Reception problems to come: minding digital P’s and Q’s
Most of the technical discussion we hear about digital television is the debate about the several format options we have--progressive vs. interlaced, 480 lines vs. 720 vs. 1080, and so on. There's no agreed-upon "right" answer, and the argument is over which wonderful (compared to NTSC) picture format to use. A nice problem to have.
But despite the optimism about soon being able to transmit these great pictures,
we have cause to be pessimistic about how well people are going to be able
to receive them.
You may recall from college statistics that the probability that an event will happen is designated as P and the probability that it won't happen (1 - P) is designated as Q. In audience terms, if your average share (P) is 5 percent, then the probability (Q) that someone isn't watching you is 95 percent.
I want to talk about two aspects of the Q side. The first is an inherent characteristic of digital transmission itself. The second is a disturbing "gotcha"--the expected, widespread misalignment of viewers' receiving antennas.
Television broadcasters know about the Grade B contour because it's what we see in the Television and Cable Factbook, and many are even familiar with the probability term that defines it. Decades ago, the FCC defined received-signal levels for VHF and UHF television channels that would produce a subjectively acceptable picture over average terrain to 50 percent of the locations, 50 percent of the time--F(50,50). The circle of these geographic points around your transmitter is the Grade B contour.
In the world of analog TV, many places that receive less signal than the Grade B level still get a usable picture because receivers tend to be better now than when the FCC established its standards. Even viewers who have a noisy picture tend to be okay about it because they realize they live a long way from the transmitter. P and Q therefore have been somewhat mushy concepts.
But the digital world is different. DTV pictures don't get progressively noisier as they fade; they stay very good until they suddenly disappear--what engineers call the "cliff effect." There might be only a decibel or two separating a beautiful picture from a blank blue screen. P and Q therefore become much more important concepts.
Recognizing this, when the FCC defined the DTV signal necessary to replicate a station's Grade B contour, it increased the time component from 50 percent to 90 percent. But the flip side of the P(50,90) standard means there's also a Q(50,10) lurking. At a distance approximately equal to your current Grade B contour, the FCC standard says that the best 50 percent of locations will have service 90 percent of the time, and no service 10 percent of the time. Viewers in the worst 50 percent of locations, of course, could lose out even more often.
Signal strengths vary both by the season and by the time of day. These variances are usually only a few decibels, and are usually barely discernable in analog television. But in DTV they can make the difference between Teletubbies at 8 a.m. and nothing at all at 4 p.m. if a receive site is near the threshold.
I won't afflict you with the math, but you can easily compute other P and Q probabilities as well. For example, on the due-west radial of my Pullman station (KWSU), the distance to the P(50,90) point is 62 miles, the P(50,95) point is 48 miles, the P(50,98) point is 31 miles, and the P(50,99) point is 20 miles. Since P(50,99) = Q(50,01), then FCC methodology predicts that the best 50 percent of locations 20 miles from my transmitter are going to lose service 1 percent of the time (about 88 hours per year!). Remember, this is no picture at all, not just a noisy picture.
But wait! For some stations it gets worse. There are other contributors to the Q side. If your viewers are located behind blocking terrain, in a high multipath environment (with high-rise buildings or mountainous terrain, for example), or if you have a "lone wolf" transmitter located away from where people have their antennas pointed, your practical reception area could be considerably less than the P(50,90) method predicts. Let's look at receive-antenna misalignment as another Q side factor.
Tests by the Sinclair Broadcast Group station in Baltimore, reported in Broadcasting & Cable and elsewhere, have shown that a Radio Shack directional antenna misaligned even 15 degrees from the DTV station could not produce a picture. Misaligned antennas introduce multipath and lose gain. People tend to aim their antennas where they can get the most good pictures and rarely install a special antenna or rotator for the occasional "lone wolf" station. So, unless the "lone wolf" is close enough for adequate indoor antenna reception, prevailing antenna orientations can cause a lot of DTV reception problems.
Again, let's use KWSU as an example. Our transmitter is located 50 miles south of the mountain where most of Spokane's television stations have their transmitters (all except for the "lone wolf" ABC station). Viewers throughout our coverage area point their antennas toward this mountain. Viewers in only one-eighth of our coverage area are likely to have their antennas pointed within 15 degrees of our tower. Based on the Baltimore experience, these are the only viewers who will be able to pick up KWSU over the air. Viewers in the remaining seven-eighths will have problems receiving our DTV signal unless they install a separate antenna for KWSU, or a rotator.
The greater the distance between your transmitter and the majority of the area's TV towers, the fewer viewers will be able to pick up your station over-the-air. This will pose especially serious problems for state networks, which tend to have a lot of "lone wolf" transmitter sites.
On the edge of our region, another "lone wolf" CBS station, in
Idaho 30 miles farther south of KWSU, has only about one-twelfth of its coverage
area within the favorable over-the-air reception area. For the ABC station
mentioned above, 25 miles from the cluster of towers, the major metro area,
Spokane, will be entirely outside the favorable area.
So, what can we do?
First and foremost, work hard for cable DTV must-carry rules--and rules that protect your entire data stream. It is the single best means of ensuring your viewers can continue to receive your programming.
Plan to assign staff to work as reception technicians. Some viewers will need to have their hands held as DTV forces over-the-air viewers to pay a lot more attention to their indoor or outdoor antennas. Properly engineered, digital paths can be very reliable. One manufacturer reportedly is planning to offer site surveys with each receiver sold.
Pursue partnerships with other stations and allocation studies that could lead to moving your transmitter closer to the cluster of stations that people are favoring with their antennas.
Investigate ways to fill-in holes in your coverage due to terrain shadowing or antenna misalignment. You'll have more holes with digital than with analog, so you'll have to do more than just establish DTV translators where you now have analog ones. There is encouraging, though still not conclusive, work being done on on-channel repeaters. Learn from cellular, PCS and LMDS; they achieve blanket coverage despite greater propagation handicaps.
mind your P's and Q's. Because not everyone will receive you or receive you
100 percent of the time, we need to learn to manage the consequences. Develop
business strategies in which the advantages of DTV overcome the loss in reliable
coverage that many stations will experience.
Although we can't do much about the shortcomings in early DTV technology ourselves, the "rocket scientists" in electronics and broadcasting need to design receiving antennas--especially indoor ones--and receivers that can handle the effects of multipath. Fortunately, DTV receivers will benefit in this regard from the frequent doubling and redoubling of computing power.
Dennis Haarsager is chairman of the PBS Board's New Technologies
Committee and a member of CPB's Digital Funding Task Force and Digital Advisory
Group. Since 1978 he has managed KWSU stations based in Pullman, Wash., including
the statewide Northwest Public Radio network and public TV outlets in Pullman
and Richland. He also serves as associate v.p. for educational telecommunications
and technology of the licensee, Washington State University. Before coming
to WSU, he worked in South Dakota and Idaho's public TV networks. On the side,
Haarsager writes and publishes software that estimates broadcast coverage.
Web page posted March 8, 2004
Current: the newspaper about public TV and radio
in the United States
Current Publishing Committee