THE RUN-UP TO DIGITAL TV

DTV channel options: How much to mux?

When public TV first learned it could air four program streams on a single DTV channel, that may have seemed too much airtime to fill. Now the system faces an abundance of high-def and other options, leaving stations to bet on those with the best chance to last and prosper. Liroff, who was v.p. and chief technology officer of Boston’s WGBH when he wrote this, examines the options and trade-offs.

Originally published in Current, Feb. 12, 2007
Commentary by David B. Liroff

At recent public television meetings, there has been much discussion about how stations should prepare for the February ’09 shutdown of analog broadcasting, and the implications for the mix of services we should be offering on our digital channels.

The need to address our DTV multicasting options has become all the more urgent as stations respond to the March rollout of the V-me (formerly Viva-TV) Spanish-language channel, and the proposed July launch of the nonfiction World channel.

In addition, there is growing concern among some stations about whether the technical quality of our current HDTV program services—which compete with multicasting for DTV transmission capacity—measures up to the increasing number of HD viewing options available from other broadcasters and from cable and satellite.

Add to those pressures the channel space needed by locally assembled channels and educational services promised to state governments, or by datacasting, with its revenue potential, and the demand for broadcast bits quickly outstrips the supply.

What’s the best mix of uses for each PTV transmitter’s 19.4 megabit-per-second digital bitstream? As a system, we need to engage in active dialogue about possible services, develop a shared understanding of our technical options and identify the factors we should consider in making these strategic decisions.

We should at least acknowledge the consequences of the considerable disparities between the choices that stations are making in allocating their DTV bitstreams. These decisions have considerable impact on services for the stations’ local communities as well as on our ability to achieve critical mass by making common cause with each other to create more-than-local program services.

One top-10 market station is multicasting one HDTV plus three standard-definition (SD) services while another has chosen to limit its digital program service to one HD and one SD. Is this difference the result of individual judgment calls, both based on solid information, or should there be a better coordinated approach to address these questions?

But first, back to basics

At the dawn of DTV broadcasting, a transmitter was capable of sending out data at the rate of approximately 19.4 million bits per second (Mbps). The data could be in the form of video, audio or digital files of various sorts—datacasting.

Today, a new DTV transmitter can broadcast at about 19.4 Mbps. That hasn’t changed.

What has changed are broadcasters’ common expectations about how much programming and data can be carried in that signal.

In the early days, some broadcasters argued that all 19-plus million bits were required to transmit a high-definition picture of acceptable quality. The purists maintained that dividing the capacity among more than one program stream would compromise HD picture quality. Many assumed a station would have to choose between broadcasting one high-definition (HD) program stream, or up to four SD program streams.

Though it seems that we signed on our DTV transmitters only yesterday, encoder technology advances in the meantime have already added significant program capacity for stations that invest less than $100,000 in new encoding equipment. Thanks to the use of best-in-class encoding equipment, for example, stations such as KQED-DT in San Francisco are broadcasting one HD plus three SD channels.

In northern Minnesota, where Lakeland Public Television is the only over-the-air broadcaster serving much of its viewing area, Bill Sanford, the g.m. and director of engineering, emphasizes programming choice over the best possible HD pictures. Lakeland PTV is offering viewers one HD plus five SDs.

Both San Francisco and Lakeland are using statistical multiplexing technology (the engineers call it “stat-muxing”), which automatically adjusts the number of bits assigned to individual multicast channels depending upon the needs of each channel’s program content.

For example, a relatively static public affairs talk show or an animated kids’ show require relatively few bits to maintain picture quality. In contrast, programs with a great deal of movement or detail in them — e.g., concerts or sporting events — need significantly more bits. The magic of stat-muxing is that it optimizes overall capacity by allocating, from one frame to the next, the number of bits given to each channel.

For stations that seek to accommodate datacasting along with multicast video services, the multiplexer can add data to the broadcast bitstream on an “opportunistic,” space-available basis.

Who’s doing what?

Public TV’s most recent industry-wide statistics about DTV multicasting were collected by PBS and APTS in November 2005. Stations reported that: 28 percent of public TV’s digital transmitters switch between broadcasting SD and HD at some point during the day; and 47 percent broadcast solely HD 24/7, with one to three SD channels, also 24/7. Of these:

83 stations broadcast HD and one SD, 24/7,
61 broadcast HD and two SDs, 24/7, and
8 broadcast HD and three SDs, 24/7.
7 percent broadcast only HD, 24/7, and
18 percent broadcast only SD. Most of these air two or more SD streams.

Since most stations were broadcasting one or more SD channels with HD, it’s not surprising to learn that 81 percent of the stations were broadcasting their HD at 15.5 Mbps or less (table at right).

There are no hard and fast rules about how many bits are required to broadcast a standard-definition or high-definition picture of “acceptable” quality.

Ultimately, subjective judgment (your eyes) should prevail. As a viewing standard, you should use a recently manufactured 1080p digital set comparable to what some of your viewers might have purchased recently. (For set advice, see the March Consumer Reports.)

What about HD quality?

Those of us in the industry understandably are enthralled by the beauty of HDTV, and we aspire to provide pictures as stunning as any on television. But we should acknowledge at this point in the growth of digital television that the number of viewers actually watching HD is still too small to meet Nielsen’s reportability standards. For example, although you are likely to see claims to the contrary, there is no reliable data about the number of viewers who watched the recent Super Bowl in HD. At least for now, they’re not being measured separately by Nielsen.

