Zenith and the HDTV Challenge^*

"Zenith is committed to establishing an American leadership position in high definition television," commented the CEO,^, Jerry Pearlman, in 1988. High resolution flat tension mask (FTM) color monitors, a technology in which Zenith has uncontested leadership, is a cornerstone in Zenith's strategic thrust.

Now it was February 1993, and the Federal Communications Commission (FCC) had just announced that more tests would be necessary before deciding on the high definition television (HDTV) standard. Japan's analog system had already been rejected. Europe's strategy was having a myriad of problems, the most recent being Philip's announcement to suspend plans to mass produce TV sets compatible to the European Community's (EC's) preferred HDTV standard.

Ironically, the United States, which abhors formalizing industrial policy, may end up with the most advanced (digital) standard. Japanese and European policy makers worried more about setting a homemade (analog) standard than about achieving a world class (digital) standard. Adopting an advanced standard and doing advanced research does not, however, guarantee significant industry participation. The success of the HDTV industry depends not only on the TV manufacturers, but also on the simultaneous adoption of the new technology by broadcasting stations, program producers and transmission companies.

Zenith lost $51.6 million in 1991 and $106 million in 1992, even though operating expenses were cut by $38 million in 1990 and $59 million in 1991. At the same time the company had committed nearly it's entire display R&D budget to FTM ($65 million over 5 years). In 1989 Zenith teamed up with AT&T, and in 1991 with Scientific Atlanta, to consolidate its position in the emerging HDTV industry.

A number of critical questions needed to be answered. Should Zenith's technological lead in FTM's be further developed or should Zenith begin developing new technologies? Even if new technologies were pursued for the long run, how could Zenith best capitalize on it's superior FTM technology today? Would HDTV really be a breakthrough product or would it remain an "amorphous" technology? Were the consumers really "asking" for a clearer picture? Would the Japanese eventually adopt the U.S. standard? Was it pragmatic to assume that the U.S. could really gain an up to five years head start by quickly adopting a digital standard? And should Zenith capitalize on its superior HDTV technology and enter into alliances with leaders in the Japanese electronics industry in order to learn and strengthen its future position in consumer electronics?

HDTV

For the first time in almost fifty years the definition of American television was about to change with the development of HDTV. Today's television will be replaced by a system that offers sharp, clear and larger images with excellent audio and is "smart" like a computer.

Initially, HDTV sets will probably be large screen models. However, HDTV will require the integration of a number of technologies: semiconductor, software, multimedia, displays, compression, VCR, compact disc (see Exhibit 1 for details).

The HDTV standards winner would gain the right to royalties on any HDTV set or broadcasting system using its technology. Dale Cripps, editor of the HDTV newsletter, estimates royalties at $5 a set or about $110 million/year assuming 22 million HDTV's will be sold at the peak of acceptance. Royalties from broadcasters could amount to a few million more. The head of the Advanced TV Advisory Committee, former FCC Chairman Richard Wiley, encouraged contenders to combine forces to move toward the standard faster. Thus, contestants may be forced to share their bounty.

Electronics Industry and HDTV

In 1992, the consumer electronics industry was waiting for a new growth stimulus. U.S. annual sales of consumer electronics were $33 billion, Japanese sales were $35 billion and European sales were $42 billion. Historically, the industry has been driven by major innovations and life cycles of the gramophone (1920s), the radio (1930s), the B/W-TV (1950s), the color TV (1960s), Hi-Fi (1970s) and the VCR (1980s). Recent product introductions such as CDs, DAT and 8mm camcorders, however, have not presented major technological breakthroughs, nor has the trend toward product proliferation refueled sales.

Table 1: World Electronics Industry¹

Product 1991 World Sales

(in $ billions)

TV Hardware 7.0

VCR 3.0

Camcorder 3.0

TV Advertising 29.0

Movies 6.0

Video Cassette Rental 14.0

Video Games 5.0

PC Hardware 50.0

PC Software 12.0

In 1991, annual sales growth rates of consumer electronics have decreased from 50% to 3.5% with 99% of all Japanese and U.S. households having at least one color TV, 80% a VCR and 65% a stereo system. Unfortunately, many fear that the move to HDTV will be far less successful than the move from black and white to color.

In response to lower margins, shorter product life cycles, lower sales growth and competition from Taiwan and Korea, the industry has been consolidating. In Japan, Akai was taken over by Mitsubushi and Aiwa became a subcontractor for Sony. Thompson, the French company (82% owned by the state), took over Thorn. Major German manufacturers, Blaupunkt and Telefuken have also been acquired by Thompson. Both firms are dependent upon product, process and competent technology from Japan. For example, DAT and VCR technology are supplied by JVC, video-disk player technology is supplied by Pioneer and comcorder technology is supplied by Hitachi. Philips, a global electronics firm of Dutch origin, reported a loss of $2.5 billion in 1991.

In 1992, three firms, Sony, Matsushita and Sharp held the strongest position in the emerging HDTV industry. Unlike their weakened world competitors, these companies sustain their R&D spending at more than 6% of sales. As a result, HDTV may be the opportunity for a few Japanese companies to take a major lead in consumer electronics.

Distribution Trends

During the 1980s, the users of VCR's and CD players drew annual increases of 15% to 20% in sales for retailers. Traditionally, manufacturers have used a two step distribution process in which the distributors provide after sales service through parts availability, advertising support and warehousing. However, this distribution system has become too expensive due to shorter product life cycles, lower margins and increased product variety. Further, as cost reductions in manufacturing have become more elusive, producers have focused their attention on distribution costs. Labor, for example, only accounts for 10% to 20% of manufacturing costs. Thus, manufacturers of late are enforcing minimum order requirements from retailers. Hitachi, for example, requires a $15,000 minimum order.

As a result of their distribution economies and purchasing power vís-a-vís manufacturers, electronic superstores have replaced small retailers, distributors and department stores. They enjoy large advertising budgets and an assortment of 2,000 to 2,500 stock keeping units (SKU) along a large continuum of price points. In addition, they receive price discounts for early purchase commitments. In 1991, Circuit City (a leading U.S. retailer), had sales of $5.5 billion through 200 stores. It dominated the metropolitan areas and offered extensive customer service, such as home installation. In 1990, 220 retailers accounted for 70% of total consumer electronic sales in the U.S. Twelve accounts alone represented 65% of Thompson's sales in the U.S.

HDTV and Multimedia Trends

As a digital system, the TV will change from a passive receiver to an integrated interactive entertainment center. As a telecomputer with multimedia capability, HDTV will affect electronic products and entertainment markets. While digital HDTV is being developed, intermediate technologies have arisen. Large monitors with integrated video recorders and CD players are a case in point. These intermediary systems are lower in both price and quality. Offering systems at this intermediate stage of industry evolution, however, educates consumers, establishes brand names, develops distribution systems, influences standards and produces feedback for future HDTV designs. Further, alliances in this market might develop standards in software and hardware which might become important in HDTV.

A computer attached to a HDTV could, for example, customize a movie's sound track, colors and ultimately the viewing perspective. Furthermore, the high demands that video processing poses on semiconductors has accelerated the rate of semiconductor innovation benefiting all areas of computing. In 1991, total world semiconductor sales were $55.5 billion per year. By the end of the decade, the consumer electronics portion of semiconductor sales alone is forecast to reach $48 billion in annual sales.

Japanese HDTV

Initially, Japan led the world by pursuing an analog transmission technology: MUSE (multiple sub-nyquist sampling encoding). This technology—developed in 1965 as a hybrid digital-analog system—is already obsolete according to some analysts.¹

HDTV is still relatively expensive in Japan. Sharp has announced a low-end HDTV which will cost one million Yen, compared to current HDTVs for Y3-Y5 million. This price reduction has been achieved at the expense of picture quality. Sharp bases its strategy on a market forecast that at a price of Y1 million, 16% of all Japanese households would buy a HDTV set. Given the trend toward large screen TVs in Japan (50% of TV screens are 32-inches or longer), the ministry of Post & Telecommunications (MPT) projects that 40% of all Japanese households will own a HDTV set by the year 2000. This represents a market value of $9.6 billion.²

Sharp's product introduction was backed by the Japanese broadcasting organization (NHK), which increased HDTV broadcasting from one hour to eight hours in November 1991. Sony and Matsushita Electric Industrial have already spent hundreds of millions of dollars developing manufacturing facilities for HDTV. Based upon historical evidence of substitutions, the price of HDTV must decrease to within three to four times the price of a color TV, for a mass market to emerge (it took the color TV twelve years after its introduction to outsell B/W). This would correspond to an installed base of about $40 billion worldwide by the year 2000.³

As of 1992, three main groups were developing the chip set for MUSE. NHK, NEC, Toshiba and Matsushita were able to reduce the number of chips needed for the Japanese HDTV standard, MUSE, to about 100. Fujitsu, Hitachi, Sony and TI working in tandem, have nearly completed co-development of a thirty chip set for MUSE. LSI Logic and Sanyo announced that they intend to reduce the number of chips required to less than ten. Logic-Sanyo were the first to attempt building a chip set with only application specific integrated circuits (ASIC). Meanwhile, LSI Logic has joined another consortium including NEC, Matsushita, Mitsubishi Electric, VLSI Technology and Sharp, which plan to reduce the number of chips down to 19. In each of these consortia, no U.S. firm contributes digital signal processing. TI is responsible for developing frame memory circuits. VLSI is working on color processing and LSI on circuits for sound processing.

