Selections from the newspaper about
public TV and radio in the United States
On the DTV decisions agenda

High-def: too-high quality, too-high cost

In the wide range of screen formats offered by digital TV, pubcasters don't need to rush to the top of the line, wrote Bruce Jacobs in 1999. He was past chairman of the PBS Engineering Committee and chief technology officer at KTCA in Twin Cities.

Originally published in Current, June 7, 1999
Commentary by Bruce Jacobs

Quality. Who can argue with it?

Many of us remember a time when it seemed the rest of the world was more committed to quality than we Americans were. Detroit learned a hard lesson when the superior quality of foreign cars caused a dramatic shift in the buying tastes of the public. We saw companies take slogans like "Quality is Job 1." We were happy when American car companies changed their ways and matched our new expectations about how well and how long a car should work.

As creative professionals, we know that quality is a subtle thing. If I were working at a car plant, it would be easy for me to think that a little excess burr of plastic on a gas-cap release lever doesn't matter that much. But if the whole vehicle is made with a similar lack of care, the little burr reminds the driver at every gas station that this is not really a great car. Or worse. And if I have a high-quality car, I can't explain all the reasons why it "feels good." To an extent, the car designers can't either. They do know they tried hard in every way to make the car great.

Many of us cite "quality programming" as a key reason we work in public television. We engineers know that great programming requires solid concepts, writing, directing and shooting. But we also know that a show suffers if we do not have a commitment to the care given to many complex and subtle technical issues for its production and distribution.

It's a constant struggle. First, there are frequent compromises that prevent using the best equipment. Second, we rarely are satisfied that sufficient staff, with adequate training and skilled leadership is available to perfectly maintain and operate the equipment.

The best approach, it seems, is to always push for better. Spend more time on a production. Conduct more training. Test equipment more often. We spend much of our energy planning and budgeting for new equipment that will perform better and work more reliably than the old. Higher quality.

In most cases the replacement equipment has provided visibly better quality. Not everyone at home could see these improvements right away. But over the years, TV receiver performance has caught up. Today, even ordinary TV sets show the difference.

We watched stations that took the attitude, "Who cares, people can't see the difference," reach a point where it became obvious their quality had gone down the drain and they had to completely rebuild their plants.

Now, we have this new technology called DTV that provides high-definition television. DTV was invented, fundamentally, to dramatically improve picture and sound quality.

Who can argue with that?

The value equation

Anyone remember the introduction of the Elcaset? It was an audio recording machine with a large cassette -- about three times the size of today's ordinary audio cassette. It provided clearly superior audio quality, compared to the best cassette machine of the day, which sounded pretty bad. Nonetheless, the Elcaset was a complete and total flop.

Remember the introduction of the compact disc? At its introduction, both the players and discs were very expensive. It provided generally superior audio quality, compared to the best LPs of the day, which often sounded pretty good. Unlike the Elcaset, the CD is a total and resounding success.

What gives?

For something new to be adopted by the public, the perceived benefits must outweigh the costs. The cost certainly includes cost of purchase and cost of ownership, plus many other factors that might require sacrifice, like size, weight, compatibility and convenience. The benefits include indirect gains like prestige, novelty and, of course, perceived improvements in quality, even if the perception is subconscious.

The poor Elcaset was more expensive, larger, less convenient, less portable and incompatible, compared with the ordinary compact cassette. Besides, the advent of Dolby noise reduction and refinements in manufacturing made improvements in the compact cassette's sound that offset the Elcaset quality advantage. Elcaset was also more expensive, larger, less convenient and incompatible, compared with the LP. Its benefits did not outweigh its costs.

With regard to the CD, there were those who argued that it would fail because people don't care about audio quality. They said the public would continue to be happy with LPs -- and cassettes, if they wanted a smaller size. But even if people only have a subconscious perception of better quality, it still makes a difference in their satisfaction. The CD's success was in a major part due to the improvement in quality, which can be heard even from a boom box.

