It seems that whatever the FCC says about 8-VSB and COFDM, the issue is not going away. I am not qualified to compare the modulation systems together in the field, but others have quite clearly demonstrated the robustness of COFDM over the 8-VSB system, particularly in multipath situations.
I am also perverse enough to believe that some of the strange things manufacturers of “correction” ICs have said in press releases puts them into the shadowy area of vaporware. After all, if you don't understand the basics of video in what you write, why should I believe you understand how to correct RF intangibles?
That aside, there was a recent survey of broadcast TV stations by SCRI International that indicates that, after excluding the 18 percent who were unsure, 45 percent were interested in using COFDM to some extent or another. Of those stations, a healthy 53 percent wanted to go there directly. Considering that the major supporters of COFDM own a considerable number of stations, we probably shouldn't be surprised by the results.
The latest wrinkle in pushing COFDM is the single-frequency network (SFN) system, where a number of stations broadcast on the same frequency in the same way that cellular phone cells operate. With the robustness of the modulation standard, proponents say, SFNs are a handy way to achieve wide coverage, fill in pockets of poor coverage and save the spectrum. These are all very laudable goals, but most of the proponents don't appear to fully understand the technology.
If we go back a couple of years, there was a wonderful apolitical “Position Paper” issued by PBS in support of continuing the implementation of the 8-VSB system. “Raising this fundamental issue of modulation systems now in the light of there being very little technical justification for it is extremely risky … any hope of funding for the public television transition to DTV will disappear.” There are no doubts where that is coming from, but the paper does include justification for the continuation of 8-VSB with its supposedly better noise immunity and statements that suggest the multipath performance isn't really that bad and that the system needs less power for the same carrier-to-noise. It does mention the large SFNs in use in Europe but rather avoids describing the technology.
But what's involved? The DVB DTTB design parameters include a guard interval (NOT a band), which is what provides the major resilience against delayed multipath signals. But building an SFN just based on the timing protection of the guard interval would not buy much. The spacing between transmitters would be quite small — which might be OK in a city, but impractical for a larger network. Other things can be done to protect for multipath at larger distances, such as varying the polarity of the transmitters, changing radiated powers, changing antenna patterns and re-timing the transmitters as needed.
The only real limitation with SFNs is that the same programming must be carried, although each carrier can have multiple programs.
I also have heard people say that an SFN is only really useful at UHF because transmission is so stable. Not so. There are DTV broadcast SFNs at VHF frequencies, and the technology also has been in use in digital radio broadcasting from 200 MHz to 1500 MHz with different modulation schemes. This is important because SFNs are not limited to COFDM. They may be easier to engineer for that modulation scheme, but they should be usable for 8-VSB as well if time was spent on the “hows.”
Why do I think that? Well, the SFN isn't exactly a new idea. Mr. H. L. Kirke joined the BBC in 1924, and in the early 1930s, he became the first head of research at the corporation. Apparently a very practical man, he guided many developments in broadcasting in short waves, 405-line television, noise-canceling microphones and a technology known as the Simultaneous Broadcast (SB) system — a topic that for many years was an item in the induction course for all new BBC employees.
The SB system uses the same medium-wave frequencies many times over across the country, all carrying the same AM programming. Yes, there are places where the signals interfere with one another, but the network was so carefully engineered that those places are generally only occupied by sheep.
I don't have a favorite between 8-VSB and COFDM any more; we are too far down the road to turn back. We have to make it work just as we can make 8-VSB SFNs work if we want them.
Paul McGoldrick is a freelance industry consultant based on the West Coast.