To arouse public interest in HDTV, the Ministry of Post and Telecommunications (MPT) provides Japanese cities with systems through the High Vision Program. The High Vision Center Project was started in 1987 to set up pilot tests in Japanese cities for public HDTV theaters and public centers for interactive access to large scale databases. The Key Technology Research Center in conjunction with the Ministry of International Trade and Industry (MITI), the MPT and the Ministry of Finance, are subsidizing development of critical HDTV components. This is achieved through low interest loans, extending a 26% permanent tax credit and allowing accelerated depreciation of assets in the electronic sector.

European HDTV

To counter the threat of Japanese domination of HDTV, the EC has invested nearly $2 billion to create a rival system, MAC. MAC has largely been developed by Thompson and Philips in a public policy vacuum. In December 1991, Thompson announced a $6000 TV which uses an intermediary D2MAC standard. On December 19, 1991, EC Telecommunication Ministers agreed not to force an immediate conversion of current broadcasters and satellite channels to MAC. Channels launched after 1994, however, will be required to cane MAC broadcasting equipment.

With the demise of British Satellite Broadcasting, the only European broadcasters currently offering D2MAC TV are Canal Plus and Antenne 2, both French companies. Thompson and BIS, a Bosch-Philips joint venture, delivered a thirteen hours per day HDTV service from the 1992 Winter Olympic Games. The project was funded by the French government, the French Telecom and the EC Commission. Without software or broadcasters, MAC has been advancing slowly.

The European market is expected to reach $20 billion by the year 2000, with an installed base to 500,000 units. By 2004, only 9% of European households will own digital HDTV sets. Since Europe has a negative trade balance of $35 billion in electronics, however, European champions like Bull, Thompson, Philips and Siemens may lobby to protect the market. Both Europe and Japan, who support the analog HDTV system would, it seems, eventually adopt the digital system.

United States HDTV

The American electronics industry is losing the war in consumer electronics according to a study commissioned by the Department of Commerce and performed by the International Trade administration. The study found that Japan has more than doubled the number of U.S. electronic patents that they have received since the mid 1970s. Further, their patents have been licensed more frequently than those of Americans.⁴ This report predicts that, at the current pace, Japan will become the world's number one electronics producer by the early 1990s. This comes as quite a shock to people who think electronics will be the largest employer by the end of the century. A National Science Foundation report on Japanese HDTV research warns that Japan is focusing on the new technology as a catalyst to gain dominance over the semiconductor, computer and telecommunication world markets.

A U.S. firm, ITT Semiconductor, sells more than half of the world's microprocessors used in conventional TVs for such features as frame within a frame. IIT's Devoe claims that many U.S. firms serving niche military markets have the technology to compete in HDTV, but want to avoid "slugging it out in the marketplace." Japanese semiconductor manufacturers were rebuffed when Fujitsu's proposed purchase of Fairchild Semiconductor was blocked by then U.S. Commerce Secretary Malcolm Baldrige in 1987. Instead, a series of alliances are forming.

Zenith

Zenith's current strategy is in keeping with the company's history of parenting standards. The company invented the remote control in 1956, the FM radio stereo transmission standard in 1968 and the multichannel television stereo (MTS) transmission system in 1986. "There are more fundamental inventions from Zenith in today's television than from all the Japanese companies combined," says James I. Magdid, a senior advisor with Needham & Co., Inc. in New York.⁵

Since 1988, Zenith has phased out or cut back operations in several key (non-high resolution) products. Total nonvideo products accounted for only $78 million in sales in 1991, down from $136 million in 1989. (See Exhibits 2–6.) In 1991, Zenith reduced capacity in magnetics production as well as announcing plans to phase out automotive electronics, low end color computer monitors and monochrome monitors. In 1990, power supply and other electronic component production was restructured and downsized worldwide. Finally, in a major move in 1989, Zenith sold its computer products division to France's Groupe Bull.

Computer products accounted for over half of Zenith's sales in 1988 and was its only profitable division. The $450 million that Zenith received from the sale allowed it to significantly reduce debt and focus more R&D spending on their FTM and HDTV technologies. These actions improved their balance sheet, but the market value of the company still continued to decline through December, 1990, reducing Zenith's market value to nearly half its book value at one point. By March of 1992, however, the market value had rebounded to about 75% of book. (See Exhibits 2-6.)

In addition to product line phase-outs, Zenith has rationalized operations worldwide. Because of the price competitive nature of consumer electronics, cost sensitive producers have chosen Mexico for its low labor costs and "the perceived benefits of the possible (North American) free trade pact." The number of workers at the Springfield, Missouri plant was reduced by 75% in March, 1992 when TV assembly was consolidated in Mexico. This plant is scheduled to be shut down sometime in late 1993.⁶ Zenith predicts that "virtually all of its television production will be located in Mexico" by the end of 1993. "The cost savings of Mexican production, where hourly wages are only about a tenth of those in the U.S., have proved too critical to pass up for loss-making Zenith." Jerry Pearlman predicts that Zenith's costs would be $400 million higher if it were not for their Mexican operations.

Following these restructurings, Zenith operates in two major segments: consumer electronics and components. Consumer electronics accounts for 85% of total sales, with color TV representing 75% of these sales. Approximately $100 million sales in color picture tubes are derived from OEMs. Component products include power supplies and low resolution displays.

During all this cost cutting, Zenith increased R&D spending on HDTV and FTM screens. Through 1991, Zenith had spent $25 million to develop the DSC-HDTV system. This is a fraction of the $1 billion spent by Japanese firms to develop what many consider to be an inferior standard. Zenith has had similar success with a shoestring budget for FTM monitors.

What remains to be seen is if Zenith can capitalize on its technology and create a manufacturing base capable of supporting demand for HDTV. Zenith's current CRT manufacturing facility is a 40-year old cookie factory that was converted to CRT screen manufacturing in 1967. The spaces are small and cramped, while the factory lacks the modern systems that typify today's high-tech manufacturing. For instance, there are no automated materials handling systems so the workers must lift bulky CRT screens on and off a conveyor belt. Similarly, workers must manually perform the exacting operation of mating the shadow mask with the phosphorous coated screen. Yet, for all their apparently antiquated work habits, in 1992, Zenith workers—running three shifts a day—turned out 3.6 million picture tubes a year. Working with aging equipment and techniques, they still manage to manufacture a picture tube in 22 hours. Matsushita, in Troy, Ohio, does it in just under 24 hours at a new, well-equipped plant.

Since current CRT plants are at capacity, Zenith needs to develop new manufacturing plants to supply an increased demand spurred by HDTV. These plants will probably be located in Mexico. A more fundamental question is whether Zenith should sink more money into CRT manufacturing when investment is needed for future technologies.

Not everyone associated with Zenith is excited about the high resolution standard. In March 1991, Sal Giordano, owner of Nycor, a New Jersey based holding company and 8% shareholder of Zenith, sent proxies to all Zenith shareholders asking them to vote for the Nycor nominees for the ten member Zenith board. The letter indicated Nycor's dissatisfaction with current management strategy and outlined their plan to turn the company around. "The most serious problem at Zenith has been and continues to be management's lack of a clear strategic plan. Though Zenith management has trumpeted the existence of a long term strategy for the company, to date they have failed to articulate to stockholders any plan which explains how Zenith will return to profitability.⁷ In an effort to fend off this attack, Zenith issued $15 million in common stock (a 5% interest) to Korea's Goldstar. Goldstar and Zenith have been working together on several trade and technology agreements. Nycor candidates were defeated two to one at the shareholders meeting, giving a strong vote of confidence for current management's strategic direction. (See Exhibit 9.)

Zenith and AT&T

Since Zenith did not have capabilities in digital compression and semiconductors to bring its system from the drawing board to the lab, the company teamed with AT&T in 1989 to develop the Digital Spectrum Compatible (DSC-HDTV) system. Under this agreement, AT&T Bell Laboratories develops video compression and AT&T Microelectronics develops semiconductors. The DSC-HDTV system has unique digital transmission properties, based on technologies developed by Zenith, to prevent interference with standard TV signals and to transmit snow-free, ghost-free pictures and compact-disc quality audio.

