Embed Size (px)
Citation preview
The Big Picture: HDTV andHigh-Resolution Systems
June 1990
OTA-BP-CIT-64NTIS order #PB90-256108
Recommended Citation:
U.S. Congress, Office of Technology Assessment, The Big Picture: HDTV and High-Resolution Systems, OTA-BP-CI’I’-64 (Washington, DC: U.S. Government Printing Office,June 1990).
For sale by the Superintendent of DocumentsU.S. Government Printing Office, Washington, DC 20402-9325
(order form can be found in the back of this report)
Foreword
As early as 1883 inventors dreamed of transmitting visual images to distant pointsextending what they had already done for written messages and voice with signals carried overthe telegraph and telephone. By the 1920s, signiilcant efforts were underway to scan andproject images. Crude television images-actually little more than shadows—were demon-strated in 1926, and by 1928 even color images were achieved. Television broadcasts of thefirst television drama, The Queen’s Messenger, were made from an experimental station inSchenectady, New York in 1928. TV was still futuristic at the time of the New York World’sfair in 1939 but finally erupted into widespread commercial use in the 1950s. Now nearly ahti-century later, technological change drives the video revolution unabated.
Television has become the dominant entert ainment medium, and has displacednewspapers and radio as the prime opinion midcer in American current affairs. Its influencereaches millions of individuals daily. It has been praised as an extraordinary educational tool,and its communication power has been credited with deeply influencing the actions ofgovernments, including the fast-moving democratization of Eastern Europe. On the otherhand, television is criticized as an agent of political manipulation that must share substantialblame for increasing voter apathy, and suffers with an image of daily progr amming dedicatedto sitcoms, violence, and trivial game shows. Television has become big business, withproduction, broadcasting, advertising, video tapes, and the cable industries ranking high in theU.S. economy. Besides entertainment, this technology now profoundly affects communica-tions, national security, research, and education.
Television technology is now on the threshold of a new evolution. We are on the vergeof combining digital-based computer technology with television. This technological marriagepromises to produce offspring that can deliver movie-quality, wide-screen programs to ourhomes with stereo sound equivalent to the best compact disks. Its importance goes wellbeyond home entertainent, however. High-definition television—HDTV as it is called-islinked with many other basic technologies important to the United States. The impacts of thedevelopment of HDTV will ripple through the U.S. economy: It will make us confront suchissues as public policy dealing with manufacturing, educational and trainin g standardization,communications, civil and military command and control, structural economic problems, andrelationships between government and business.
This background paper was requested by Congressman George Brown, as a Member ofthe Technology Assessment Board. It is a primer of HDTV technology and its relationship tohigh-resolution computer systems. OTA gratefully acknowledges the contributions of themany experts, within and outside the government, who reviewed or contributed to thisdocument. As with all OTA publications, however, the content is the responsibility of OTAand does not necessarily constitute the consensus or endorsement of reviewers or theTechnology Assessment Board.
~f~a’L# ‘ >
JOHN H. GIBBONSW Director
. . .Ill
Workshop on High Definition TV and High-Resolution Systems, October 1989
Lynn ClaudyStaff EngineerNational Association of Broadcasters
Robert CordellDistrict ManagerVLSI Systems Research
Birney DaytonPresidentNVision
D. Joseph DonahueSenior Vice President, Technology and Business
DevelopmentThomson Consumer Electronics
Bruce L. EganSpecial Consultant and Affiliated Research FellowColumbia University
Kemeth FlammSenior FellowBrookings Institution
hn GeldingVice President, Systems EngineeringHughes Network Systems
Dale N. Had3eldPresidentHatiield Associates, Inc.
John G.N. HendersonHead, Systems TechnologyDavid Sarnoff Research Center
Wayne LuplowExecutive Systems, Electronic Systems R&DZenith Electronics Corp.
David G. MesserschmittProfessor, Department of EE/CSUniversity of California, Berkeley
Howard MillerSenior Vice PresidentBroadcast Operations & EngineeringPublic Broadcasting Service
Michael RauVice President, Science and TechnologyNational Association of Broadcasters
G-ifiith L. ResorPresidentMRS Technology, Inc.
Robert L. SandersonTechnical Assistant to the DirectorImaging Information Systems-ResearchEastman Kodak Co.
William F. SchreiberProfessor, Am ProgramMassachusetts Institute of Technology
John SieSenior Vice PresidentTelecommunications, Inc.
Laurence J. ThorpeVice President, Production TechnologySony Advanced Systems
Arpad’G. TothChief ScientistAdvanced Television TechnologyPlanning and Policy DevelopmentNorth American Philips
Don WalkerDirector of Technical ProgramsMotorola, Inc.
John WeaverPresidentLiberty Television
Philip WebrePrincipal AnalystNatural Resources and Commerce DivisionCongressional Budget Office
Barry WhalenSenior Vice PresidentMicroelectronics & Computer Technology
Corp.
NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the workshop participantmembers. The participants do not, however, necessarily approve, disapprove, or endorse this report. OTA assumes fullresponsibility for the report and the accuracy of its contents.
OTA Project Staff: The Big Picture:HDTV and High-Resolution Systems
John Andelin, Assistant Director, OTA Lionel S. Johns, Assistant Director, OTAScience, Information, and Natural Resources Energy, Materials, and International Security
Division Division
James W. Cu.rlin, Program Manager Audrey Buyrn, Program ManagerCommunication and Information Technologies Industry, Technology, and Employment
Samuel F. Baldwin, Senior AnalystEnergy and Materials Program
Mark S. Nadel, AnalystCommunication and Information Technologies Program
Myra Gray, Executive TraineeOPM Women’s Executive Leadership Program
Kenneth R. Donow, Editor
Administrative Staff
Liz Emanuel
Jo Anne Price
Karolyn St. Clair
OTA Contributors
Robin Gaster, Senior AnalystIndus~, Technology and Employment Program
Katherine Gillrnan, Senior AssociateIndustry, Technology and Employment Program
Julie Gorte, Senior AnalystIndustry, Technology and Employment Program
Charles BostianClayton Ayre ProfessorElectrical Engineering DepartmentVirginia Polytechnic Institute and State University
Robert B. CohenSenior Economist, Industry Cooperation CouncilDeputy Director & Senior EconomistNew York State Financial Services Advisory
CommissionEconomic Advisor to the Director of Economic
DevelopmentUrban Development Corp.
Kenneth DonowNational Center for Telecommunication and
Information PolicyPublic Service Satellite Consortium
Mark F. EatonDirectorAssociated ProgramsMicroelectronics & Computer Technology
Corp.
Frederick L. IkensonAttorneyFrederick L. Ikenson, P.C.
Clark E. Johnson, Jr.Independent ConsultantFormer Congressional Fellow
Alan McAdamsAssociate Professor of Managerial EconomicsCornell University
Roy L. Beasleyconsultant
Jules A. BellisioDivision ManagerVideo Systems Technology ResearchBellcore
Charles W. BostianClayton Ayre ProfessorElectrical Engineering DepartmentVirginia Polytechnic Institute and State University
Joseph A. CastellanoPresidentStanford Resources, Inc.
Robert B. CohenSenior Economist, Industry Cooperation CouncilDeputy Director & Senior EconomistNew York State Financial Services Advisory
CommissionEconomic Advisor to the Director of Economic
DevelopmentUrban Development Corp.
Reviewers
Lee McKnightResearch AssociateResearch program Communications PolicyMassachusetts Institute of Technology
Suzanne NeilResearch AssociateMedia LaboratoryMassachusetts Institute of Technology
Peter D. NunanSenior MTSKufa-g Technology DevelopmentSematech
James ParkerJapanese Industry AnalystTechSearch International, Inc.
Ron PowellScientific AssistantCenter for Electronics and Electrical EngineeringNational Institute of Standards and Technology
Matt RohdeU.S. Customs Service
Marko Shmi.rczukprogram ManagerHigh Definition Display TechnologyDARPA
E. Jan VardarnanPresidentTechSearch Intermtional, Inc.
Lawrence R. WaldersGraham & James
Contributors
Rhonda CraneAdvisor for Science and TechnologyUnited States Trade Representative
Kenneth R. DonowNational Center for Telecommunication and
Information PolicyPublic Service Satellite Consortium
Mark F. EatonDirectorAssociated ProgramsMicroelectronics & Computer Technology
Corp.
Charles H. FergusonResearch AssociateCenter for Technology, Policy, and Industrial
DevelopmentMassachusetts Institute of Technology
William FinanDirectorQuick, Finan & Associates
vi
Larry FrenchCorporate Vice President for TechnologyNorth American Philips
Jeffrey FreyProfessorDepartment of Electrical EngineeringUniversity of Maryland
Jack FuhrerDirectorTelevision Research LaboratoryDavid Sarnoff Research Center
William GlennProfessorDepartment of Electrical EngineeringFlorida Atlantic University
David B. HackAnalyst in Information in Science and TechnologyCongressional Research Service
Heidi HoffmanInternational Trade SpecialistU.S. Department of Commerce
C. Edward Holland, Jr.Program ManagerOUSD(A) ResidentSemiconductor Research Corp.
William C. Holton, Ph.D.Director, Microstructure SciencesSemiconductor Research Corp.