The number of viewers actually watching HD is even smaller than the number who own HDTV sets. According to an In-Stat report in December, approximately 60 percent of U.S. HD set owners (15 out of 25 million) don’t access any source of HD programming at all. They tell researchers that they bought their new sets not to watch HD programs, but because the sets were digital—with clearer pictures than their old sets—and they liked the wider screens. Even though they now own HD-capable sets, they report that they’re not willing to pay the additional fees charged by cable and satellite providers for HD services. Apparently, they are not taking advantage of free over-the-air HD viewing options, either.

There’s no question about the direction the market is heading, however. HDTV eventually will become the default television standard, much as color replaced black and white. How many years it will take to reach that “tipping point” is the subject of much speculation. But despite reports from consumer electronics retailers that HDTV sets are flying off the shelves (particularly between Thanksgiving and the Super Bowl), ownership of HDTV sets is still far from universal.

Bruce Leichtman of Leichtman Research Group has been tracking HDTV for the last four years. He reported most recently that only 7 percent of households with annual incomes below $50,000 own an HD set, compared to 26 percent of those with annual incomes north of $50,000. And, for one-quarter of HDTV purchasers, the sets are the second (or third) HD set that they’ve purchased, Leichtman reports. So many receiver sales don’t increase the number of HD-capable households.

Placing too much emphasis on HD—to the exclusion of other multicast service options—is at best premature. Since HD viewership is still small, public TV stations face a fundamental public service question: At this stage in the migration toward HD, how much DTV bandwidth do we want to devote now to HD, when it limits the number of other programming services we can provide to our communities?

We should keep in mind that the prin-cipal drivers of the current promotional hyperventilation about HD are (1) the heated competition for higher-end subscribers among cable, telco and satellite program services, (2) ditto for the sale of big-ticket HDTV sets, and (3) ditto for program service providers’ desire to hold on to cable and satellite shelf space — and to attract viewers — by providing HD services.

The broadcast networks have indeed placed their bets on HD instead of multi-casting. But that results at least in part from their parent companies’ ownership of dozens of cable program services that give them abundant multicast-like program service options.

To make the point: If NBC, CBS, ABC and Fox had to use DTV multicasting to send MSNBC, MTV, ESPN or Fox News and other cable siblings to viewers, the network stations would be broadcasting a lot less HD.

The bandwidth cost of up-converting

The higher cost of HD production has limited the hours of HD programming available, so PBS up-converts about two-thirds of the programs on the PBS HD channel from their original SD formats. Curiously, it turns out, this workaround to ease our way into populating an HD service may be more costly in broadcast bandwidth terms than most of us understood.

The encoders that create the HD broadcast bitstream are unable to distinguish between desirable video and near-random noise in the up-converted programs, and so they attempt to encode and compress whatever signals they are given—noise and all. The end result is that the HD stream occupies significantly more of the bitstream than if it were shot as HD. A station trying to broadcast SD channels in addition to this noisy HD signal pays the price with a cramped bitstream and each program stream looking worse than it might otherwise.

DTV pioneer David Felland, director of engineering and operations at Milwaukee Public Television, estimates that the greater efficiency resulting from higher PBS technical standards for HD might free up sufficient bits at local stations to let each add two or more SD program services, while improving broadcast HD picture quality as well. Felland acknowledges that moving program producers to HD production would be costly, but urges analysis of the costs and benefits.

The news here is that, in the longer term, as the National Program Service eventually moves toward “true” HD, it might free up additional local station bandwidth for SD multicasting.

Local answers to local quandaries

In deciding how to allocate the DTV bitstream, as in many choices, public TV stations will have varied mixes of content from market to market, as well as differing opportunities to avoid a digital traffic jam.

Some stations may find that their local cable operators would be willing partners in distributing program streams that don’t fit—in full quality, anyway—in the 19.4 Mbps bitstream. Most viewing of our DTV programming is likely to be via cable, rather than broadcast, and some stations now deliver programming (both analog and digital) directly to cable headends. With the agreement of your local cable operator, the number of program services you provide need not be limited by your over-the-air DTV bandwidth.

For now, cable carriage of multicast channels may be tighter in markets with more than one public TV station, though not forever. Public TV’s 2005 carriage agreement with the National Cable and Telecommunications Association has different provisions before and after the FCC’s analog shutdown. Pre-shutdown, the agreement obligates cable systems to carry up to four program streams of digital programming from only one public DTV station per market, and only if those four streams are contained within the station’s 19.4 Mbps DTV signal.

After analog shutdown, in these overlap markets, the agreement calls for carriage of multicasts on all local public DTV signals (again, up to four program streams contained in each station’s 19.4 Mbps digital signal). But to win carriage of all four, the stations must satisfy a standard designed to reduce program duplication. If stations collaborate to minimize duplicative content, they jointly will have more capacity on local cable systems.

As we did as a system in anticipation of the arrival of digital broadcasting in the late’90s, and more recently for the Affinity Group Coalition’s work on the future of the local PTV station, we should convene a representative working group to develop alternative scenarios for the continuing evolution of digital broadcasting, multicasting and HDTV.

All of us would benefit from a better-informed decision-making process to choose among the options available, including an assessment of renewed emphasis on free over-the-air broadcasting as a competitive advantage. Informed by better tracking of the consumer marketplace, technological improvements, and audience measurement and behavior, we will know which business models for multicasting and HDTV are viable now and, perhaps more important, which ones are likely to be sustainable. Absent that more rigorous analysis, it’s difficult to develop any consensus about where to place our bets.