AT&T and Zenith have become extrememly successful bringing the HDTV system together. As noted by Paine Webber analyst Jack Grubman, AT&T uses compression technology in its core long-distance business. "They know how to do it better than anybody, if they can transfer that to another application, so much the better."⁸ In addition to technology transfer, Zenith's $25 million investment in DSC-HDTV "has been more than matched by the substantial investments of (their) HDTV partners, particularly AT&T."⁹ Over a thirty-nine month span, this partnership was able to develop and improve Zenith's paper system into a working prototype.

Harris Allied Broadcast Division and Hewlett Packard (HP) announced plans in 1993 to develop broadcast equipment for the DSC-HDTV system and to license DSC-HDTV technology if the FCC adopts the Zenith-AT&T system. Harris would support the rapid deployment of HDTV modulation and transmitting equipment to broadcasters following FCC adoption of the new broadcast standard. Likewise, HP would develop HDTV encoders (the electronic systems used by broadcasters to process and compress signals).

In 1993 the digital HDTV Grand Alliance was formed. It merges technologies from the three groups that had developed all-digital HDTV systems for consideration as the U.S. standard: Zenith and AT&T, General Instrument Corporation and the Massachusetts Institute of Technology, and a consortium composed of Thomson Consumer Electronics, Philips Consumer Electronics and the David Sarnoff Research Center.

Zenith and the other Alliance members are pooling their resources, skills and technologies to create a "best of the best" solution for the HDTV system being proposed to the FCC. The system elements for digital video compression, transport, scanning formats and audio all were selected by the Grand Alliance in October 1993.

In early 1994, after extensive competitive laboratory testing, the Grand Alliance selected Zenith's VSB (vestigial sideband) technology as its terrestrial broadcast and cable HDTV transmission subsystem. The 8-VSB system assures broad HDTV coverage area, minimizes interference with existing analog broadcasts and provides immunity from interference into the digital signal.

AT&T has become particularly aggressive in the last few years in its bid to excel in semiconductors, an increasingly important element within the communications industry. AT&T's Bell Labs was a pioneer in the software industry and developed both the C language and the UNIX operating systems to meet in-house software development needs. Large internal demand for semiconductors has made AT&T the world's largest producer of standard cells, with annual sales of about $250 million. Purchasing the world's third largest standard cell producer, NCR, propelled AT&T even further into the lead. In 1992, AT&T had eleven manufacturing plants and twelve design centers for semiconductors. Despite AT&T's large internal market, their penetration into external markets has been limited.

Antitrust concerns initially hampered AT&T's ability to compete in the computer industry. In 1985, AT&T finally entered the merchant semiconductor industry but with little success. Two years later, over fifty middle and senior level managers with extensive industry experience were brought in to revitalize AT&T's semiconductor initiative. AT&T is developing semiconductors not only for HDTV, but also for video conferencing, cellular telephones and network computing.

GoldStar and Scientific Atlanta

Support for Zenith's DSC-HDTV has also come from GoldStar and Scientific Atlanta. In addition to GoldStar's $15 million capital investment in Zenith stock, GoldStar has entered into a licensing agreement to market Zenith's FTM screens for televisions in Korea. The agreement calls for GoldStar to pay a fee to Zenith for each TV CRT, on top of a substantial up-front fee. The deal allows Zenith to expand its technologies to previously untapped overseas markets.

In late 1992, Zenith and GoldStar announced the joint development of a digital high-definition VCR for home use. The new digital HD-VCR, developed for the DSC-HDTV system, is designed to record HDTV signals on standard Super-VHS videocassettes. The HD-VCR also would be able to record and play back programs in today's TV format using standard VHS tapes.

Scientific Atlanta is a technological leader in satellite and cable transmission of television signals. In December 1991, Scientific Atlanta pledged its support for the DSC-HDTV system as the U.S. standard. Under agreements developed with Zenith and AT&T, Scientific Atlanta develop the transmission and conditional access systems and equipment necessary to allow secure satellite transmission of the DSC-HDTV signal to make HDTV programming available to all markets and to support a timely market introduction.

In addition to satellite signals, Zenith and Scientific Atlanta announced that under a separate agreement they have developed a common transmission structure for carrying the Zenith/AT&T digital HDTV signal and Scientific Atlanta's digitally compressed conventional TV signal through cable plants to the home. This will combine Scientific Atlanta's Vector Quantization digital compression system for conventional TV signals with Zenith's four level vestigial sideband modulation technology for cable transmission. Similarly, Scientific Atlanta will utilize a common transmission structure in the satellite system for carrying the DSC-HDTV digital signal and for Scientific Atlanta's digitally compressed conventional TV signals.

Hence, both the HDTV signal and the Vector Quantization digital compression signal will have end-to-end transmission capability over satellite and down the cable to the home. Programmers and cable operators will benefit from the flexibility and cost advantages. The U.S. consumer will share in these benefits not only because of transmission efficiencies, but also because the same high definition television set tuner will accept HDTV signals off the air or down the cable, resulting in lower cost TV sets and cable boxes. By combining forces (with AT&T and Scientific Atlanta) on digital compression and transmission, Zenith will be able to bring high performance, low cost digital video to the cable industry, and at the same time lay a compatible transmission foundation for the smooth adoption of HDTV.

HDTV Technology

HDTV systems mandate the integration of various technologies, from the production and transmission of programs, to the television terminal at the user's home. There are many emerging technologies capable of achieving the requirements of HDTV. With rapid development of high resolution displays, compression technologies, DSP chips and VLSI capabilities the arena is wide open when considering which technology would ultimately be dominant (see Exhibit 1 for details). Following is a brief description of each parameter.

High Resolution Displays

High resolution displays are much more than just a component of HDTV systems. As Robert Duboc of Coloray puts it, "Displays become the tail that wags the dog."¹⁰ Within the HDTV industry, the high resolution display is the single most expensive component and the component that most effects image quality as perceived by the viewer. Outside of the HDTV industry, high resolution displays are already used for everything from computer screens to car consoles. The defense, airline, space and medical imaging industries have employed high resolution displays for years. Sales are expected to reach $9.3 billion by 1996, 40% of which will be in consumer electronics with the balance in computers.

High resolution displays for HDTV must be defined with numerous parameters. Resolution is quantified as the number of horizontal strings of dots (scanning lines) that fill a screen (frame) and the rate by which new screens appear (refresh rate). Screen dimensions are stated as the ratio of the screen's width to height (aspect ratio).

There are various technologies in the development stage that provide improved clarity and contrast from that of the traditional CRT terminal (see Exhibit 1 for details). In 1992, Zenith had the lead in flat tension monitors (FTMs), but considering the research being put into this area and the options available, the future is highly uncertain with regards to which technology will emerge most effective, yet economically viable.

Software Availability

An important variable in determining HDTV demand will be the availability of software. HDTV broadcast equipment will require a large initial investment by broadcasters not to mention larger camera crews (up to 3 times as large) since flaws are magnified. Some of these extra costs are mitigated by decreased production costs realized by eliminating film.

A study by the Transition Scenarios Group of the Advanced TV Advisory committee has determined that broadcasters in large markets will be ready to broadcast HDTV within eighteen months of a standard. Smaller broadcasters will, however, be at a disadvantage due to capital demands for new equipment. A TV station might have to pay between $12 million and $22 million to convert to HDTV. The average price of a TV station sold between 1987 and 1990 was only $19.7 million. HDTV stations might also service a smaller area causing perhaps a 20% decline in revenues.

To alleviate conversion costs, Zenith-AT&T have developed a simple up-conversion process that will allow local TV stations to convert their existing NTSC equipment to Zenith's DSC-HDTV format. Through a two-step process the conventional NTSC signal is stepped up to a high definition format compatible with DSC-HDTV. "With a minimal investment, TV stations would be able to move to high definition signal and the up-converted locally produced signal."¹¹

The FCC has given away HDTV licenses to 1700 stations to encourage investment by the existing broadcasters in HDTV. There will be no competition for HDTV licenses. The value of these licenses is estimated at $1 billion. Alfred Sikes, chairman of the FCC, states that "it's not a give away, it's a transition from one format (of broadcasting) to another."¹² Two months before the license announcement was made, 100 broadcasters petitioned the FCC to require cable to carry HDTV to ensure an adequate audience to justify investment. It is feared that without fresh competition, existing broadcasters will cling to the NTSC standard. Public demand for HDTV is likely to materialize mainly in response to broad program selection. Studios will be hesitant to produce programs in two standards until broadcasters are able to transmit HDTV.

HDTV & Semiconductors

HDTV represents a major boost to the semiconductor industry. Not only will HDTVs use large quantities of semiconductors internally but they will stimulate sales of complimentary semiconductors for applications ranging from multimedia to photography. (See Exhibits 7 and 8 for related information on the semiconductor industry.)