James HurdPresident and Chief Executive OfficerPlanar Systems, Inc.
Clark E. Johnson, Jr.Independent ConsultantFormer Congressional Fellow
Al KelschStrategic Business DevelopmentLinear Product MarketingNational Semiconductor
David M. LewisTechnical Assistant to the DirectorElectronics Research LabsEastman Kodak Co.
Mike LiebholdApple Computer
Andrew B. LippmanAssociate DirectorThe Media LaboratoryMassachusetts Institute of Technology
Alan McAdarnsAssociate Professor of Managerial EconomicsCornell University
ke McKnightResearch AssociateResearch Program Communications PolicyMassachusetts Institute of Technology
Dave MentleyDirectorDisplay Industry ResearchStanford Resources, Inc.
Suzanne NeilResearch AssociateMedia LaboratoryMassachusetts Institute of Technology
W. Russell NeurnanDirectorAudience ResearchThe Media LaboratoryMassachusetts Institute of Technology
Peter D. NunanSenior MTS/Manufiict~g Technology DevelopmentSematech
Yozo OnoSenior EngineerNHK-North America
James G. ParkerJapanese Industry AnalystTechSearch International, Inc.
Matthew RohdeU.S. Customs Service
Eric J. SchirnrnelVice PresidentTelecommunications Industry Association
Richard SolomonResearch AffiliateResearch Lab of ElectronicsMassachusetts Institute of Technology
David StaelinProfessor of Electrical EngineeringMassachusetts Institute of Technology
Lawrence E. Tannas, Jr.PresidentTannas Electronics
David L. TennenhouseAssistant Professor of Computer Science and EngineeringMassachusetts Institute of Technology
Antoon G. UyttendaeleDirectorAllocations & R.R. SystemsBroadcast EngineeringABC Broadcast Operations & Engineering
E. Jan VardarnanPresidentTechSearch International, Inc.
Lawrence F. WeberSenior Vice PresidentPlasmaco
Robin S. WhkkinExecutive DirectorBIS Mackintosh
vii
PageChapter 1. Introduction and Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
suMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .- +...”””””””””””” 1Manufacturing 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....”” ““”””Communications Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . .. .. .. .. .. .. .. ... ... ...”OO” .OO” ””O” 2Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ””””””””.”””.””””””””””. 4
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4The Historical Development ofHDTV(Ch. 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7CommunicationsT echnologies (Ch. 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Television Technology (Ch. 4).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Technological Linkages (Ch. 5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9AdvancedTelevision Markets (Ch. 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9U.S. Manufacturing ofHDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
POLICY ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13NATIONAL SECURITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16THECOMMUNICATIONSINFRASTRUCruRE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Recorded Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Terrestrial Broadcasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Cable Television . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Direct Broadcast Satellites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Mobile Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Telephone Companies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Market Share . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20costs . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
THE STANDARDS-SETI’ING PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Standards Setting and Marketplace Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23The Thnetable for Establishing Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Providing Full Opportunities for All Serious Proponents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Hybrid Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Chapter 2. The Historical Development ofHDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27JAPAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Research and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Financial Supports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Market Promotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
EUROPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Market Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Eureka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
UNITED STA~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Production Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Transmissions tandards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36New Initiatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Chapter3. Communications Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39TEHTRIAL RADIO COMMUNICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40SA~LLITECOMMUNI CATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44COAXIAL CABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45COPPER TELEPHONELINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46OPTICAL FIBER CABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Chapter 4.TVand HRS Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49CONVENTIONALTELEVISION: PRODU~ION,TRANSMISSION, AND RECEPTION . . . . . . . 49ADVANCEDTELEWSION SYSTEMS: IDTV,EDTV, HDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52VISUAL RESPONSE ANDHIGH DEFINITIONTV .......,.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52HDTV SYSTEMS: PRODUCTION, TRANSMISSION, RECEPTION . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Production Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Transmissions ystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
PageReceivers for HDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
HIGH-RESOLUTION SYSTEMS AND TECHNOLOG~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Chapter 5. Linkages Between HDTV, HR& and Other Industries ● *.*..... . . . . . . . . . . .m...**..* ● 61INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61SEMICONDUCTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Digital Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......... .......... ...*..”.”. “...”” 63DRAMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Gallium Arsenide andOther Compound Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Semiconductor M.anufa*g ..1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 67
DISPLAY TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68STORAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72COMMUNICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73PACKAGING/INTERCONNE~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Chapter 6. Advanced TV Markets and Market Uncer@inties . 0 . . 0 . . 0 ● . 0 0 . . 0 . 0 . . . . 0 . 0 . 0 ...0...0 79INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79SCENARIOS FOR MARKET DEV.ELOPMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............... 79
Consumer Demand forHDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Consumer Demand for Interactive Video . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Government Policies and Market Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
MARKETPROJE~ONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83MARKET PROJECTIONS ANDTHERELATIVEmRT~CE OF ~m . . . . . . . . . . . . . . . . . . 85
Extrapolating Current Growth Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Comparing the Semiconductor orOther ComponentRequirements for
Assumed Market Sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86Comparing Ultimate Market PenetrationRates and Comespon~g Needs peruser . . . . . . . . . . . . . . 87
TECHNOLOGYAND MARKET TRENDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Appendix A. The Decline of the U.S. TV Industry: Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Appendix B. TheDecline of the U.S. TV I.ndustry: Trade... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93AppendixC.TheDeclineoftheU.S.Dw~dus~:Manufac~g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95AppendixD.TheDecli.neoftheU.S. DIWMIndustry: Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Appendix E. TheDevelopment ofthe Japanese Computer ~dus@y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101Appendix F. Research and Development Consortia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104Appendix G. Acronyms and Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
BoxesBox Pagel-1. Digital Video Information and Telecommunications Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53-1. Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414-1. Inside Today’s NTSC Analog Television . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504-2. Inside Analog/DigitalTelevision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534-3. hvelsof Receiver Compatibil.ity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... . . . . . 57
FiguresFigure Pagel-1. Erosion ofU.S. hdership in Semiconductor and Relat~ Technologies . . . . . . . . . . . . . . . . . . . . . . . 33-1. Amplitude andFrequency Modulation ofaCarrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413-2. ADigital Signal With Its Comesponding Numerical Equivalent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413-3. Techniques forModulating an Analog Carrier in OrderTo Send Data in aDigitid Format . . . . . . . . 423-4. Selected Allocat.ions of the Radio FrequencySpectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433-5. Map ofa CellularRad.io System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........................ ~4-1. AnalogtoDigital Conversion Using Pulse Code Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-2. Comparison oftheFieldof View for anHDTVandT*y’s NTSCTVTo Havethe Same
Resolution forthe Viewer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....”... 555-1. Selected Applications of Digital Signal processors v. the Speed of@emtion . . . . . . . . . . . . . . . . . . . 645-2. ProjectedGrowth ofthe Digiti Signal ProcessorMarkeL 1989-92 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655-3. ’’ResosRtie”e” for Active Mati Liquid c~stiDisplays............ . . . . . . . . . . . . . . . . . . . . . . . 69
ix
Page5-4. Data Transmission Rates and Holding Times for Different Types of Data Communications . . . . . . . 745-5. Levels of Packaging/Interconnect ... ................”.....””””””””””” . . . . . . . . . . . . . . . . . . . . . . 755-6. Fraction of Integrated Circuits Used in Surface-Mount Packages for Japan, Europe, and
the United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776-1. Sales of VCRs, VCR Prices, and the Growth of VCR Tape Ilmtal Stms . . . . . . . . . . . . . . . . . . . . . . 816-2. Penetration of Color TVs Into American Households, Average TV Prices, and the Share of
Prime-Time Progra mming Devotedto ColorTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ... .. ”.. .....s 826-3. Sales and Projections ofVideoEntertainm ent Technologies in the United States andJapan . . . . . . . 84
TablksTable Page2-1. Selected HDTV Technology Development Effofis by Japanese Companies . . . . . . . . . . . . . . . . . . . . . 292-2. Chronology ofTVDevelopment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....................... ~~2-3. Private U.S. Groups Involved With HDTV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-1. TheU.S. Communications ~fiastmcture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395-1. Status ofU.S. Producers ofFlat Panel Displays in the 1980s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715-2. Recent Investments in AM/LCD Production Facilities inJapan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Chapter 1
Introduction and Summary
SUMMARYDuring 1989, High Definition Television
(HDTV) moved from obscurity to center stagein the ongoing debate over the role of the FederalGovernment in U.S. industrial competitiveness.HDTV and related High-Resolution System(HRS) technologies in the computer and com-munications sectors may significantly impactU.S. electronics manufacturing, accelerate fun-damental restructuring of the U.S. commu-nications infrastructure, and provide a host ofvaluable services.