This is not to say that quality alone was enough for the eventual success of the compact disc. We also gained convenience because of the smaller size and random accessibility. The final turning point came when it was possible to manufacture players as cheaply as turntables and compact discs as cheaply as LPs. Then, it just happened, and CDs were everywhere. Try to buy a boom box without one.

We value quality, but as only one factor of many.

The value of HDTV

Is there a viable consumer value equation for HDTV? What success will this new technology have in the marketplace? Is the improvement in quality worth it? Will HDTV be the Elcaset or the CD? How much can or should broadcasters try to drive the market? How important is HDTV for the success of DTV?

A number of factors diminish HDTV's perceived value and conspire against its general adoption by the consumer. A major barrier is the inability to provide a consumer display that can show real HDTV quality.

I can hear the difference between a cassette and a CD even on a cheap boom box. But I cannot see the difference between HDTV and "standard definition" digital on most consumer "HDTV" displays. The quality improvement achieved by these displays seems to come entirely from the elimination of the many NTSC analog impairments and the wider screen -- characteristics of lower definition DTV as well as HDTV. This is a problem, because there is no way to argue for the need to transmit real HDTV, if there is no perceivable advantage!

Why is this the case?

Picture tubes don't work for real HDTV. I spent 30 minutes observing Sony's latest 24-inch studio monitor at close range, and could not see any difference between real HDTV content and the same content converted to "standard definition" 480-line progressive (480p) DTV. The Sony and Panasonic's 30-inch, $40,000 monitors cannot come close to showing the full resolution of HDTV. They can only show about one-third HDTV. To make matters worse, none of these cathode ray tubes (CRTs) are bright enough for home viewing -- the brighter ones have even poorer resolution. The picture tube is fundamentally limited to what you might call "Medium Definition Television" (MDTV).

Real HDTV requires a giant display. Even if it were possible to make a bright enough 30-inch CRT that could really do HDTV, it would not be big enough to show HDTV at home. Even 20/20 vision has its limits.

To see the resolution of HDTV with good eyes, the screen height must be at least one-third of its distance from the couch. Let's put this in real terms: in a living room where the distance is about 10 feet, seeing real HDTV requires an 80-inch (diagonal) display! It is a fair generalization that the true viewing of HDTV requires a wall-sized display. Only front projectors offer pictures that big.

Video projectors require a dark room. Projectors are rapidly gaining resolution and becoming more affordable. There will be projectors that can do real HDTV in the home on a wall-sized screen. But there are a few problems. First, there needs to be a place to put the big screen.

Second, we do not perceive resolution. We perceive sharpness, which is influenced by both resolution and contrast, which requires a dark room. To get enough contrast for real HDTV, the windows need to be covered, and the lights need to be off.

Besides, projectors have light bulbs. You can't leave them on all the time, or they burn out. Lots of people don't have a place for a big screen, or don't want to keep the room dark, or don't want to change projection lamps.

Today's plasma TVs are not HDTVs. The latest great development in home displays is the plasma screen -- marketed as "flat TV." Models on the market are only 3 inches thick and around 40-50-inch on the diagonal. Soon, they will become affordable for the consumer. It is argued that this technology will make it possible for more people to have large screens mounted to a wall. However, in the near term, the plasma displays that will be affordable will not be HDTV. They will be SDTV in the 480p flavor. Today, there is just one plasma panel that is internally HDTV -- the Pioneer 720p display that lists for $19,000. It is only 50-inch -- not large enough to see HDTV at the couch.

Some "HDTV" shows have no HDTV content. If a production does not have many scenes with high detail, viewers will notice no difference between an SDTV and HDTV version. These shows will not entice them to purchase a true HDTV display.

These factors will prevent most of our audience from seeing an advantage to our transmitting HDTV, because they do not have space for a large enough display, or because of the cost of such a display, or because they see little HDTV content on many shows, even if shot in HDTV. They will not see real HDTV quality.

HDTV is not the only choice

When the CD was introduced, what if there had been a competing technology that was nearly as good and nearly as convenient but which cost only half as much? We know what the public would have done.

What if we were able to offer viewers a quality improvement that viewers judge equal to HDTV, at far less cost?