In the capital intensive semiconductor industry, large up-front capital expenditures have encouraged increased consolidation. The investment in R&D and manufacturing for 16 bit dynamic random access memory (DRAM), for example, exceeds $1 billion. It is estimated that companies have to produce at least 5 million DRAM chips per month to recoup their initial investment. As a response to high investment and short life cycles, companies specialize in one stage of the R&D process or in a particular product. Toshiba, for example, supplies memory chips to Motorola and Motorola reciprocates with microprocessors. Further, Toshiba transfers production technology to Siemens.

HDTV looks like it will transform to digital the last high volume analog consumer electronics segment, TV. Analog refers to information represented in wave form and digital refers to information represented in numeric form. Sound and image naturally occur in wave form, thus digital systems must first translate waves into numbers and later recreate the waves based upon the numeric descriptors. Digital HDTV systems have many advantages over analog including less distortion and more flexibility. The major disadvantage of a digital system is the incredible volume of numbers that must be processed to relay HDTV images.

One second of HDTV data requires one billion binary (0s and 1s) numbers of code. To put the volume of numbers to be processed into perspective, it would require appproximately 360 personal computers with standard 40 megabyte hard drives to store a two hour HDTV program uncompressed. Secondly, in terms of broadcasting bandwidth, uncompressed digital HDTV would require 100 times the signal capacity available on current 6 MHz band width channels. The solution lies in compressing the signals before transmission. This solves the bandwidth problem as well as reduces the volume of numbers that must be processed, but increases the number of operations that must be performed on each signal, thereby developing the need for parallel processing and digital speed processing (DSP) chips. It should be noted that the increased resolution and aspect ratio of HDTV would still surpass the current channel capacity by a factor of four or five times even if we use conventional analog signals.

Digital HDTV was long thought unfeasible because it would require the processing power of a super computer. Recent semiconductor advances in video compression (VC), digital speed processing (DSP), digital to analog converters (DAC) and very large scale integration (VLSI) make it highly likely, however, that a digital standard for HDTV will be implemented in the U.S. in the next few years.

Video Compression

General Instruments announced that it had found a way to compress HDTV signals to a manageable size in June 1990, two days prior to the deadline set by the FTC for standards proposals. Three other consortiums, Zenith-AT&T, M.I.T. and Philips-Sarnoff (the Grand Alliance) soon announced competing digital standards. The digital video compression strategies that have been devised involve several components, including coding options, quadrature modulation, motion compensation and error correction (see Exhibit 1 for details). Together, these compression techniques allow an HDTV signal to be compressed by a ratio of approximately 100:1.

Digital Signal Processing

Compression reduces the HDTV signal size to a manageable 1.2 MB/second but it vastly increases the number of operations that must be performed on each signal. It must be encoded, compressed, decompressed and debugged. During the encoding process, a series of compressions must take place sequentially. Each compression, in turn incorporates the results of multiple calculations for such compensatory measures as motion correction. These calculations may take place in parallel. The same process occurs in decoding followed by an elaborate debugging process. Thus, broadcast HDTV is constrained not only by bandwidth, but also by affordable processing power. Digital speed processing (DSP) chips promise to deliver the most appropriate computing power for the least price. This is achieved by their multi-tasking capability that allows them to perform several applications simultaneously, such as playing music and receiving faxes.

Several semiconductor manufacturers have already introduced high performance DSPs for HDTV image processing. National Semiconductor has introduced a 50 MHz DSP capable of 100 million instructions per second (MIPS). AT&T Microelectronics has already produced a DSP for HDTV and is currently working on a DSP capable of 10,000 MIPS. Due to its ties with Bell Labs, AT&T has unique strength in algorithm design. In 1990, AT&T achieved a 10% market share in the $300 million DSP market, which is expected to grow to $4 billion by the end of 1996.

Digital Analog Converters

DACs are required to ultimately translate the digital signal back to an analog signal to drive the speakers and the display. Fully integrated and temperature tolerant 12 bit DACs have only been available a few years at low cost, further increasing the price performance ratio of HDTV.

Very Large Scale Integration

Considering the vast increase in the number of chips required for processing HDTV data, VLSI along with faster DSPs, will be important catalysts for reducing HDTV costs. Initial systems required about 2000 chips, arranged on what has been likened to shelves in a refrigerator. Thus, semiconductor components currently comprise 33% of HDTV costs. Miniaturization advancements in both manufacturing and design will drastically lower the number of chips required for HDTV. Faster processors will be able to perform more tasks per frame, further reducing the number of chips required. Fewer and smaller chips translate to a greater HDTV price-performance ratio.

Trade Lobbies

Various trade associations, directly or indirectly effected by the emerging HDTV industry, independently lobby to safeguard and promote their interests.

Semiconductor Equipment and Materials International represents firms that build the machine tools for primary HRDs. Their main goal is thus to develop a U.S. manufacturing base for HRD. Their lobby efforts focus on the Defense Advanced Research Project Administration (DARPA) funding for HRD, and only indirectly on HDTV.

The American Electronics Association (AEA) concluded in 1989 that the Japanese would use HDTV technology to overtake the U.S. in semiconductors.¹³ Further, they felt that losing in HDTV could jeopardize the U.S. electronics industry. To ease the transition, AEA recommended moving to EDTV (enhanced) as an intermediate step before going to HDTV. This concept is supported by the Advanced TV Research consortia, however the FCC has made it clear that they will not accept a piecemeal standard.

A study commissioned by the AEA in 1990 recommended a $500 million government loan program and another $500 million in loan guarantees to support U.S. HDTV research. AEA spokeswoman Sue Wirth notes that this study has never been formally adapted by the AEA. The AEA's lobbying activity supports the component technologies that produce HDTV. In fiscal year 1993, the AEA lobbied for $75 million from DARPA and an additional $5 million from other organizations. Further, the AEA has established an Industry-Government Advisory Board to encourage domestic electronic manufacturing.

The National Academy of Sciences recommended in March 1992 that a new self-funding Civilian Technology Corporation (CTC) be created. The CTC would promote HRD technologies, notably HRDs and microelectronics. Funding would be provided by a onetime $5 billion appropriation, with the rest of the financing coming from licensing fees and patents. Senator Hollings, Democrat from South Carolina and chairman of the Senate Commerce Committee endorses the CTC. This is considerably more than the $100 million effort from which Sematech was founded.¹⁴

The National Association of Broadcasters (NAB) made the mistake of inaction with aid stereo and they do not want to repeat this mistake with either digital audio (DAB) or HDTV. This group has moved to protect industry interests by successfully lobbying the FCC to reserve unused spectrum for HDTV. The NAB questions if the FCC proposed 15 year conversion period from NTSC to HDTV is realistic.

Government Lobbies

The U. S. does not have a comprehensive national policy addressing the basic research, commercialization or manufacture of the electronics industry. By default, defense concerns and a myriad of special interest constructs serve as public policy. Throughout the cold war the federal government supported basic research with defense applications. DARPA coordinated the mainstream of this effort.¹⁵

HRD in particular has been targeted by DARPA. The agency has financed R&D projects for such companies as Texas Instruments, Projectivision, Photonics, Ovonic, Norden Systems, MRS Technology, Microelectronics and Computer Technology Corp., Tektronix, Sarnoff Research Center and Zenith.

Legislative Branch

Trade Barriers: The 1988 Trade Act allowed the U.S. to target unfair traders and impose retaliatory measures, the retaliatory provision is commonly known as Super 301. In 1992, the Bush Administration threatened to use it against Japan, and it is credited with opening Japanese lumber, semiconductor and super computer markets. In 1990, the Super 301 clause was removed (vís-a-vís Japan) and replaced with the Structural Impediments Initiative (SII).¹⁶ This move was considered a "serious tactical mistake" by Senator John Danforth (R-MO), the ranking Republican on the Senate Commerce, Science and Technology Committee and a key author of the bill.

In August 1991, the FTC levied a 63% duty on LCD displays from Japan which comprised 97% of the U.S. supply. A group of small U.S. producers filed a petition under the Tariff Act of 1930 asking for tariff protection, alleging that the Japanese manufacturers were selling screens at below fair market value.

Since this tariff does not apply to finished products, such as laptop computers, manufacturers have been forced offshore. In 1991, IBM and Apple petitioned the Commerce Department to set up a foreign trade subzone for imported Japanese AMLCDs, to avoid paying the 63% tariffs. An approved foreign trade zone would allow firms that manufacture in the zone to import parts duty-free for assembly.

Antitrust Policy: One option is to relax antitrust enforcement on domestic competitors, relevant market is defined globally and antitrust enforcement is based on world share. This assumes that domestic firms will become stronger, however it does not assure reciprocal access to foreign markets. Another path is to export U.S. antitrust policy to ensure that foreign markets encourage new entrants. Policies are tempered to local economies, attacking vertical monopolies (e.g., the Japanese Kierstu), as well as to traditional horizontal monopolies.