Manufacturing
HDTVS must process huge quantities ofinformation at speeds approaching those oftoday’s supercomputers in order to display areal-time, full-color, high-definition video sig-nal. HDTVS are able to do this at relatively lowcost through the use of circuitry dedicated tospecialized tasks. In contrast, supercomputersare software programmable and do a muchbroader range of information handling andcomputation.
HDTV is driving the state-of-the-art in anumber of technologies that will be important tofuture generations of computer and communica-tions equipment. These include certain aspectsof digital signal processing for real-time videosignals; high-performance displays; fast, high-density magnetic and optical data storage;technologies for packaging and interconnectingthese electronics; and, as with all high-volumeconsumer electronics, processes for manufac-turing these sophisticated products at affordablecosts.
Consumer electronics has long been theprincipal driver of important aspects of theseand other technologies. For example, televisionhas long pushed display technology. VCRs,compact disks, and digital audio tape havedriven important data storage technologies.
Products such as calculators, watches, and LCDTVs have been important in the development ofpackaging/intercomect technologies such astape automated bonding and surface mount.Light-emitting diodes (LEDs) in watches, calcu-lators, and indicators; and diode lasers inCD-players are examples of important optoelec-tronic technologies driven by the consumermarket. Finally, important manufacturing tech-nologies, such as automatic insertion equipmentto place components on printed circuit boards,were developed for the high-volume consumerelectronics industry.
Video entertainment markets will be worthbillions of dollars, whatever form they take inthe future. The economies of scale realized inproducing for markets this large combined withthe technological linkages noted before may aidmanufacturers in penetrating other marketsusing similar products and technologies, partic-ularly the computer and communications sec-tors.
Consumer electronics is characterized byfierce competition, large volume production,and low profit margins. Because of this, con-sumer electronics may be the equivalent of the“coal miner’s canary” for manufacturers ofelectronics—providing a sensitive indicator oftheir managerial and technical performance indesign, production, and quality; of the health ofthe environment they operate in (macroeco-nomic, regulatory, and structural); and of theeffectiveness of government policy towardsforeign trade practices. Consumer electronicsmanufacturing is nearly dead in the UnitedStates. Much of what remains is domestic‘‘screwdriver assembly’ of components andsubassemblies produced abroad.
Congress might question the wisdom of anyfurther government involvement in HDTV ifthey view the technology in a narrow sense—asnothing more than a near-term improvementover today’s TV. However, if it is viewed
–l–
2 ● The Big Pictwe: HL3TVand High-Resolution Systems
broadly—as a possible first step back intoconsumer electronics manufacturing; as a pri-ncipal driver of HRS technologies for futurecomputer and communications equipment; or asa component of a national fiber informationnetwork with HDTV or related products servingas the home terminal-then Congress may findthat HDTV and related HRS technologies couldcontribute to several national goals.
HDTV may also bean instructive case studyof the difficulties facing the United States inreversing the erosion of U.S. leadership in manyelectronic technologies and in global and do-mestic electronics markets (figure l-l). TheUnited States is seriously lagging technicallyand/or in market share in semiconductor materi-als; ceramic packaging; DRAMs, gate arrays,CMOS and ECL devices generally; LCD dis-plays; optoelectronics; and floppy disk andhelical scanning drives (VCRS); to name only afew. The United States will not long remain aworld leader in electronics technologies if itstechnological foundation continues to crumblein this manner.
HDTV and related HRS will not by them-selves determine the fate of the entire U.S.electronics industry. They will only have adirect impact on technologies and products forhandling visual information, and these impactswill not begin to be felt for several years. In thenear- to mid-term, the U.S. electronics industryfaces substantial challenges. Many technologiesmust be developed—including materials, x-raylithography, large-area lithography, optoelec-tronics, packaging/interconnect, and others—and much greater effort must be devoted tomanufacturing with quality at low cost.
The responses to these challenges should notbe viewed as independent efforts. The electron-ics industry is a complex and highly interde-pendent whole. For example, the design ofleading-edge microprocessors requires access tohigh performance computers which, in turn,depend on leading-edge materials, semicon-
ductor manufacturing equipment, packaging/intercomect, and other leading-edge technolo-gies. Manufacturing these microprocessors willbecome increasingly difficult if vertically inte-grated foreign competitors control the basicmaterials and semiconductor manufacturing equip-ment. Selling them will be similarly difficult ifthese foreign competitors control most othercomponents (memory and other chips, opticaland magnetic storage, displays); have superiorpackaging/intercomect and assembly technolo-gies (e.g., chip on glass); and prefer to use theirown microprocessors instead of purchasingthem from U.S. manufacturers. Although thefragmented entrepreneurial U.S. electronics in-dustry is remarkably imovative, giant foreigncorporations can easily invest more heavily incritical technologies or simply buy out the U.S.entrepreneur.
HDTV and HRS should thus be viewed asonly part of a larger and more comprehensiveeffort to understand and resolve the problemsfacing the U.S. electronics industry. Theseinclude: the high cost of capital; lack of verticaland/or horizontal integration; inattention tomanufacturing process and quality, poor designfor manufacturability, and the separation ofR&D from manufacturing; poor or adversarialrelationships with suppliers; weakness in theU.S. educational system; industrial and tradepolicies in other nations that have aided compe-titors; inadequate foreign protection of U.S.intellectual property; and foreign trade viola-tions and closed foreign markets. These issuesare discussed in greater detail in a recent OTAreport.l
Communications Infrastructure
The communications industry is currentlyundergoing dramatic, technology-driven changethrough the increasing use of digital electronicsand fiber optics. In the midterm, the continuedincorporation of advanced electronics and pho-tonics into the existing public telephone net-
Iu.S. con~ss, office of Technology fkessmen~ Making ?’/zings Beffer: Competing in A4aw.facfun”ng, OTA-~43 ~~tigto~ ~: Us.Government Printing OffIce, February 1990).
Chapter 14ntroduction and Summary . 3
Figure l-l—Erosion of U.S. Leadership in Semiconductor and Related Technologies
Devices
AutorwtsdSysti3m
Assemblystatisticalprocesscontrol
nics
EnablTechnol
Ciaawoorn t!ompMer-
twhrmfogy integratedmanufacturing
Materials $iiiwnmatwfais
GaAs andeptoeiectronic
matwiais
Uitra-pureprocessingcimmicais
Twhnoiogies in which the United States is lagging Japan are shown in light tones; those in which the United States and Japan are roughiyequal are shown in medium tones; those in whioh the United States is leading Japan are shown in dark tones.SOURCES: Adapted from the Report of the Federal Interagency Staff Working Group, ‘The Semiconductor Industry,” National Academy of Sciences, 1987;
National Research Council, Advanced Processing of Hectionic Materiak in the United States and Japan (Washington, DC: National AcademyPress, 1988); W.C. Holton, J. Dussault, D.A. Hodges, C.L. Liu, J.D. Plummer, D.E. Thomas, and B.F. Wu, “Computer Integrated Manufacturing(CIM) and Computer Assisted Design (CAD) for the Semiconductor Industry in Japan,” JTECH Panel Report, Department of Commerce, ScienceApplications International Corp., December 1988; Manufacturing Studies Board, The Future of E/ectronhs Assernb/y: Report of the Pane/ onStrategk E/ec&onics Manufactwing Ttino/ogies (Washington, DC: National Academy Press, 1988).
work will make many new and improved making the transition to terrestrial HDTV broad-interactive information services available. Radio casting, but digital technologies offer the pros-frequency communications are also rapidly pect of more efficient use of the limitedchanging due to such imovations as cellular available spectrum than is possible with today’stelephones and other products. TV system—which is based on 40-year-old
analog electronics technologies. Some of theHDTV could accelerate these changes in the broadcast spectrum might thus eventually be
communications infrastructure. The greater infer- freed for other uses. For example, it has beenmation content of HDTVS broadcast signal has suggested that if sufficient spectrum becameraised the most significant issues for radio available, cellular telephones could becomefkequency spectrum allocation in decades. HDTV more broadly competitive with today’s phonebandwidth requirements present problems for system for voice communications.
4 ● The Big Picture: HDTIJ and High-Resolution Systems— . . — .
