Association CEMA in its FCC petition for a COFDM-based mobile multimedia broadcasting system (MMBS) on TV Channels 60 and above designed for commercial use. The Performance Comparison stated that in RF tests, COFDM was 4 dB worse in C/N, but theoretically that difference could be as little as 1.7 dB.
Dr. Wu, however, points out that because the two systems have different data rates and the threshold is defined differently, Eb/No (carrier-to-noise ratio per bit) is a better measure. In comparison with 8-VSB, in terms of theoretical Eb/No performance, COFDM was 1.3 dB worse with a payload data rate of 17.4 Mbps and 2.3 dB worse at 19.6 Mbps in theoretical comparisons than 8-VSB.
For the lower data rate, coding was 2/3, with a guard interval of 1/16. For the higher data rate, coding was 3/4, also with a guard interval of 1/16. In RF tests, the margin widened to 3.6 dB and 4.6 dB (estimated) at the 17.4 and 19.6 Mbps data rates, respectively.
Dr. Wu was careful to point out improvements are possible with both systems. Trade-offs affect the results. Please refer to Dr. Wu's paper, available at the ATSC Web site, for the entire story and the actual C/N and Eb/No ratios.
What about transmitter power? For the same coverage, if the C/N (or Eb/No) performance is worse, the transmitter power needs to be increased to compensate for it. This would imply that average effective radiated power would have to be increased by 4 dB (about 2.5 times) for COFDM to match the 8-VSB coverage.
That isn't the only price - COFDM has a higher peak-to-average ratio than 8-VSB, so the transmitter should be capable of delivering about 2 dB (1.6 times) more peak power. (Dr. Wu's paper mentioned 2.5 dB (or 1.8 times) the peak-to-average ratio of 8-VSB.)
As COFDM and 8-VSB technologies approach their theoretical performance limits, this power penalty will decrease. Looking at the numbers in Dr. Wu's paper, it appears the ATSC implementation is within a half dB of its theoretical limit, while the DVB-T implementation is about 3 dB worse.
Sinclair used the same average power for the COFDM and 8-VSB transmissions. This should have put COFDM at a disadvantage, but it wasn't evident in the demonstration. Sinclair has conducted some tests at locations in fringe reception areas and found 8-VSB and COFDM had similar performance. Long term, in the presence of varying signal levels or interference, the Sinclair results may have been different.
IMPACT ON ANALOG
As I previously noted, a lot of work has gone into the 8-VSB ATSC standard. One of the more difficult areas to test was the impact of digital broadcast signals on NTSC analog TV reception. Testing required subjects to view a series of images and rate the degradation under different interference conditions and levels. The results of these tests formed the basis for the FCC's DTV allocation table.
Some characteristics of the ATSC 8-VSB system, such as the switchable comb filter to reduce co-channel interference from NTSC, were added to deal with the congested U.S. TV band. Dr. Wu included a table comparing protection ratios for various conditions for the Canadian, U.S. and EBU systems. As expected, the ATSC system does better with co-channel analog interference.
The extensive amount of coding used in the ATSC system gives it an advantage over DVB-T for impulse noise. This was also observed in Sinclair's Baltimore demonstrations.
If COFDM power is increased to make up for the apparent 4 dB C/N disadvantage, interference to analog TV signals will increase. Some would counter that the current interference thresholds may not be appropriate, even for 8-VSB.
Stanley J. Salamon, Charles W. Rhodes and Charles W. Einolf, Jr., from the Advanced Television Technology Center in Alexandria, Va., presented a paper at NAB '99 titled DTV Taboo Channel Interference into NTSC at High-Power Levels. The paper reported on tests that showed under strong signal conditions, "interference may occur at significantly lower power conditions than what is predicted by the FCC Planning Factors."
This not to imply that the existing FCC Allocation Table should be revised. It may not even be possible accommodate all channels if it were changed. However, because every TV station I know is receiving the vast majority of its income stream from its analog channel, we need to be very careful not to create a condition that interrupts that stream.
NEXT MONTH á
Look for Part 2 of this discussion next month. Different stations have different priorities, and that may determine which transmission system your station would rather use.
Receiver manufacturers are making improvements. Take time to visit the Web sites listed below to learn much more than I can cover in these short articles.
I strongly recommend Dr. Yiyan Wu's paper comparing 8-VSB and COFDM and J.H. Stott's The How and Why of COFDM. Visit the ATSC and DVB Web sites and read the claims and counterclaims about each system. Follow the battle on Dale Cripp's HDTV Web site.
If you have a DTV on the air, I'd be interested in knowing what your experience has been with consumer DTV reception. Are the telephone calls coming in from digital TV viewers complaints or compliments? Let me know at firstname.lastname@example.org
Doug Lung is vice president and director of engineering for the Telemundo Group of stations.