Committee Work: Many measures have gained limited support within Congress. For example, the American Trade Preeminence Act contained provisions for suspending antitrust, a $400 million research provision to create a commission to study HDTV, the hi-tech industry capital costs and to evaluate a new cabinet post for technology policy, passed the House Science, Space and Technology Comrnittee by a vote of 49-0 in March 1990. The primary author of the bill was Congressman Don Edwards (D-CA), the antitrust provisions were written by Congressman Tom Campbell (R-CA). House leader Richard Gephardt (D-MO) favored a "managerial industrial policy" to restore American competitiveness. Toward this goal he, together with Senator John Glenn (D-OH), supported the formation of a civilian DARPA. Congressman Edward Markey (D-MA) chairman of the House Subcommittee on Telecommunications and Finance (which has jurisdiction over the FCC) proposed federal funding of HDTV research in February 1990. In 1992, none of these initiatives had received the necessary support and funding, so HDS is funded primarily through defense associated grants.

Information Gathering: Within Washington, various agencies have been established to filter information for an informed legislative decision. The Congress Office of Technology Assessment was created to assist legislators in determining the potential of new technologies. The office is administered through the Commerce Department and congressional direction comes from the House Committee on Science, Space and Technology. In June 1990, James Curlin, program manager for Telecommunication and Computing Technologies, issued a report concluding that development of a HDTV industry could reverse the erosion of leadership in many electronic technologies.

Research Costs: Various proposals have been introduced to lower the cost of research by a reduction in the capital gains tax, by encouraging tax free savings plans and by formally changing the mission of DARPA. Federal Labs would be allowed to commercialize their research through licensing agreements. Currently Federal labs employs 18% of the scientists in the U.S. Thus, they enjoy a large portion of the government's $65 billion annual research budget.

Tax Considerations: Many observers have asked for a reconsideration of section 861 of the IRS code. This section requires that companies allocate a certain portion of their R&D to foreign income, thus creating an incentive to move R&D operations offshore.

Executive Branch

The Bush Administration talked about "moving more toward an investment on applied R&D at the margin" and emphasizing support of "high performance computing...battery technology...[and] ...magnetic levitation."¹⁷ The American Technology Preeminence Act avoided a veto only after a $10 million loan program to help companies' research of commercial applications was deleted. Further, the two year old Advanced Technology program (ATP), which funds small and medium sized companies engaged in fundamental research, is expected to spend $45.8 million in grants in the early 1990s. A 38% increase is sought for the ATP in fiscal 1993.

The Bush Administration considered HDTV to be the seed of industrial policy and opposed support for a technology with such a direct product outcome. Administration officials often risk their careers when they take too bold a position on industrial policy. For example, Craig Fields, a former DARPA director, lost his job over HDTV. This dismissal drew a strong reaction, particularly from Senator Al Gore (D-TN), chairman of the Senate Commerce, Science and Technology Committee. In response to industrial policy allegations, the HDTV concept was renamed High Definition Systems (HDS) to emphasize the much wider applications of the technology. In addition, Congressmen Mel Levine (D-CA) and Don Ritter (R-PA) co-chairs of the Congressional HDTV Caucus changed its' name to the Congressional High Technology Caucus.

Within the executive branch, various agencies exist to monitor technological developments. In the Commerce Department, the Technology Administration is responsible for analyzing and removing barriers to technology commercialization. This department consists of the Office of Technological Assessment and the National Institutes of Standards. The President's Council of advisors on Science and Technology helps inform the Office of Science and Technology on policy, which in turn reviews the research and development budgets of the Department of Energy, Commerce, Interior and the National Science Foundation. These reviews are forwarded to the Office of Management and Budget. The National Science Board sets policy for the National Science Foundation. In the Defense Department most of the activity takes place in DARPA, although there is a civilian run agency, the Office of Research and Advanced Technology, that is responsible for the guidance, policy and oversight of the Science and Technology Program.

Regulatory Environment

The Federal Communications Commission's (FCC) has among other charters, the goal of protecting consumer investment. This surfaced over forty years ago when the FCC mandated that color TV's be backward integrated with black and white sets. The FCC currently stipulates that the HDTV standard be compatible with current TV transmission.

Zenith wants to set a fixed time after adoption of the HDTV standard to terminate NTSC. Such a date should be at least seven years after adoption of the standard. To protect consumer investments, Zenith argues that HDTV must be scalable (through converters) to accommodate resolution on the 13-inch to 20-inch sets. Zenith forecasts a 1% penetration one year after the adoption of the HDTV standard. Additional lobbying by Zenith includes a proposed foreign excise tax. The company recommends that those sets with a domestic content of 50% or more, enjoy reduced licensing fees.

The competition for a transmission standard will be judged by the Advanced TV Advisory Committee composed of colliding interests including: chief executives from TV manufacturers, broadcasters, cable TV companies and suppliers. The Committee will finish testing and recommend a system to the FCC in early 1993. Digital system proposals are now being tested at the Advanced Television Test Center (ATTC). These tests and those of analog proposals will be presented to the FCC. A standard should be chosen by mid-1993.

Industry Players

Sony lost about $160 million in 1991. Thus, it cut its capital expenditures by $500 million in 1992. Together with Fujitsu, Hitachi and TI, Sony has been co-developing a chips set for the MUSE decoder. The chips set, which reduced the number of devices in the decoder from 150 to 30, and the cost to less than $1500, has been available in commercial volumes since April 1991.

Sony is the leading developer of professional HDTV equipment. In July, 1992, Sony introduced a 32-inch HDTV in Japan incorporating a full-specification MUSE decoder. In April 1993, Sony also introduced in Japan a low-priced, basic feature 32-inch model. Both of these units are capable of receiving HD broadcasts and are equipped for linkage to HD laserdisc players and other HD equipment that will be sold in the future. In 1994, Sony introduced in Japan a 28-inch model equipped with a newly developed full-specification MUSE decoder, at a price lower than that of any of Sony's existing models.

In digital video editing systems, Sony's D-2 standard has been adopted by 5000 broadcasters worldwide. Sony also works on HDTV optical videodisk systems with a two hour playing time. In 1992, Apple and Sony set up General Magic, a venture which will develop multimedia players. Additionally, Hitachi, Sony and Matsushita co-developed an analog, uncompressed base-band VCR format for recording HDTV signals and for playing back prerecorded tapes. They plan to make this format compatible with other HDTV products. Since Sony and Matsushita had been rivals in the VCR format, this announcement surprised industry analysts.¹⁸ Sony also developed HDTV equipment for use in car production at Ford's plants in England.

Hitachi introduced a HDTV with a 46-inch screen, the largest for HDTV sets targeting consumers. In 1991, the company had an output of 1000 AMLCDs per month, only 10% of the quantity promised to OEMs.

Toshiba and Applied Materials will jointly develop LCD technology based on chemical vapor deposition. Also, Toshiba produces the TFS-800 system which can store and retrieve 240 HDTV pictures. In 1991, Toshiba introduced a $26,000 HDTV system and a professional HDTV camera with 2 million pixels. Furthermore, the company committed itself to reducing the price of an 11-inch AMLCD to $385 by the end of 1995. Display technologies, a 50-50 joint venture between IBM and Toshiba, began full scale production of TFT-LCDs in November 1991. Investment by both companies in the joint venture totaled $1.2 billion. As an alliance partner with Motorola, Toshiba is involved in the development of chips for the MUSE decoder.

In the field of industrial HDTV equipment, Toshiba developed the world's first 3/4 inch digital, high-definition VCR, which is based on a universal digital high-definition recording format jointly developed with Broadcast Television Systems GmbH (BTS).

Toshiba is currently cooperating with Scientific Atlanta to develop a digital set-top converter for Time Warner Cable's experimental full-service, interactive home video network in Orland, Florida. In 1994, Toshiba also supplied HDTV systems to the Warner Brothers Studio Store in New York City.

Sharp has marketed a receiver since March 1992 for $8,000, 25% of the price of existing HDTV sets. The set has a 36-inch picture tube with 1,125 scanning lines, a screen with 16:9 aspect ratio and uses a Muse NTSC converter. Sharp also introduced a 16.5-inch color LCD which handles 1.2 million pixels, allowing the use of computer graphics and multimedia. With a 1991 production capacity of 200,000 passive monochrome LCDs per month, and 10,000 color LCDs a year, Sharp is the market leader in HRD technology. Sharp hopes to decrease the cost of a 10-inch AMLCD down to $400 by the year 1995. In February 1992, Apple and Sharp set up an alliance. Both companies will develop a notebook with video and sound capability with Sharp supplying color AMLCDs.

Matsushita, in 1991, demonstrated a 15-inch amorphous silicon, thin film transistor AMLCD which displays 16.7 million colors. Together with NHK, the company developed a VCR for professional broadcast studios. With Dai Nippon Printing and Oki, Matsushita will develop a 55-inch gas plasma panel TV screen. Further, the company entered into an agreement with Gain Technology to develop object oriented multimedia software.