HDTV might also speed the extension of fiberto the home by stimulating market demand forhigh-quality video entertainment. In the mid-term, a mixed cable TV and telephone companynetwork might provide a modest level of interac-tive video services. Whether or not such asystem can evolve into a national two-waybroadband fiber network is unclear. This is animportant consideration in establishing a finan-cial and regulatory framework for the communi-cations industry.2
Services
HDTV and related HRS technologies mayoffer a host of important services to individualsas well as to business and industry. Interactivevideo, medical imaging, desktop publishing,and computer graphics are examples of servicesalready in advanced stages of development or onthe market for high-end commercial users (boxl-l). Other services may not be developed formany years, and many more remain to beinvented. Video information technologies willbecome increasingly important in computer andcommunications systems in order to make bestuse of the most important human sense—vision.
The United States has the opportunity toestablish a more powerful and flexible HDTVsystem than those currently being developed andintroduced in Japan and Europe. America hasmany strengths in HDTV and HRS-relatedcomponent technologies, as well as in highlyinnovative HDTV transmission standards andreceiver designs now under development in anumber of U.S.-based facilities.
The window of opportunity for U.S. firms toenter or to strengthen their position in thesemarkets could close quickly, however, if thestrategies of foreign competitors are successful.Many U.S. firms seriously lag in manufacturingpractices—both managerial and technological—and there is little consumer electronics manufac-turing remaining in the United States on whichto build. These deficiencies could be overcome-—
the Japanese surmounted far greater obstacles indeveloping their domestic computer industry(app. E)—but doing so would require consid-erable effort and discipline on the part of bothU.S. industry and government.
INTRODUCTION
HDTV is one of several possible forms for thenext generation of home video entertainment;following Black and White (B&W) TV in the1940s, color TV in the 1950s, and VCRs in the1970s. It promises to deliver pictures to thehome as clear as those seen in movie theaters,with sound comparable to compact disk players.
But HDTV is far more than just a prettypicture. It is part of an ongoing evolution inhome electronics toward computer-like digitaltechnologies. This evolution began with suchthings as automatic electronic tuners on stereosand TVs, compact disk players, and electroniccontrols on microwave ovens and many otherhousehold appliances. It continues today withthe introduction of Improved Definition TV(IDTV) that uses computer memories and otherdigital techniques to provide a much betterpicture even with today’s broadcasts. The evolu-tion will continue in the future with HDTV orother adv:~nced video entertainment products.
HDTV 1s also at the leading edge of a muchbroader, though less well publicized, transitionin computers and communications equipment totechnologies that can create, manipulate, trans-mit, and display high-quality visual informa-tion, including full-motion video as seen ontelevision receivers. In many respects, Ad-vanced TV (ATV), interactive video, comput-ers, and communications are all gradually merg-ing technologically. Known generically as HighResolution Systems (F-I’M), these systems willallow the user to interact with that beingdisplayed, and will have profound implicationsfor educat]on, entertainment, and work in thefuture.
2See, U.S. Congress, Office of Technology Assessment, Critical Connections: Comnunicatl(m jor the Future, OTA-CIT-407 (TVashingto~ DC: U.S.Government Printing Office, January 1990).
Chapter 14ntroduction and Summary ● 5
Box l-l—Digital Video Information and Telecommunications Services
Digital video information and telecommunications technologies (DIVITECH) may potentially offer a varietyof services beyond entertainment. Some of these are listed below.
Telemedicine-13 ecause of its high resolution and true rendition of color, advanced video communicationstechnologies could be used to transmit medical images such as x-rays, CAT scans, or color pictures of tissue toleading experts in distant cities for instant diagnosis. A distant expeti might even observe and provide advice duringa critical operation. People in rural areas with little access to world class medical facilities could particularly benefit.
Education—Recent advances in manipulating digital video data allow the viewer to interact directly withreal-world images. For example, the viewer can “stroll’ through ancient Athens at will, the computer selecting anddisplaying the appropriate audio and video signals (prerecorded on location) in response to the viewer’s direction.The viewer could similarly examine the effect of different strategies on the outcome of a battle; take apart and rebuildan auto engine; or dissect a frog, with detailed information available on demand on how each part works individuallyand with other parts. This ability to interact with what is being displayed will make these technologies far moreimportant to education than today’s TVs. Advanced video technologies could thus be used widely in education, frompre-school through medical school.
Simulation—Engineering simulation, including computer aided design of structures, electronic componentsand equipment, aircraft, and a host of others is already a vital market. As for interactive video, recent advances incomputer-generated images could extend simulations to such things as building design-where a prospective clientmight take a realistic “walk” through a proposed design.
Photography—Pictures taken by electronic still cameras could be displayed on a screen or sent over a networkfor immediate printing at a distant film developer. With computer assistance, photographic-quality images anddigital audio might one-day be edited almost as easily as words are today.
Telecommunications-A host of new services, ranging from videophones and teleconferencing totelemarketing new goods could become available to the consumer. One might even design a personal telenewspaperby using a computer program to search the news services, TV news, PBS educational programs, and others forwritten or video information of particular interest that could be stored and later displayed-for example, at breakfasttime.
Publishing-Desk-top publishing using personal computers has already revolutionized the business.Advanced video technologies will accelerate this by allowing the transmission and display of high-quality visualmaterial. NYNEX is now experimenting with the transmission of advertising copy between agencies and clientsover a fiber-optics network in New York City.
Defense-HRS technologies could enhance many of today’s defense technologies. Examples might include:using electronic cameras for recomaissance to eliminate the delays and logistics in processing film; providing highresolution maps for targeting and/or very close-in air support of ground troops; or improving cockpit displays forpilots+verlaying information about incoming missiles, ground fire, or aircraft with fuel availability and weapons’status on a display of the upcoming target.
SOURCES: Paul Kemezis, “HDTV Slowly Moves Into Medical Applications,” New Technology Week, Aug. 15, 1988; Arch C. Luther, “YouAre There. . . And In Control,” IEEE Spectrum, September 1988; William Booth, ‘‘BendingReality’s Borders,” Washington Post,Oct. 23, 1989, p. A3; “What’s Ahead in Phontx and High-Resolution TV,” Science Focus, vol. 3, No. 2, Fall 1988, New YorkAcademy of Sciences; U.S. General Accounting Offke, “High-Defiition Television: Applications for This New Technology,”December 1989.
Early generations of these technologies are video material in an electronic encyclopedia. Atalready affecting our professional lives in im- home or in the office, this fusion of computer,portant ways. High-performance graphics- communications, and imaging technologies willbased workstations and personal computers, make widely accessible a host of new serviceslaser printers, copiers, fax machines, and other (box l-l).technologies are revolutionizing the office. In-teractive video systems are becoming available Similarly, many government activities couldthat allow an architect to “stroll” through a benefit from HRS technologies. For example,building being designed, or a student to call up the Federal Aviation Administration (FAA) has
6 ● The Big picture: HDWarul High-Resolution Systems
already contracted with Sony (Japan) for high-resolution displays to monitor air traffic. Themilitary could use these technologies for train-ing simulators, command and control centers,teleconferencing, and aerial recomaissance—eliminating the delays and difficulties inherentin processing film. NASA could use HRStechnologies for deep space exploration, remotesensing of the Earth, and for monitoring launches.For example, higher resolution pictures wouldhave aided the analysis of the Space ShuttleChallenger tragedy. In March 1989, NASAconducted its first test of HDTV by videotapingthe launch of the Discovery shuttle and transmit-ting the pictures within the Space Center and asfar away as Orlando, Florida, by fiber-opticlinks.3
The role of HDTV as a consumer product inthe future information society remains unclear.Skeptics portray HDTV as simply providingbetter entertainment for ‘‘couch potatoes,” andclaim that there will not be a sufficient consumermarket to support a more complex or capabletechnology. Advocates portray HDTV as thebasic technology platform on which tomorrow’shome and perhaps even office informationservices will be built—a veritable keystone forthe video information archway of the future.
Although such scenarios suggest that HDTVmight eventually become the home informationcenter—providing entertainment, computer, andtelecommunications services-it is perhaps morelikely that these different services will insteadcontinue to be primarily provided by separate,specialized pieces of equipment. People simplywork that way. While the teenager is on thevideophone, one parent could watch video on abig screen in the family room, while the otherparent could use the computer in the study tobalance the monthly finances.
HDTV might thus be one of three platformsfor home information services, the others beingthe computer and the videophone. (In homes
that would not otherwise purchase a computer,the HDTV might serve as an affordable meansof providing some computing power and wouldthen open a wide range of services.) All threetypes of equipment will probably evolve com-mon basic designs that allow easy exchange ofinformation among them and could significantlyoverlap in the services they provide. Overtime,it may become increasingly difficult and moot todistinguish these different types of digital equip-ment from each other. The large, high-qualityscreen of the HDTV might be the most notabledifference.