Wide-screen digital passes for HDTV. If you have not done this, go out of your way to view the same content displayed in both 480p SDTV and 1080-line HDTV. Panasonic has a great tape that shows this. You cannot tell the difference between them unless: (1) the HDTV production is top notch, (2) the scenes have a lot of detail, (3) you are viewing a display of 30 inches or larger, (4) you are very close to the display and (5) the display is really good (and expensive).

The reason HDTV has wowed so many people has more to do with it being digital then it does with it being HDTV. In fact, many of the demonstrations called "HDTV" fall far short of real HDTV. The display used for the subjective testing at the Advanced Television Standards Committee in Washington cannot reproduce true HDTV. Most of the compression tests have been done with the resolution intentionally reduced from true HDTV (in order to minimize compression artifacts). The majority of "HDTV" consumer CRT displays, rear-screen projectors and plasma panels, are much closer to SDTV than HDTV. We have seen a lot of SDTV and MDTV masquerading as HDTV.

Just going digital eliminates a host of severe artifacts inherent in the old NTSC analog system. All of the noise and ghosts are gone. The brightness resolution goes up 30 percent. The color resolution doubles. The pictures can be widescreen.

Compare widescreen digital to HDTV. On the vast majority of consumer "HDTV" displays, and on even better displays with much of our content, you cannot tell the difference. Going digital alone is a huge quality improvement over NTSC -- enough to be called "high definition." This quality increase alone is a great reason to adopt DTV.

Why all the HDTV hype?

Why is there such a spoken industry commitment to HDTV, if the value equation is so skewed against it and there is something less expensive that the public will judge equal? The most complete answer can be found in Defining Vision by Joel Brinkley. Suffice it to say that the push for HDTV was not based primarily on any measure of public demand for it. And it remains politically dangerous for the major commercial networks to back away from HDTV, since they insisted to Congress that HDTV is the reason they needed to hang on to so much of the broadcast spectrum.

Today, the manufacturers see their largest early profits in high-dollar displays that they can label "HDTV," even though only one model comes close to true HDTV resolution. The companies who funded the immense amount of HDTV R&D, promote it heavily in order to recoup at least some of their investment.

No one is speaking for the public. But the public will speak with its pocketbooks.

The true cost of HDTV

If DTV were like yesterday's television, this entire argument questioning the value of HDTV could be blown out of the water. Using yesterday's thinking, we would suggest that the producer should decide what quality their program deserves. If they find the money to provide a quality level that only a few people can appreciate, that's their decision. If the producer finds the money for higher-definition formats--1080i or 720p--what possible reason would there be not to transmit this program in the same format all the way to the home?

So what, if only a tiny fraction of the audience is watching a real HDTV display? So what, if most people have a display that is too small, too far away from their chairs, of too low a resolution, or in a room too bright for them to really tell the difference? Why not just go for the best?

Because of the value that viewers would lose. DTV is not yesterday's television, and a purist approach to quality now carries a big, ongoing price.

NTSC provides one channel per station with a fixed, rather low limit on quality. DTV gives each station a giant bitstream that can carry multiple programs of unimagined number and variety, with a fixed upper limit of quality that is extraordinarily high. Dedicating more bits to one channel leaves fewer bits for other channels, for other data and for other enhancements. Dedicating more bits to the quality of one channel reduces both the quantity and quality of other channels. This trade-off is the most radical aspect of DTV. We have only started to talk about how to grapple with it.

DTV can spawn the creation of multiple channels to serve diverse public TV audiences. Limiting some streams to daytime hours, as many public TV leaders expect, would limit the value of these channels. But shouldn't even a high-school channel continue throughout the evening -- offering repeats of daytime services? Shouldn't a "kids" channel continue into primetime?

DTV will also spawn the creation of countless additional "micro-channel" services akin to the proliferation of content on the Internet. We haven't explored this much, but it may be the real reason DTV succeeds -- akin to the way millions of web sites launched the Internet. Each microchannel also needs bits.

If we limit our quantity of services in order to benefit the quality for what the public perceives to be an elite group -- those who can afford giant displays in big rooms -- the public will think we have done them a disservice, and they may be right. For the first time, the elitist charge against public television might be a fair one.