Sanyo developed an HDTV system with a 400-inch screen and 4.5 million pixels based on the MUSE transmission system. Together with LSI Logic, Sanyo is developing a chip set for the Japanese HDTV system that should reduce the number of chips from the current 46 to 10. Ownership of the chip design will be held jointly by both firms. Under the agreement, the chips would be manufactured in Japan by LSI Logic's affiliate, Nihon Semiconductor.

Philips, in 1992, was awarded $41.7 million in R&D subsidies for liquid crystal displays (LCD) research under the European EUREKA and JESSI programs. The company will convert a factory to flat panel production by early 1993. Philips is seeking other European partners for its LCD project. Thompson, however, had already decided to develop and manufacture LCDs in its own Grenoble facility. Philips announced its first HDTV sets in the D2-MAC standard for 1994. Despite $2.5 billion provisions for restructuring expenses, Philips has committed more than $2 billion to R&D. Much of these research efforts will go into digital compact cassettes, CD-I and HDTV. Together with Thompson and AEG, Philips is establishing a research consortium to develop LCDs based on ferroelectric technology. This new technology is expected to have higher manufacturing yields than AMLCDs.

Apple, in response to a decrease in net income of 40% in FY 1991, plans to diversify into the consumer electronics industry. Mr. Sculley, the CEO, envisions designing the "Paradigm 3 TV scalable, interactive and personalized." In addition to accessing the hardware skills from Sony and Sharp, Apple set up a joint venture with IBM (Taligent) to develop software standards and file formats for multimedia applications. Sony and Philips support this venture. The future of the IBM-Apple relationship is uncertain. Many expect an Apple-Sony relationship to be better suited to Apple's diversification plans.

Coloray received financing for the development by Micron Technology, a company that produces memory chips. Micron, with $450 million sales in 1991, will also build the pilot plant. Coloray expects to commercialize 10 inch diagonal displays for computers by 1994.

Microelectronics and Technology Corporation, a member of the International Electronics Devices Association, established a consortium of U.S. companies involved in field-emitter CRT technology. The consortium includes Optical Imaging Systems (OIS), Cherry Display Products, Electro Plasma Inc., Magnascreen Planar Systems, Plasmaco Inc. and Tektronix. In 1991, the consortium received $15 million; $7.5 million from its own members, and $7.5 million from the Advanced Technology Program. The research program will cover automated inspection and repair technologies in flat panel manufacturing. Under a DARPA contract, MCC also develops displays based on field emission technology. The first commercial screens from MCC is expected around 1995.

Photonics Technology has developed 19-inch displays for the U.S. military. The firm's proprietary technology handles 262,000 colors and achieves flicker-free imaging with wide-angle viewing. The company was part of a group that lobbied for 4.6% antidumping duties levied by the Commerce Department on Sharp and Toshiba. Photonics has been supplying 60-inch monochrome screens to the military for $100,000 each. In 1992, only approximately fifteen screens were produced per year. Photonics estimates that the cost of production could fall to as low as $1,000 each if mass produced.

Planar Systems received a $2.5 million contract from DARPA for research and manufacturing in its thin film electroluminescent (EL)technology. Since its merger with Finnish Finlux Display Electronics, Planar has become the largest maker of EL technology in the world. The company expects to begin pilot production of 8-inch VGA panels in late 1993. Further, Planar will develop EL displays for DEC workstations. Like Photonics, the company successfully lobbied for antidumping duties.

In-Focus had 1991 revenues of $37 million. Additionally, Compaq paid In-Focus $650.000 for the rights to use its Stacked PMLCD technology. In-Focus is working on reducing, by threefold, the response time. Since In-Focus imports passive LCD glass screens from Japan, in 1991 it joined Compaq in lobbying against the 62% duties imposed on imported LCDs. The company announced it would pursue licensing options in military, instrumentation and medical computers.

Tektronix has been selected to carry out research for military applications for its PALC technology under a DARPA contract.

Footnotes

EXHIBIT 1

Technology Details

HDTV & Compression

One second of HDTV signal requires approximately one billion binary (0s and 1s) numbers of code or a gigabit. Bits are arranged in groups of eight called bytes. Thus, a gigabit equals 120 megabytes. Microprocessors perform operations on one or more bytes at a time (a word) depending upon the width of the data path measured in bits. Within the Intel chip family, the 8088 processes 8 bit words, the 80286:16 bit words, the 80386: 32 bit words and the i860:64 bit words. One or more instructions are performed on these words per clock cycles measured in thousandths of a second or megahertz (MHz). Overall chip performance is measured in millions of instructions per second (MIPS).

The digital video compression strategies that have been devised involve several components, including coding options, quadrature modulation, motion compensation and error correction. Together, these compression techniques allow an HDTV signal to be compressed by a ratio of 100:1.

Coding refers to the numeric description of pixels. At the most basic level, coding transforms continuous waves into discrete numbers (quantization). Quantization is achieved by dividing waves into nonoverlapping subranges and assigning discrete numbers to each subrange. Coding can also reduce the number of signals that must be processed and transmitted by taking advantage of pixel redundancies. Redundancy occurs both within a frame and across time.

At any one time, large blocks of pixels, such as the sky, have little variance. Thus, Ioterpolative Coding can send a subset of pixel infonnation upon which can be extrapolated during decoding to recreate a close approximation of the original pixels. Transform Coding involves transforming pixel intensity into frequency coefficients for blocks of pixels. Adaptive Transform Coding changes the transformation process as a function of either picture statistics or as a function of subjective quality requirements.

Across moderate- to low-action frames, most pixels change in a predictable manner. Differential Pulse Code Modulation (DPCM) takes advantage of evolutionary changes by using past trends to predict future changes. A decoder is able to make the exact same predictions. Thus, only the common prediction error must be transmitted.

Quadrature Modulation allows two distinct signals to be broadcast over the same frequency by insuring that the two signals are 90 degrees out of phase. Quadrature is currently used in the NTSC standard to transmit chrominance (color) information.

Motion Compensation can correct for the distortions introduced by compression tools such as interlace scanning. As discussed under HRD, scouting provides a 2:1 compression ratio, but it distorts moving objects. Motion estimation algorithms extract three-dimensional motion parameters from a sequence of two-dimensional images. These motion parameters, in turn, can be used to compensate for the distortions that motion introduces to compression.

Error Correction involves detecting faulty bits in a word of information and forming new bits. It is essential that transmission errors be corrected due to the magnifying effect of decompression and the lingering effect of differential transmission. The corrected bits are interpolated from the preceding and/or following words. Forward Error Correction (FEC) can correct for as much as 33% error in a word by storing a sequence of words upon which to make predictions.

Proposed Compression Standards

General Instrument proposed fitting a digital HDTV signal into a 6 MHz channel by first using a Transform Coding algorithm. The signal is further compressed by only sending the difference between the predicted signal and the actual signal using DPCM. To increase the accuracy of predicted signals, Motion Compensation is used to forecast movements. Lastly, the chrominance is averaged in groups of 4 pixels by 2 pixels while the luminance (brightness) is broadcast for each individual pixel. Chrominance is broadcast at a relatively lower resolution due to the eye's relatively greater sensitivity to luminance.

Zenith-AT&T (DSC-HDTV) have also developed a compression algorithm that squeezes an HDTV signal into a bandwidth of 6 MHz. Again, DPCM is used to broadcast only the discrepancy between the new image and the predicted image. The difference between the two algorithms lies largely in their prediction accuracy. The DSC-HDTV algorithm forecasts movements in the form of vectors by tracking luminance. Luminance moving in blocks across frames is translated into vectors using a hierarchical block-matching motion estimator. Before transmission, the signal is compressed a second time using Transformation Coding. Since luminance and chrominance are encoded separately, the signals must be synchronized, as well as error corrected by the decoder.

Philips-Sarnoff's Advanced Digital Television (ADTV) adopts the collection of compression methods outlined by the Moving Pictures Expert Group (MPEG), a committee of the International Standards Organization. The methods include motion estimation, motion-compensated predictive coding, adaptive Transform Coding, quantization and variable-length coding-decoding. A prioritization layer determines how the various compression methods interact so that the most important video data is transmitted with the greatest reliability. MPEG promises the least distortion, but it also broadcasts the lowest resolution of the four proposed digital standards (1050 lines interlaced).

MIT's American Television Alliance System (ATVA) also adapts the compression strategy to the circumstance. When there is little to moderate motion, the difference between predictions and actual images are compressed (DPCM). To increase the accuracy of predictions, motion estimation and compensation techniques are used. When there is significant motion or scene changes, DPCM compression is disabled and the image frame is encoded in bands of 8 pixels by 8 pixels. A weighted average based on luminance and chrominance determines which bands are most important and thus first compressed. The process continues until the band width is saturated with compressed data.