These video information services will neitherreplace today’s media quickly, nor will theyuntil they provide significantly greater function-ality at an affordable price. For example, simplyreprinting a newspaper story on a bulky, hard-to-read ATV or computer display wiIl not inducepeople to give up the convenience of newspa-pers, which can be carried around and readanywhere. But video information services thatdeliver more in-depth information on a newsprogram upon request; can send a movie clip toa friend; or provide electronic yellow pages thatinclude video clips of restaurants that viewenmight want to try could attract a great manynewspaper readers (a possibility of obviouslygreat concern to the newspaper indust~).
Despite the potential impacts of HDTV andrelated HRS technologies on U.S. electronicsmanufacturing and on the U.S. communicationsinfrastructure, and, despite the opportunitiesthat new video information services may offer,the United States significantly lags Japan andEurope in developing and manufacturing manyof these products.
This report focuses primarily on consumerHDTV for several reasons: HDTV raises thornypolicy issues; HDTV is driving a number oftechnologies and manufacturing processes forHRS more generally; and HDTV may signifi-cantly impact our communications infrastruc-
3MCM Dohe~, 6’HDTV c~er~ capture Shuttle LaunclL” EZecrronic Engineen”ng Times, m. 20, 1989, P 19.
.$tbsho~d tie U.S. R= tie Baby Bells?” Business Week, *. 12, 1990, P lx.
Chapter l-introduction andSu~w ● 7
ture through possible radio frequency spectrumreallocation and perhaps by accelerating thedeployment of fiber-optic systems.5
MS attention is paid to intermediate forms ofAdvanced TV such as Improved Definition orExtended Definition TV (IDTV, EDTV): IDTVhas little impact on the communications infra-structure; EDTV’S impact on communicationsis modest and EDTV may be bypassed ifsimulcast systems are chosen; and neither IDTVnor EDTV are driving technologies as hard asHDTV is today.
kss attention is also given to alternativeforms of video entertainment and informationsystems such as interactive video: it has littleimpact on the communications infrastructure; itmay become an element of the discussion ofHDTV if flexible designs for HDTV receiversare chosen as the U.S. standard; and it facesmany of the same questions about consumerappeal as HDTV.
There is similarly little discussion of videoprogram production: it provides the UnitedStates a net $2.5 billion trade surplus—compared to a $5 billionb trade deficit forconsumer video equipment—and it has littleimpact on either U.S. manufacturing perform-ance or the U.S. communications infrastructure.
Finally, the market for video productionequipment is smaller than that for householdvideo equipment. There is less emphasis on themuch larger generic category of High Resolu-tion Systems—which represent much of futurecomputer and communications systems—thanfor HDTV. These areas will be referred tobriefly throughout this study, and will bediscussed in more detail in future OTA reports.
The Historical Development of HDTV (Ch. 2)
The Japanese have been selling HDTV studioproduction equipment since 1984 and are nowgearing up large-scale commercial productionof HDTV receivers. The Europeans began a
crash program in June 1986 and now lag theJapanese by just 2 to 3 years. In contrast, theUnited States will not even begin testing toestablish HDTV transmission standards untilmid- 1990 and U.S.-manufactured HDTVS couldnot likely be commercially available until 1993or 1994.
The Japanese effort is particularly notewor-thy. Japan considers HDTV to be an importantstep into the future information society. Itforesees numerous technological linkages be-tween HDTV, High Resolution Systems, andother parts of its highly successful electronicsindustries. As in other countries, the Japaneseface the “chicken-and-egg” problem of whoinvests first: consumers will not buy HDTVSuntil the price comes down and there are enoughHDTV programs to watch, but manufacturerscannot reduce the unit price until sales volumesare large, and movie producers and broadcasterswill not provide HDTV programs until there isa sufficient audience. The Japanese believe thatthe best way to overcome this problem is bysharing the costs and risks between the gover-nment and the private sector. The JapaneseGovernment has therefore spun a complex webof direct and indirect R&D, financial, andmarket promotion efforts to stimulate the devel-opment of the HDTV market.
In contrast, all U.S.-owned firms, exceptZenith, have abandoned the TVreceivermanufac-tunng business. Many factors contributed to thisexodus of U.S. firms, including: the relativelypoor manufacturing performance by some U.S.fms (see app. A); and the failure of the U.S.Government to protect U.S. industry fromforeign trade violations (see app. B). As a resultof the loss of the U.S. TV market, the little workdone on HDTV in the United States has largelybeen by or for foreign-owned consumer elec-tronics firms or by small, underfunded univer-sity programs and entrepreneurs. The DefenseAdvanced Research Projects Agency’s (DARPA)planned R&D program is the most significant
5~e tem DW i5, on oca5iom Ud loosely to include Advanced TV system gener~lY.
6sa~ ml, ~temtio~ Trade Ahs~tioq Us. @~@ of co~er~, persod cOmrtNUlhtiO~ NOV. 16, 1989.
8 ● The Big picture: HL3TV and High-Resolution Systems
recent step to reverse this, but there are seriousquestions about the Administration’s commit-ment to this effort.
Communications Technologies (Ch. 3)
Historical, economic, and technological fac-tors have combined to provide the United Statesfive major electronic communication media:terrestrial and satellite communications usingthe radio frequency spectrum; coaxial cables forTV; twisted copper pairs of wires in thetelephone network; and, recently, optical fiberfor the high traffic “backbones” of both thetelephone and cable TV networks.
The greater information content of HDTV’Sbroadcast signal will require changes in theallocation and use of the existing TV broadcastchannels and could potentially free spectrum foruse in mobile communications and other serv-ices. HDTV could spur the use of DirectBroadcast Satellites (DBS) and might also speedbroader use of fiber optics in cable and possiblytelephone networks by stimulating market de-mand for high-quality video entertainment.
Television Technology (Ch. 4)
Video entertainment systems may take avariety of forms in the future, including (in orderof increasing picture quality) Intermediate Defi-nition TV, Extended Definition TV, and HighDefinition TV; or perhaps various forms ofinteractive video either in conjunction withthese Advanced TVs or as separate systems. Ofall ATVS currently under advanced develop-ment, HDTV may have particular consumerappeal because of its greater potential forproviding viewers the feeling of ‘being there’that one sometimes gets in watching a high-quality motion picture up-close and, for exam-ple, having the sense of moving with a stuntplane when it makes a fast turn.
Television systems have three general func-tions—production, transmission, and receptionof TV programming—all of which will requiresubstantial changes from today in order to makethe transition to HDTV.
Production
International efforts are currently focused ondeveloping common production formats thatwill allow easy conversion between differentregional standards. Earlier attempts to establisha single global production standard founderedon the lack of compatibility between existingsystems and the cost of converting from one toanother. Earl y recognition by European interestsof the potential competitive threat to theirdomestic electronics manufacturing posed byhaving a single common standard based on theJapanese system was also a factor in stoppingthe establishment of a single global productionstandard.
Transmission
It is difficult to squeeze the greater informa-tion content of an HDTV signal into the charnelbandwidths allocated to terrestrial broadcasting,especially given the inefficiencies of conven-tional color TV signals. The Japanese andEuropeans have therefore opted to insteaddevelop HDTV services through Direct Broad-cast Satellite (DBS) systems operating at higherfrequencies not currently heavily used. In theUnited States, the greater importance of theexisting broadcasting system, issues of localismand programming diversity, and other factorsmake the development of a terrestrial HDTVbroadcasting capability for HDTV more impor-tant than it is in Japan or Europe.
Terrestrial transmission systems proposed forAdvanced TV services in the United States arebased on augmentation of the existing NTSCtransmissions with an additional 3- or 6-MHzsignal; or on simulcasting (simultaneously broad-casting) in the charnels now left vacant toprevent interference between stations (taboochannels). Although the NTSC system was aremarkable triumph when it was developed—given the electronics technologies of the 1950s—more efficient use of the broadcast spectrum isnow possible with today’s electronics technolo-gies. Augmentation systems will continue touse, in part, the NTSC system and may thus tendto lock into place less efficient use of the
Chapter 14ntroduction and Summary ● 9
broadcast spectrum. In contrast, simulcast sys-tems might eventually allow a large amount ofradio frequency spectrum now reserved forNTSC broadcasts to be vacated and used forsuch services as mobile communications.
Receivers
Advanced TV receivers of greatest interest inthe near-term are the Multiport Receiver and theOpen Architecture Receiver (OAR). Multiportreceivers would be adaptable to a limited rangeof predetermined broadcast standards and wouldprovide limited access for adding voice/data/video communications.7 OARS follow the pathpioneered by the personal computer industryand would be adaptable to a much broader set ofbroadcast standards, personal communications,computer functions, or other services that mightbe of interest to consumers. This could open ahost of new markets for entrepreneurs.
Advanced TV systems are evolving naturallyfrom today’s conventional, largely analog sys-tems through the increasing use of computer-like digital electronics. This evolution of TVtechnology to digital electronics is inexorable,even if its speed is uncertain. Similarly, com-puter and communications systems are evolvingtowards greater use of still and video images asnow seen primarily on TV. Advanced TV,computer and telecommunications systems areexpected to continue to evolve towards reasona-bly common forrns— HRSS. The impact ofHDTVS and HRSS if connected to a nationalfiber communications network could revolu-tionize information services.