What if every freeway were converted from four lanes to two, to benefit owners of new, wider trucks? What if your car cost four times as much, so that it performed better, though only when driven over 100 m.p.h.? What if the newspaper cost four times as much, so that the ads and feature photos could be printed in collector-edition quality on bond paper? What if your telephone long-distance rates quadrupled, so that you could talk to Aunt Betty in CD-quality stereo?

In all of these cases, even though we would acknowledge that quality is a good goal, our outcry would reflect a judgment that the value equation for each proposition was totally out of whack.

It is high time, in the development of DTV, for us to be exploring how wide each lane should be, what quality of paper to use, what speed the engines should be optimized for, when we really do need to transmit true HDTV quality.

What direction for DTV?

Bitstream trade-off
From highest quality
to highest quantity

Real HDTV
1080i or 720p
Electronic cinema
Right for wall-sized screen
'MDTV'
480p or 480i
Consumer "HDTV" sets
Right for window-sized screen
NTSC TV
525i
Ordinary analog TV
Right for painting-sized screen
Videoconference
Right for book-sized picture
Internet TV
RealVideo
Right for cameo-sized picture

Let's recognize that the situation is both better and worse than we like to think. It's worse, because we have so many choices -- even more than is generally acknowledged. Choice is messy for consensus and for system interoperability. But the situation is also better than we think, because digital technology allows us to have so many choices about how to serve the public. I think we should begin to focus on these four:

Make component digital the minimum standard. Here is a big quality step that has no downsides and many advantages. Converting from NTSC systems to component digital actually makes compression easier (read, more channels) while making pictures visibly better in a big way. Go a step further with a wide-screen version, and you'll have viewers saying "Wow! HDTV!" Our PBS technical center and satellite system is already component digital. With today's digital formats, it's relatively cheap to convert a station to component digital, once we are on the air with DTV. Let's go component now!

When you want higher quality, use 480p and keep multicasting. Component digital is also called 480i. If we use the same number of scanning lines, but go to "progressive scanning" instead of our current "interlaced" scanning, we get 480p. This format offers much better vertical resolution and looks far better than 480i on progressive displays like computers and plasma panels. It also looks much better than 480i when upconverted for 720p or 1080i displays. We can probably do three 480p signals on one DTV transmitter. Perhaps more. The DTV audience will see a huge quality improvement over NTSC, without requiring us to use most of the bits for one show. Transmitting 480p does not require that shows be produced or distributed in 480p.

Judge HDTV. Let's ignore the hype and watch "HDTV" pictures and the public response to them for ourselves. Let's be sure we know when the picture really is HDTV, and how much other factors like dim lights and great sound skew our perceptions. Know when what is being shown is really just MDTV, or even just SDTV. Decide when HDTV is really worth it. Consider the value equation.

Classify your content and be selective about using HDTV. No one suggests that all public TV content should be HDTV -- we agree that the daytime schedule does not require it. And, yet, what of the primetime schedule does? Are we going to produce HDTV versions of every primetime show? If not, which deserve that much bitstream? Which should be 480p, instead? We can only make these decisions when we have a good understanding of the quality provided by 480i, 480p, and HDTV on a range of home displays. We must make this decision with an awareness of what is lost in quantity when we commit many more bits for a minimally perceived increase in quality.

No longer can we simply insist on "nothing but the best," because the penalty for always using the highest transmission quality is too great. We must begin to ask, and answer, for every program, "how much quality increase is plenty enough?" It will be painful. It will be contentious. It requires flexibility and speed. It will pit quality diehards against quantity mavens. It should. Let's get started.

Web page posted June 26, 1999
Copyright 1999 by Current Publishing Committee

EARLIER ARTICLE

Switch to 16:9 format, HD producers advise in 1997.

LATER ARTICLES

Jacobs on the line between high standards and counterproductive mania for "excellence."

By 2004, with tech advances including viewers' screen options, Jacobs admits, he thinks HDTV is ready for priime time.