Screen Measurements

Refreshing may occur every other line (interlaced) or every single line (non-interlaced). Keeping the scanning frequency constant, interlacing doubles the number of lines and thus doubles the resolution. Though resolution increases, image quality decreases since the screen is displaying simultaneously two consecutive frames out of synchronization by 1/30th of a second. Motion is thus blurred and jagged edges result. The traditional standard (NTSC) calls for interlacing, while the Zenith/AT&T HDTV standard proposes non-interlacing.

The scanning frequency is measured in megahertz (MHz). NTSC is broadcast at 59.94 MHz, which translate to thirty interlaced frames per second. The Zenith/AT&T HDTV standard also proposes broadcasting at 59.94 MHz, but in a non-interlaced format that corresponds to sixty frames per second.

A line consists of a string of colored dots known as pixels. Each individual pixel is uniform in color and shade. When arranged as a tightly packed mosaic, however, these monotone pixels form pictures of all colors and shades. How seamless the pixels blend together is directly proportional to the resolution and inversely proportional to the screen size. NTSC consists of 525 inter-laced lines per frame. HDTV proposals standards that range from 700 non-interlaced lines to 1575 interlaced lines per frame.

Lastly, the NTSC aspect ratio calls for the screen width to be 4/3rds as long as the screen height. This near square proportion was adopted largely to add stability to the cathode ray tubes (CRT) first used as screens. A CRT's contains a vacuum in a large glass tube. The more cylindrical the tube, the lower the stress introduced by the vacuum. The Zenith/AT&T HDTV standard is modeled after the 35mm film or cinema format, which has a width 16/9ths as long as the height. Though HDTV screens will initially be made from CRTs, improved technology will allow adequate stability in the rectangular format.

Attaining HDTV resolution is a trivial task for today's CRT technology. The challlenge in HRD lies in improving a screen's production cost, color fidelity, image contrast, viewing angle and physical size. The 90-year old CRT technology is likely to be eclipsed in the medium run with alternate HRD technologies currently under development. By the end of the decade, several flat screen technologies are expected to out perform CRTs in every respect, except perhaps price.

Cathode Ray Tube

CRTs have been the standard display technology from the birth of television to the birth of the HDTV. Modern advances in CRT technology can compensate for most of the inherent drawbacks of small CRT displays, however, large CRT displays are likely to remain awkward due to their depth and weight.

CRTs consist of a large glass cone (tube) with an electron gun (cathode) on the narrow end and a coating of phosphor dots on the inside of the broad flat end (screen). The electron gun emits electrons (rays) which are attracted to the positively charged screen. The phospors on the screen emit colored light when they are excited by the electrons. The sides of the cone contain electron coils (yoke) which channel the flow of electrons into vertical and horizontal lines by virtue of the magnetic field which they create.

Color CRTs require three electron guns. Each excites a distinct phosphor dot in a cluster. The cluster of phosphor dots, representing the three primary colors, form a pixel (the intersection of a vertical and horizontal line). To insure that an electron ray hits only the intended primary color within a cluster, a sheet of metal in front of the screen full of minute holes (shadow mask)absorb errant rays.

Traditionally, CRT color saturation and contrast have been limited by the amount of electron energy the shadow mask could absorb. To improve saturation and contrast, a polarizing filter can be applied to the screen. This filter, however, reduces the light intensity producing a dimming effect. Increasing the intensity of the electron beam compensates for the dimming effect, but the increased errant energy expands the metal show mask causing image distortion. Zenith has solved this problem in moderate size screens by placing the shadow mask under such high tension that the heat fails to distort the holes. Zenith continues to expand this technology to ever larger screen sizes.

Another CRT weakness that Zenith has been able to overcome in smaller screens is the traditional curvature of CRT screens. Hemispherical screens restrict the angle of view and increase the problem of glare. Screens are curved along the circumference of the electron beam to insure that the cross section of the rays remain circular and in focus from the center to the edge of the screen. Zenith has overcome these difficulties in smaller screens bv adjusting the ray focal length as it scans across the screen and altering the distance between dot clusters. Other manufacturers have been able to flatten their screens using similar technology, but Zenith remains in the lead in both size and degree of flatness.

Despite these advances, CRTs remain prohibitively heavy and bulky. The larger the screen, the further back the electron gun must be placed and the thicker the glass tube must become for strength. If the gun is not recessed, the beam must be moved through larger angles exacerbating ray distortions. As the glass tube becomes larger and less cylindrical, it must be made thicker to avoid breakage. Thus, a 3 x 5 foot screen, a likely format for HDTV, would weigh approximately 800 lbs. The neck of its tube would be too long to fit though the door of a home.

CRTs are also toxic. The radiation emitted by CRTs are suspected of adversely affecting health. This is of growing concern, especially to people who work around computer displays all day. The flicker caused by the sweeping beam of electrons is also suspected of aggravating the sight of computer operators over extended periods of time. Lastly, the radiated tube and phosphor are too toxic for landfills.

To date, Zenith has been unable to apply its flat tension mask (FTM) technology to screens above 34-inches in diameter, and then only at high cost. Thus, initial HDTV screens will be limited in size and quality by CRT flaws. Zenith's CEO Jerry Pearlman may well be correct in his prediction that "Over the next 10 years, the CRT is going to be the most cost-effective display, and by far the most cost-effective high definition display."

Projection System

A current alternative to CRTs is a projection system which basically folds a projected image via a mirror. The image is first created by any of a number of technologies, including CRT-Liquid Crystal Display (LCD) or a lightvalve. The image is then bounced off a mirror and projected onto the back of the viewing screen. These systems have the advantage of a large flat screen, but the disadvantages of poor image contrast and a lack of brightness. Hughes Aircraft and Hughes Electronics have traditionally produced projection systems with the highest contrast and brightness, but at very high cost. A lower cost system is expected for HDTV.

Field Emission Display

Several companies are developing a radical variation on the CRT theme called Field Emission Display (FED). Instead of the rays from one electron gun exciting millions of pixel dots, ten to one hundred microscopic electron guns (emitter tips) illuminate each individual pixel dot. Thus, several emitter tips per pixel can be defective without compromising the quality of the screen.

The emitter tips are mounted on the intersections of row and column electrodes which drive the cluster of electron guns. The electrodes in turn are mounted on a thin base plate. Parallel to the base plate, is the face plate which is coated on the inside with dots of phosphor in cluster formation to form pixels. The entire sandwich is only a few millimeters in depth.

The advantages of FED over CRT are numerous. FED screens are perfectly flat and very thin. They also require much less energy, reducing unwanted radiation, lowering power consumption and reducing heat buildup. The fault tolerance built in through emitter tip redundancy increases yields, decreasing production costs. The disadvantage is that the largest FED produced to date is only six inches in diameter.

Microscopic cathodes were developed by Capp Spindt at SRI International in the late 1960s. LETI Labs in France advanced the technology in 1983 by producing the first video screen with the cathodes. Spindt left SRI in 1989 to develop FED screens in his own firm, Coloray.

Active Matrix Liquid Crystal Display

LCD technology was developed in the U.S. by RCA in 1963. Westinghouse Electric Corporation built the first active matrix liquid crystal displays (AMLCDs), in the early 1970s. Up until the late 1980s, several major U.S. corporations including IBM, Apple, AT&T and General Electric, had R&D projects aimed at commercializing AMLCDs. A few U.S. corporations, such as Ovonic, Kodak and Xerox, continue to research AMLCD, however, Japanese giants such as Sharp, Toshiba, Hitachi, Fujitsu and Matsushita, dominate the application of the technology. In fact, U.S. companies have found it difficult to attract financing not only to AMLCD, but also to alternate HRDs due to Japan's perceived lead.

Every cluster or pixel is controlled by four thin-film transistors (TFT). One TFT for each primary color and white. Current VGA screens (642 x 480) for color portable computers have 308,160 pixels times 4 TFTs per pixel or 1,232,640 transistors. HDTV resolution will require approximately 2 million TFTs. TFTs control the flow of light through liquid crystals by applying an electric current to the crystal which untwists the nematic material making it translucent. To color one pixel, a cluster of four TFTs vary the amount of current passing through their crystals to produce the intensity of each primary color needed to form a specific color intensity.

Many believe that the future of HDS lies mainly in AMLCDs. It is estimated that the market for AMLCDs will reach $15 billion by 2000.¹⁹ In addition, it is estimated that 50% of all computers will use AMLCD technology by 1995.²⁰ Many Japanese firms are betting on AMLCD screens dominating HDS markets. It is estimated that Japanese firms will have invested over five billion dollars in AMLCD research and development by 1993. Though Japanese firms have been unable so far to commercialize screens larger than 15 inches in diameter, these small screens have been in high demand. Given the history of the underlying semiconductor technology, it is likely that AMLCDs will rapidly increase in size and decrease in price as a function of production volume.

AMLCD technology is often characterized as a memory chip for a screen. Indeed. a AMLCD has millions of transistors that are individually addressed like a memory chip. AMLCDs do not emit light, rather they selectively allow pinpoints of light to pass through both the liquid crystals and a color filter.