Technological Linkages (Ch. 5)
HDTV is driving the state-of-the-artin certaindigital signal processing, data storage, display,packaging/intercomect, and other technologies,as described above. As with consumer electron-ics generally, HDTV will also push the limits ofcost-effective manufacturing. This could be oneof the most important impacts of HDTV for theUnited States.
The expectation of a large market is forcingpotential HDTV manufacturers to push thestate-of-the-art in several areas of HDTV-related technologies. These technological link-ages could assist HDTV producers in other HRSmarkets. Simply developing technologies doeslittle good, however, without markets to sell in.
In the past, the United States has tended toassume that if technologies were developed,markets would follow. Faced with large, oftenvertically integrated and aggressive foreigncompetitors, market shares may be as importantas technology development. These foreign firmsare much more likely to use their own internallydeveloped semiconductors and other compo-nents than to buy them from a U.S. firm (asmight a vertically integrated U.S. firm). Someforeign firms are also more likely to buycomponents they don’t make internally fromlocal suppliers with whom they have long-termpreferential relationships. These relationshipscan be very difficult to crack, regardless of theprice or performance of the ~echnology offeredby the outsider.
Controlling the market, however, is notenough. Even with a strong technology base and70 percent of the world’s personal computermarket, the United States still lost the (mer-chant) DRAM industry to Japan. This was dueto a variety of factors, including: relatively lessefficient manufacturing by some U.S. firms(app. C); foreign trade violations and closedforeign markets (app. D); and industry and tradepolicies in Japan that encouraged heavy invest-ment in and rapid development of their domesticindustry.
Advanced Television Markets (Ch. 6)
Forecasts based on analogies with past suc-cessji.d products, project U.S. sales of HDTVs at10 to 15 million sets annually within 15 years.Some expect a large business/industry marketfor HDTV equipment to develop much sooner.Yet others suggest that the HDTV market will
7AxpadG. TothandJosephDonahue, ‘‘An MuMport Receiver: Preb.inary Analysis, ’ First Report of the EIA ATVMulfiportReceiver Com”ttee,Sept. 11, 1989.
lo ● The Big picture: HDTVand High-Resolution Systems
not develop and that consumer needs can be metby intermediate products such as ImprovedDefinition TV or Extended Definition TV, orinstead byproducts such as interactive video. Infact, there will probably be markets for all ofthese products. The large uncertainties in howvideo entertainment markets will develop in thefuture should not obscure the underlying trendin consumer video towards digital electronics,high-performance fiat panel displays, high-density optical and magnetic recording, andother key technologies.
The HDTV market projected by these fore-casts varies between $5 billion and $12 billion(1988 dollars) by 2003. VCRs, movie produc-tion, and broadcasting equipment increase thesepotential values. The overall U.S. consumerelectronics market was worth $30 billion infactory sales in 1987 and is growing rapidly.8The High Resolution System market will belarger yet, encompassing a broad array ofimaging and image processing markets in thecomputer, consumer electronics, and telecom-munications sectors.
The relative importance of the future HDTVmarket has also been compared with that of thecomputer sector. Such comparisons are not veryuseful; there are simply too many uncertainties.Regardless of the precise form consumer videoproducts take in coming years, the consumerelectronics sector will continue to be a large andimportant market, and video entertainment willcontinue to be its most important component.Computers and communications equipment willalso make greater use of still and video images,but they will probably lag consumer video indriving some of the key technologies such asdigital signal processing, high-performance dis-plays, optical and magnetic storage, and certainmanufacturing processes.
U.S. Manufacturing of HDTV
Proponents argue that government supportfor HDTV and related HRS R&D might serveseveral national goals.
HDTV might serve as a stepping stone forU.S. firms to reenter consumer electronicsmanufacturing. Consumer electronics is a largemarket; it also supports many upstream indus-tries. For example, roughly one-fourth of Japa-nese semiconductor output is currently used inconsumer products, and TVs and VCRs are amajor portion of this.g Similarly, 70 percent ofthe $8 billion (1988) world display industry isfor consumer products. There will continue to bea large demand for video entertainment andother consumer electronics equipment, regard-less of the specific form these technologies take.As noted above, consumer electronics is also animportant driver of manufacturing processes.
HDTV might serve as a principal driver ofmany High Resolution System (HRS) technolo-gies due to the exceptional demands it places ondisplay, video processor, storage, and othertechnologies. At one time, U.S. fiis couldignore the consumer electronics market at lessrisk because analog electronics were used. Withthe shift of consumer equipment to digitalelectronics, the linkages to the computer andtelecommunication industries are becoming muchmore important. U.S. firms can no longer ignorethe consumer market with impunity.
HRS technologies will be very important tothe computer and communication industries incoming years. As a forerunner to video, manu-facturers sold roughly 9 billion dollars’ worth ofhardware and software in the United States in1988 for commercial graphics applications, withsales expected to rise to over $25 billion by1993. 10 There are similarly large markets fordisplay technologies; for imaging equipmentsuch as facsimile machines; and for telecommu-
8E1=moIIic Industies Association Consumer Electronics Annual Review, 1988 cd., W@@Von, DC.
g“JapanElectronics Almanac,” Dempa Publications, various years; Kenneth FlarnnL Brookings Institution personal communication VCRs aloneaccount for 12 percent of total Japanese IC production: “TV’s I-Iigh-Stakes, High-Tech Battle,” Fortune, Oct. 24, 1988.
IO~C~~ &apfics Revolutio@” Business Week, NOV. M 1988.
Chapter I-hu’reduction and Summary ● 11
nications equipment, including that using wide-band switching and fiber optics. American firmsare increasingly lagging foreign competitors inmany of these technologies; more R&D isneeded, together with greater attention to manu-facturing with quality at affordable prices.
Finally, HDTV or related HRS might serve asthe home terminal on a national fiber informa-tion network. In the midterm, the IntegratedServices Digital Network (ISDN) will makepossible a host of information services short ofhigh-quality real-time video. In the longer term,a national two-way fiber network would makepossible many desirable video information serv-ices. If a national fiber network is made anational goal, then policies to aid its implemen-tation should be put into place in the near-term.These might include a framework to encourageadditional effort in developing and manufactur-ing video information systems.
Skeptics insist that if HDTV is important,industry will invest in it independently and willdo so more wisely than the government could;that there is no need for government support.
Skeptics argue that HDTV is likely to be arelatively small market (at most $30 billion in 20years) compared to the entire world electronicsmarket (which is already $450 billion or more)and can therefore be ignored. The same argu-ment could be made about almost any segmentof a market and ignores the relative importanceof specific products in driving the state-of-the-art in important technologies. For example, thetotal U.S. supercomputer market was just $1.4billion in 1988, but supercomputers are veryimportant in driving a number of leading-edgetechnologies. DRAMs were just a $2.5 billionmarket in 1987—5 percent of the total worldsemiconductor market-but DRAMs drive manyimportant semiconductor manufacturing tech-nologies. HDTV and related HRS are similarlydriving many important technologies (ch. 5).
In contrast to American firms, many foreigncompetitors seem much more cautious aboutabandoning markets. This may be due to: thesignificant financial and technical skills needed
to reenter high-technology manufacturing; thepotential linkages with other existing markets;or the new opportunities that being in a marketmay create. Being in many parallel markets canalso provide economies of scale in R&D andproduction of the underlying components, andtends to insulate a firm from downturns in anyparticular market segment.
Some skeptics argue that trying to outguessthe market by backing a specific product isfoolish: instead of HDTVS, consumers mayprefer lower-cost systems with somewhat lessresolution, or systems that provide much moreinteractivity. This point may prove to be true.Consumer markets will likely develop aroundeach of these as well as other applications.Neither industry nor government can guess theprecise form that these markets will take; nor isit necessary to do so. There is a clear trendtowards video information systems, which in-volve many of the underlying technologies nowbeing most strongly driven by HDTV.
Other skeptics suggest that it doesn’t matterif HDTV and related electronics products aremanufactured abroad. They even speculate thatit helps the American consumer if these productsare dumped in U.S. markets-that this effec-tively gives us something for nothing. Viewednarrowly, consumers agree. Consumers wantthe best possible HDTV programs and picturesat the lowest possible price, and without havingto pay any subsidies-+ither through taxes tosupport R&D, or added fees to cable companiesor other distributors.
This argument is based on a questionabledefinition of consumer interest. If U.S. consum-ers are to buy these goods and maintain a highstandard of living, they must have high-payingjobs in a strong economy. The United States willlose potential jobs-especially the skilled jobsneeded to ensure a high standard of living in theUnited States-if the electronics or the displaysfor HDTVS and related HRSS are producedoffshore. Consumers in other countries haveoften paid taxes and price subsidies in order to
12 ● The Big picture: HDTV and High-Resolution systems
maintain their jobs and develop their manufac-turing sector.
