the myth of HDTV

HDTV is a myth

THE WHAT, WHY, and WHEN of HDTV

What is HDTV? As technology has blossomed over the years, so has the definition of TV image quality grown to match the limitations of the available apparatus. It now seems likely that engineers can produce images, under proscribed viewing conditions, whose quality exceeds the resolution capabilities of human vision. Picture quality is probably exponentially tied to cost and the problem facing the designers and regulators of Home TV is the biggest bang for the buck.

The FCC, on Sept. 1, 1989, with due appreciation of the available state of the art in television engineering and concern for the consumer pocketbook, adopted a tentative decision in "the Matter of Advanced Television Systems and Their Impact on the Existing Television Broadcast Service". Every serious student of HDTV should possess a copy of this excellent overview of the present state of HDTV design and the options open to the FCC for implementing this service in the U.S.

The commission announced three major tentative conclusions. First, the benefits of Advanced Television would be realized by the public most quickly if existing broadcasters are permitted to implement ATV. Second, it will require the initial ATV signals to be either compatible with existing TV (NTSC standard) receivers, or that ATV broadcasters duplicate the program on another channel if the ATV system is incompatible with NTSC. Third, if additional spectrum is needed for ATV, it should be found within the existing
VHF and UHF bands.

You may have noticed that the FCC uses the term ATV (advanced television) to characterize all forms of improved video and audio techniques. The Commission states that "HDTV generally refers to systems that provide quality approaching that of 35 mm film, whereas EDTV (enhanced definition television) refers to systems that perform better than NTSC but not on a par with 35mm film".

A good way to characterize the quality of a television presentation is by the number of pixels (picture elements) contained therein. The number of horizontal lines is another commonly used standard of quality. The NTSC color pixel number is about 250,000, the new HDTV pixel number will be around 1,200,000, and a quality 35 mm static print film jumps to 20,000,000(!), which takes about 4000 lines of electronic means to replicate. Thus, 35mm film would seem to far surpass HDTV, except for the fact that mechanical and film processing deficiencies hold the actual showing to about 800 lines, still better than most proposed HDTV electronic schemes. The color quality and contrast ratio (difference in brightness between black and white) of film is measurably better than the best of 1000 line HDTV. Finally, INmax or OMnimax, a 70 mm creation of Canadian science, has a film area 10 times larger than 35 mm, thus 10 times more pixels, together with a rolling gate, vacuum stabilized projector whose superb pictures have hypnotized audiences in every Disneyworld-like theme park in the world.

The key to more pixels is more bandwidth, or a more effective use of the limited broadcast spectrum authorized by the FCC. While it is true that other transmission competitors are blessed with more space, a common standard, or at the very least, a compatible mishmash, is desirable so that all viewers could share the lower cost arising from mass production of the TV receivers.

Psychophysics is concerned with the interpretation of the Human Visual System in physical terms which can codify the reaction of the human to his environment. An intimate knowledge of psychophysics is indispensable in the formulation of an HDTV transmission system which can induce the same response in the viewer on a lower spectrum budget. In my view, the first and most obvious defect in the attempt of all of the proposed HDTV systems to reproduce the original illusion of human vision is the fact that 2-D displays are not a substitute for 3-D human eyes and brain cells. An illusion of depth can be approximated by a clever director who manipulates perspective, size and color -objects appear distant if they are small and dark and shadowed by the larger players. However, if the screen is enlarged to subtend a viewing angle of 90 degrees or more, the human visual system can be induced into a state of pseudo stereo vision. Witness the effect on the audience watching such a large screen when a roller coaster ride is shown. The screams from the audience mimics the real thing, whereas the same show on a small TV would be boring.

There are several physchophysical aspects of the Human Visual System that helps the HDTV system designer to reduce the bandwidth without affecting the perceived picture quality to the viewer. Flicker is not objectionable if the field rate is above 45, although with high brightness and contrast, 60 cycles is advisable. For a small sized image, say under 30 inches diagonal, the present aspect ratio of 4-3 is eminently satisfactory, as generations of movie goers have proven, with the extremely positive benefit of continuing a past practice at the lowest cost to the consumer. As an old designer of competively priced electronics, I find it very painful in the pocketbook to read the very complex and obviously expensive proposals to do no more than widen the picture a minuscule amount, with darn little improvement in entertainment value. The biggest barrier to HDTV is the visual acuity of the eye which is about 1 minute arc of solid angle for one eye pixel. Practically speaking, this limit to human vision results in no improvement to NTSC if the subject is seated 10 times the height of the picture screen away. But, big brother is on the way (large screen TV), and HDTV is inevitable! The eye is insensitive
to motion which sees motion as a blur, thus the designer can use motion detection and cut the bandwidth sharply as it occurs. Try counting the number of fingers in your hand as you move it rapidly past your eyes! Color flicker can be tolerated down to 12 cycles, and the eye's resolution for color is much poorer than for luminance, thus bandwidth was saved here for NTSC as well as it will be for HDTV. Interlace was introduced as a sort of electronic shutter to raise the flicker rate to 60 cycles and thus to imitate the mechanical shutter of the motion picture projector. Unfortunately, eliminating the flicker brought on many other ills which became more and more obtrusive as the other elements of TV improved their quality.