The advantages of AMLCDs are numerous. Image quality is superior to CRTs because there is no need for a curved screen, there is no concern for focus and there is no problem applying a polarizing filter. AMLCDs are so thin and light that they can be hung on the wall like a picture. AMLCDs are also nontoxic. Analysts believe that within the next decade, AMLCDs will become more cost effective than CRTs in formats larger than 40 inches.

Problems with AMLCDs include a relatively high power consumption and low manufacturing yields. Power consumption is a major issue only for battery operated screens used for such applications as note book computers. Manufacturing yields, currently at only 30% for 10-inch screens. are expected to rise significantly. Sharp, which in 1990 held a worldwide market share of 40%, hopes to increase yields of 10-inch screens to 80% by 1995, thereby reducing costs by 66%, to approximately $400.

The problem with yields stems partly from the fact that only one of the over two million thin film transistors (TFTs) need be defective to render the entire screen defective. Additionally, the glass substrate upon which the TFTs are mounted is much more susceptible to damage than the silicon used as a platform for most semiconductors. Preliminary indications are that the manufacturing learning curve is not as steep as with other semiconductors, such as memory chips. In 1991, for example, AMLCD manufacturers produced only 25% of their yearly volume forecast, due largely to slower than anticipated yield improvements.

Canon may emerge to lead the Japanese pack, based upon a new technology it disclosed in October 1991. Canon claims its new feroelectric AMLCD is larger, sharper, easier to produce and thus less expensive than conventional AMLCDs. Many U.S. corporations hope to overtake the Japanese firms by developing alternative technologies.

Plasma Addressed LCD

Tektronix has modified AMLCD by eliminating TFTs in its plasma-addressed liquid crystal (PALC). PALC varies current to LCDs via ionized and deionized gas. The gas is ionized with plasna switches. Theoretically, PALC will deliver the performance required by HDTV, but in 1990, Tektronix had only been able to produce a 7-inch screen with 300 lines.

Passive Matrix Liquid Crystal Displays

Passive matrix liquid crystal displays (PMLCDs) or supertwist LCDs differ from an active matrix in that the color dots within each pixel cluster are not addressed individually, but rather through strips of electrodes along the x and y axis. This less efficient process reduces the screen response rate. It also reduces the contrast and the viewing angle. These screens have been widely used in portable computers.

In-Focus Systems has addressed these inherent flaws by designing a system whereby three PMLCDs, one for each primary color, are stacked upon one another. When current flows through the three dot clusters aligned vertically, the pixel takes on the white color of the backlight. By selectively reducing the current, primary colors can be blocked to produce a pixel with whatever color is desired.

While the stacking innovation does improve image quality, it also adds weight and bulk. More importantly, stacked PMLCDs retain the disadvantage of their slow refresh rate. Still, PMLCD will continue to thrive at least in the portable computer market until AMLCDs production yields improve.

Gas Plasma Display

Gas plasma display (GPD) is perhaps the contemporary system that comes closest to meeting HDTV needs. The basis of the technology is a gas that illuminates when charged with an electric current. To produce color, clusters of phosphor coatings of each primary color are placed on the inside of the screen. The gas illuminates the dots necessary to produce the color required of each pixel. One of the technological problems that must be overcome is how to compensate for the red-orange light emitted by the gas. It may be difficult to attain a full range of colors on large screens.

Fujitsu already produces a 16-inch GPD for Sun workstations. NHK has demonstrated a 33-inch color GPD, but it only has l/3rd of the pixels necessary for HDTV. A model compatible with HDTV is expected soon. Other companies developing GPD technology include NHK, Matsushita, Sharp, Thompson and Hitachi.

Electroluminescence

Electroluminescence (EL) is similar to the preceding technologies except for the fact that there is no intermediary between the electric diodes and the color phosphors. When current is supplied, the phosphor illuminates directly. The problem with this technology so far has been finding a phosphor that illuminates with the proper blue color.

Planer Systems currently produces an EL with a 19-inch diameter for computers. Planer's manufacturing system, vacuum deposition, allows high-quality and high-volume and hence low prices. Planer Systems was spun off from Tektronix who still holds equity in the venture. Other EL competitors include Sharp and Hitachi.

Deformable Mirrors

Texas Instruments (TI) has developed a radically new approach to illuminating phosphor dots on the screen. A chip with hundreds of thousands of independent mirrors, called a deformable mirror device (DMD), directs simple light rays to strike individual phosphor dots, as needed, to build whatever pixel colors are required. By varying the electrical charge around a mirror, the orientation of the mirror can be changed to one of three positions. In 1990, TI had only been able to build a display with 16,000 mirrors, well short of the million mirrors that are required to fit HDTV parameters. Other problems include manufacturing complexity and image contrast. The first product likely to use DMD technology will be a laser printer employing 840 mirrors in a linear array.

Omnivision 3D

TI is already developing the technology that might render all flat panel displays obsolete: 3 dimensional (3D) screens. TI first demonstrated a prototype in 1990. The screen can be viewed at great angles and doesn't require any additional peripheral equipment such as glasses. The prototype is capable of 750 scanning lines and could be upgraded to HDTV standards.

The viewing surface is a translucent double helical disc that rotates at 600 rpm, presenting an area continually varying in depth, like threads of a screw. Three laser beams scan the disc as it turns, thereby projecting onto it the color TV picture. The first applications are likely to be for the military and air traffic control.

Zenith's Important Events

First Quarter 1992

• Industry color TV sales down 9% coupled with the production start-up of new big screen flat tension mask (FTM) screen color display tubes that will contribute to an expected first half loss for 1992.
• In February, began shipping new 15 inch and 17 inch FTM color computer monitors with multi-scanning capabilities.
• In March, reduced employment by 75% at Springfield , MO plant to consolidate color TV final assembly and plastic cabinet operations with plant in Mexico.
• In March, received $5 million defense contract to develop FTM screens.
• In March, the FCC began testing of DSC-HDTV.

• Lost $51.6 million on $1.32 billion in sales.
• Industry color TV sales sown 4%.
• Gain in market share for color TVs due to revamped line of products.
• Cut operating expenses by $58 million for year.
• Sold Taiwan monochrome montior plant to consolidate operations with Mexican plant.
• Sold magnetics plant in Keils, Ireland to consolidate operations with the magnetics plant in Chihuahua,Mexico.
• Announced plan to phase out automotive electronics by mid-1992.
• Announced plan to phase out low-end color and monochrome monitors due to extreme market competition from Far East suppliers.
• Re-entered the audio business to capitalize on demand for home theater systems.
• Zenitt/ATT develop digital filtering technology to eliminate interference from standard TV signals.
• In March, Nycor (who owns 8% of Zenith's stock) and its owner Sal Giordano, Jr., send proxies to shareholders in an attempt to win control of the board of directors.
• In April, issued $15 million in common stock to Korea's GoldStar as a white knight in proxy fight.
• In other dealings with GoldStar during April, the Korean TV maker signed licensing agreements to market FTM technology in TVs in Korea and agreed to promote the DSC-HDTV system as the Korean standard.
• At April's stockholders meeting, current management wins 2 to 1 support, successfully fighting off the Giordano proxy fight.
• In September, introduced first TVs in the world with built-in closed caption decoder. (Taking this action two years ahead of U.S. regulations requiring all manufacturers to have models with this capability by June 1993.)
• In December, Scientific Atlanta endorsed DSC-HDTV for the U.S. standard and is developing equipment for satellite delivery and testing.

• Lost $65.2 million on $1.41 billion in sales.
• Cut $38 milion in operating expenses for fiscal year.
• U.S. defense advanced research project agency provided $1.25 million for research on high-definition displays.
• State of Illinois gave $1 million grant for the development of new productiion processes.
• Power supply plant in ireland phased out.
• Components operations downsized worldwide.
• Development breakthrough allowing FTM shadow masking to be stretched to larger size screens with greater precision at costs lower than standard tube production.
• In April, received $15 million last payment from Groupe Bull for comuter products business.
• In October, joint Zenith/GoldStar task forces began working to expand the vender customer relationships between the two.
• In December, Zenith and ATT announced the development of an all digital version of Zenith's Spectrum Compatible HDTV called Digital Sectrum Compatible HDTV (DSC-HDTV).

• Lost $66.4 million on $1.55 billion in sales.
• TV unit shipments increase 9%.
• Applied for ten U.S. patents on HDTV bradcast related inventions, and more than thirty related to FTM display technologies.
• In February, entered into a partnership with AT&T to develop an all digital HDTV system based on Zenith's Spectrum Compatible HDTV system
• In December, sold computer products division to Groupe Bull.

• Return to profitability making $11.7 million on $2.69 billion in sales.
• Twenty-three percent increase in computer sales; consumer electronics sales were flat.
• Added 600 employees at the Chicago-area picture tube plant which operated seven days a week.
• In September, announced proposal for HDTV system called Spectrum Compatible HDTV.