Finally, some skeptics insist that if HDTV ispotentially such a large market and so importanta driver of technology, then industry wouldenter it. The implication is that U.S. industry’shesitation to enter indicates that HDTV is likelyto be a turkey. This might prove true, yetJapanese and European industry have embracedHDTV. This difference in attitude may, in part,be due to: the history of the U.S. consumerelectronics industry; the supports Japanese andEuropean industry receive from their gover-nments; and the barriers now facing prospectiveU.S. entrants.
U.S. industry has largely abandoned con-sumer electronics manufacturing. In today’scolor TV industry, for example, the only signifi-cant American-owned f~m remaining is Zenith,which has just 14 percent of the U.S. (2.8percent of the world) color TV market. In orderto remain in the TV business, Zenith recentlysold its highly profitable computer division toGroupe Bull, a 90 percent French Govemment-owned firm.ll In contrast, there are currently 10Japanese fums, 3 European firms, 2 Koreanfirms, and 1 Taiwanese firm producing TVs andcomponents at some 32 locations in the UnitedStates.12 A significant portion of this work isscrewdriver assembly of electronic componentsand subassemblies manufactured elsewhere.
The history of the U.S. consumer electronicsindustry is long and tortured. Numerous factorscontributed to its decline. Appendixes A and Bfocus on two of these many factors for the colorTV industry: less-competitive manufacturing bysome U.S. firms than their foreign competitors;and trade violations by foreign firms coupledwith a failure of the U.S. Government toadequately protect U.S. industry. Given this
history, U.S. industry has good reason to becautious about reentering consumer electronics.
American manufacturers face significant bar-riers if they are to reenter the consumer electron-ics industry and manufacture Advanced TVs:
Luw Market Share—Foreign-owned fmscontrol the U.S. domestic TV market.Foreign competitors can and do use thisbase to hone their manufacturing skills,build their production and distributioninfrastructure, and generate revenues fordevelopment of ATVS. This might alsoenable foreign firms to quickly initiatelarge-scale production of any imovationdeveloped for the U.S. market-perhapsmore quickly than the U.S. innovator.Lznv Profi”ts-The U.S. TV market todayprovides little or no profit. Zenith, forexample, has not made a full-year’s profiton its television business since 1984.13 FewU.S. firms could justify entering such abusiness to their stockholders.Large Capital investments—Manufactur-ers must risk large upfiont investments andwithstand years of losses in order to createan ATV market. These investments arelarge and are increasing rapidly as manu-facturing processes grow more complex.Forexample, capital equipment foraminimum-efficient-scale, state-of-the-art DRAM fab-rication facility now costs perhaps $300-$400 million.Manufacturing Skills—Many U.S. firmslag behind foreign competitors in a numberof manufacturing technologies importantto the production of HDTV (ch. 5).Inequities inForeignTra&-U.S. manufactur-ers may not trust the government to ade-quately protect them from foreign dump-ing; they may also have little faith that theywill be able to enter or export to the
1 IEvelyn Richards, “French to Purchase Zenith Computer Unit,” Washington Post, Oct. 3, 1989, p. Cl.
lzElectronic ~dus~es Associatio~ HDTV Info Packe~ and Suzanne Heato~ Electronic Industries AssociatiorL perSOfId Commtiatioq Dec. 27,1988.
13Jew K. pe~~m, test~onY at h~~gs ~fore tie House su~ommi~ee on Tel&omm~catiom ~d Ftice, Committee on Energy andCommerce, Sept. 7, 1988.
Chapter I-Introduction and Summury ● 13
Japanese or to export to the European ATVmarkets. Without being able to penetratethose markets, U.S. manufacturers may beunable to realize the same economies ofscale as their foreign competitors, who cancompete in the United States. Foreignproducers with protected home marketscan expand production with much greaterconfidence that it will pay off than U.S.producers who have no such assured mar-kets for their sales.
POLICY ISSUESU.S. industry thus faces significant barriers to
entering HDTV manufacturing. Japanese f~msstruggled under somewhat different, but inmany respects even greater, disadvantages whiledeveloping their computer industry in the 1960sand 1970s. Yet through a variety of mechanisms(app. E), they have developed world-classcapabilities across the spectrum of computertechnologies and products.
Developing competitive Advanced TV, con-sumer electronics, or HRS manufacturing indus-tries in the United States carries many potentialbenefits, especially in strengthening manufac-turing abilities in electronics and in developingtechnologies that have important spillover appli-cations in other branches of the electronicsindustries. To succeed in consumer electronics,fums must meet demanding tests of manufactur-ing excellence: the ability to turn out reliable,well-designed goods at high volume, whilekeeping costs competitive. What firms learn inmeeting these tests for consumer electronics canthen be applied to other products, such ascomputers and telecommunications equipment.
Technological linkages between HDTV/HRSand other industries are equally significant. Theknowledge gained in developing core technolo-gies for these systems may sometimes providea critical edge in competition for other marketsin this fast-paced industry. U.S. firms might alsofind it difficult to get access on equal terms tosuch technologies if they are developed by
foreign competitors. For example, if a smallgroup of like-minded foreign companies were togain control of an important component, such asflat panel displays, U.S. fms might be vulnera-ble to overpricing or outright denial of sales.This is not an unheard of practice. Fujitsu, whichproduces both semiconductors and supercom-puters, is reported to have delayed for manymonths in supplying critical semiconductors tothe U.S. supercomputer company Cray Re-search, Inc.; and Nikon reportedly withheld itslatest and best lithographic stepper from U.S.semiconductor manufacturers.
For the reasons outlined in the previoussection, the prospects look poor for U.S.-ownedfirms to reenter manufacturing in the consumerelectronics field. A few foreign-owned firmsbased in the United States are pursuing HRSresearch here. Possibly, with some form ofgovernment encouragement, either in develop-ing technologies or in rebuilding manufacturingcapability, or both, U.S. firms might becomemore interested in risking their own capital andefforts in the field as well. That possibilityimmediately raises the question of what exactlyconstitutes a U.S. firm? This question is ex-plored below. But U.S. efforts to support ATVor HRS technologies have been fragmented andabortive so far, in part because such supportraises policy issues that have not yet beenresolved in public debate. This report does notattempt to resolve them either, but the followingissues should be addressed if Congress wishes topursue options to support a domestic HDTV/HRS industry.
The existence of growing international competi-tion—and the clear decline of U.S. manufacturi-ng competitiveness—has raised the generalpossibility of increased government support forindustry. This makes the question of corporatenationality central. Government support shouldgo to serve the interests of American citizens inthis new global environment, but, in RobertReich’s words, “ Who 1s Us?”14
14Ro~ Reich, “W%O IS US?” Harvard Business Review, January-February 1990, pp. 53-62.
14 ● The Big picture: HDTV and High-Resolution systems
This question has become increasingly press-ing because there have been two fundamentalchanges in the American economy. First, manylarge American firms have become global: theydo an increasing proportion of their businesselsewhere, and U.S.-owned firms are doingmore sales, manufacturing, and even design andR&D off-shore, in markets like Japan or the EC,or in low-cost platforms like Malaysia andThailand. Second, foreign-owned fms havediversified into the United States, and many nowhave substantial manufacturing efforts in theUnited States. Foreign-owned R&D and designfacilities are also opening in the United States,and some foreign firms have acquired U.S.high-tech businesses in electronics and otherindustries, and support their R&D as well.
Corporate nationality is an exceedingly com-plex issue, and only a brief overview is possiblehere.15 Ownership has traditionally been thesole criterion of nationality. The fact thatAmerican firms were owned by Americans,tended to site their production in the UnitedStates, and made almost all their sales in theUnited States allowed the ownership criterion torepresent all the other attributes of nationality.In addition, there was the unspoken assumptionthat American firms would in some sense act tomaintain American national interests. But theshift toward off-shore production, technologydevelopment, and contracting by U.S. manufac-turing fms has undercut this assumption. Andownership itself is not an unambiguous concept,as ownership does not always mean control: aminority share ownership can exercise control insome cases while a majority ownership may notbe sufficient to take key decisions in others.There is little sign that U.S. fms are in factputting the national interest before their own—
partly because that would be a breach of theirfiduciary duty to their shareholders.lb
Thus an alternative view of corporate nation-ality is now emerging, where the key criterion isthe contribution of the fm to the nationaleconomy, or national competitiveness. Owner-ship is only one part of that contribution; not onethat is easily quantified. While profits are arelatively small part of a firm’s overall directcontribution to the economy in which it oper-ates, they can be a critically important source offunding for R&D, and growth and wealth. Thus,even though their share of value is small, profitsmay be disproportionately important. Nor is theflow of profits clear: they can in part berecaptured by foreign fms plowing back profitsinto R&D or capital investments in the UnitedStates or by Americans who are fast increasingtheir holdings of foreign securities. The same istrue for foreigners, whose increasing holdings ofU.S. securities cause profits to be repatriatedelsewhere.