Psychophysics is the key to reduced bandwidth HDTV, but it is an extremely esoteric and obtuse science for most engineers to master, including me. Any design team, struggling to stuff HDTV into 6 mhz, should work closely with the psychotheoretician and ask critically, "is this minor improvement worth the cost and possible transmission artifacts?". The FCC desires two designs, Enhanced TV (improved NTSC), and High Definition TV ( roughly defined as possessing twice the resolution in both horizontal and vertical directions), and it anticipates an open ended improvement in perceived picture quality for the indefinite future.

Enhanced TV has many options to play with, some so new that their value can only be ascertained with extensive testing under field conditions. My experience is that any new idea takes at least two years from the concept to the production line, and one should add at least one year of user testing before the product is deemed a success. The announced goals of the ATS committee are wildly unrealistic and they and the FCC should be willing to dispense with minor quality but costly increments to the Enhanced TV format.

Apparently the best suggestion for enhancing NTSC is to add a frame store to the receiver and read out the picture at twice the normal speed in a non-interlaced presentation usually called progressive scan. The vertical resolution is increased by about 40%. Interline flicker is reduced and motion interpolation eliminates most of the the spatial artifacts that plague the interlaced picture. Frame storage has another worthwhile feature; if the studio starts with an HDTV camera and transmits the picture elements in four sequential NTSC frames, the frame store receiver can then merge the information into a reasonable replica of the original HDTV picture. I believe that no one has yet solved the motion artifacts.

The worst suggestion, in my opinion, for enhancing NTSC is augmentation via a non-contiguous frequency band. It involves more than the nuisance of a possible second antenna and a receiver, it introduces variables like the signal level, noise content, time delay, shifting ghosts for which there have not been reported sure cures.

Then there have appeared all sorts of ideas to add information via "holes" in the spectrum while retaining full compatibility with NTSC. These schemes add new subcarriers to the visual carrier, mix modulation methods, time share the color subcarrier, add new subcarriers, and on and on. Rigorous field tests will prove the value of these ingenious schemes, but I suspect that the added picture quality will not be worth the cost.

Combing and processing the multiplexed NTSC color and luminance at the transmitter and receiver is now an old art and worth the effort.

Incompatible TV is another ballgame. The state of electronic art has advanced beyond belief -and maybe beyond understanding by us old scholars- and there seems little doubt that a new NTSC committee would create a 1000 line format based on a suppressed midband carrier, quadrature modulated with the high frequencies and with the low frequencies digitally modulated on the vertical back interval, blessed with 4 digital audio channels and formulated for secure scrambling and addressability. If the FCC approves a 6 MHz incompatible HDTV system, then the odds are that in one generation, all receivers will be capable of dual reception and good old NTSC will be no more!

Why do we need HDTV? If technology stood still, if our viewing areas remained the same, if the programming retained its intimate style, NTSC could withstand all criticism. But, none of the above is true. I have been informed by impeachable sources that the next reasonable large screen size will be 1 meter (3 ft diagonal) on which NTSC will look like the projector operator is asleep again. Movies will then be fed from a laser disc looking as good as the neighborhood movie theater. Every other means of delivery except terrestrial broadcasting will have the bandwidth needed for HDTV.

When will HDTV arrive? Once ,upon a distant time, NTSC was called HDTV by its proponents. Currently, the FCC is defining HDTV as possessing approximately 1000 lines of resolution. At the present pace of researching, designing, testing until a winner is selected and made the U.S. standard, we will be years behind the systems in vogue abroad. There cannot be, there will not be, a hiatus on R&D,HDTV will be improved for ever. America is now a country with negligible research funds for TV, and WHEN WILL HDTV arrive is in the hands of others.