Aside from ownership, most value created bya firm comes from research, development,manufacturing, related services, and sales. Eachof these elements can be located in the UnitedStates or elsewhere, with differing contributionsto the national economy. Recent testimonybefore Congress has stressed the importance ofperformance-oriented criteria for determiningwhether a firm qualifies as American-theextent to which the fm provides jobs, taxreceipts, R&D, technology transfer, and otherbenefits to the United States, or contributespositively to the U.S. trade account.17
An extension of this view is offered by Reich,who argues that the critical contribution of afm to the economy lies in its support for aworld-class work force. The kind of business
15A ~ql~~ ~ysis ~~ & d~~l~@ ~ the fofi~q O’J’A rqrt on ~de ~ indus~ policy. This ~ ~ the tid and hid report Of theassessment on Technology, Innovation and U.S. Trade, the fmt two of which were: U.S. Congress, Office of Technology Assessmen4 Paying the BiZ/:Manz@actun”ng andAmerica’s TradeDej7cz”t, O’K4-ITE390 (Washington, DC: U.S. Government Printing Office, June 1988); and U.S. Congress, ~lceof Technology Assessmen4 Making Things Better: Competing in Manufactun”ng, op. cit., footnote 1.
16Reic4 op. cit., footnote 14.
17Jo~ -e, ta~ony at h-s &fore the House su~o~tt~ on science, Re~ch ad Technology, COmmittee on Science, SpaCC, ~dTwhnology, and the House Subcommittee on International Scitmtiilc Cooperation Committee on Science, Space, and Technology, Nov. 1, 1989.
Chapter 14ntroducti0n and Summary ● 15
being conducted in the United States will haveimplications for the kind of work force beingproduced, and hence for the attractiveness of theUnited States as a site for high-technology, highvalue-added manufacturing.18
Some witnesses and some Representatives atthe hearing19 also underlined the importance ofreciprocity between the treatment of foreign-owned firms in the United States and thetreatment of U.S.-owned firms in the corre-sponding foreign country. These commentsreflected views similar to those embedded in S.1191, the FY 1990 appropriations bill for NIST
Corporate nationality becomes an extremelycomplex issue once the simple but perhapsoutmoded criterion of ownership is abandoned.First, the various activities of business—production,R&D, support activities, sales, trade—must be weighed against each other. Is R&Dintrinsically more valuable to the United Statesthan an equivalent amount of value added insales? How should quantitative criteria (e.g., acertain percentage of value added) be balancedagainst qualitative criteria (e.g., a commitmentto doing R&D in the United States)?
These problems can be solved in principle,but practical application could be difficult. TheEuropeans have already found difficulties evenin determining the percentage of locally pro-duced content in some goods, and recently losta key anti-dumping case before a GATT tribu-nal, partly over this issue.20 Qualitative judg-ments are even more difficult. The combinationcould threaten the viability of the GA~ in thefuture..
Finally, the complexity is compounded by thedifferent definitional contexts. Definitions thatmay be appropriate for controlling foreign direct
investment may not be useful when qualifyingfirms for direct government support, or R&Dcontracts.
Despite the difficulties, nations have foundways in which to discriminate on the basis ofcorporate nationality. The EC is funding anumber of R&D consortia. It appears that atwo-tier system of discrimination maybe evolv-ing: firms which are foreign-owned but whichact as good corporate citizens (developing afully integrated manufacturing complex with theEC, even exporting back to their country ofownership) appear to have better access togovernment support than foreign-owned firmswhich are not such good citizens. The bestgovernment support may be reserved for firmswhich are not foreign-owned-even though theexplicit authority for such discrimination issometimes hard to find. So far, decisions havebeen made case-by-case.
Advanced TV is clearly a case where theUnited States must play catch-up if it is to be inthe game. All but one U.S.-owned companyhave left the business of making televisions, andthe remaining company is not financially strong.The questions are whether and how to support anindustry that is practically nonexistent, andwhere foreign producers are clearly ahead. In thepast, countries that have succeeded in doing thishave relied heavily on protection from foreigncompetitors—in the case of Japan, protection ofdomestic markets not just against imports, but inmany cases from foreign direct investment aswell. In Europe, which is also playing catch-upwith the United States and Japan in a number ofindustries, the inclination against foreign firms’participation is less pervasive, especially insome countries. American and Japanese firmshave been permitted to participate in some
18~Reich*~ ~cle, T~d Hixon and ~ch “Wrnball distinguish between importers, simple screw-driver assembly (like the plants that make mostcmsumer electronics products in the United States), plant complexes (which produce and even modify existing designs, but do not provide a full lineof R&D in support of new ones), and fully integrated business operations; the latter is a relatively undefined concept which stresses tight integrationof basic research development designj and production. Reichj op.cit., footnote 14.
q+~w ~fom tie House Su&o~ttW on Sciace, Research and Technology, Committee on Science, Space, ~d T~huoIogY, and tie HouseSubcommittee in International Scientific Cooperation Committee on Science, Space, and Technology,“What is a U.S. Company?” Nov. 1, 1989.
-illiamDullforce, “JapanScores Victory overEC onDuties,” Financial Times, Mar. 29, 1990; Peter MontagnonandLucy Kellaway, “ECRefuses‘Ib Adopt GATT Report on Dumping, ” Financial Times, Apr. 41990.
16 ● The Big Picture: HDTV and High-Resolution Systems
EC-funded R&D programs, and foreign invest-ment is now encouraged, especially in industrieslike electronics and motor vehicles. None-theless, domestically owned firms still seem tobe favored for access to government-and EC-funded programs aimed at hastening technicaldevelopment.
The European and American governmentsface different problems in HDTV; Europe stillhas a domestic, European-owned consumerelectronics manufacturing industry. There arecompanies that produce televisions, many ofthem with production and even research facili-ties in the United States. Most are foreign-owned. Sony, Philips N. A., and Thomson areinvolved in ATV, and would probably bewilling and able to take advantage of anygovemment-supported program to foster HRStechnology development. Once again, that raisesthe question of what criteria the U.S. Gover-nment would establish to determine the participa-tion of companies.
The interests of firms and the interests ofnations are not always the same. Americanfirms, like European and Japanese firms, areincreasingly likely to be involved in a variety ofinternational cooperative agreements with otherfins. Multinational firms have many choices ofwhere they will perform R&D and manufactur-ing. Right now, the United States is not anattractive location for developing and manufac-turing televisions and other consumer elec-tronics products, except for foreign-owned firmswith external sources of capital and otheradvantages stemming horn foreign bases ofoperation. To generate interest among domesticfnms in reentering the business of developingand making televisions, the government proba-bly will need to change the rules for operatinghere, including altering the capital and invest-ment market for manufacturing in the UnitedStates.
In the long run, the most promising way ofassuring that domestic efforts to support anynew technology will result in domestic valueadded, is to make the United States a moreattractive location for manufacturing. The UnitedStates now has disadvantages in cost comparedwith many developing nations, and disadvan-tages in its financial environment and quality ofhuman resources compared with many ad-vanced nations, including Japan and much of theEC. In addition, the EC and Japan, amongothers, have substantial government programsto support new technology development anddiffusion of manufacturing technology. TheUnited States compares poorly here, too. Im-provements in these areas will help to ensurethat any government support of new technolo-gies or infant industries are more likely to leadto domestic development and manufacturing.
NATIONAL SECURITYHigh Resolution Systems (HRS) and related
technologies are likely to play an important rolein future military systems—analyzing the bat-tlefield, targetting the enemy, or parrying theenemy’s attack could all benefit from high-quality, real-time video information (see boxl-l). HRS technologies will, however, probablybe driven primarily by the needs of the commer-cial sector.
The defense strategy of the United States haslong relied on technologically superior weaponsto overcome numerically superior foes. Elec-tronic technologies are a critical element in thisstrategy. The broad loss of U.S. leadership insemiconductor and other electronics technolo-gies, particularly in their manufacture, raisessignificant concerns for U.S. defense capabili-ties in the future.21
Dependence on foreign-sourced technologiesgenerates risks to the U.S. defense posture:supply lines might be disrupted during a crisis;
zl~w Ofthe Under Secrew of Defemefor Acquisition “Report of Defense Science Board Task Force on Defense Semiconductor Dependency,February 1987; U.S. Congress, OffIce of Technology Assessment, The Defense Technology Base: Introduction and Overview, O’E4-ISC-374(Washington, DC: U.S. Government Printing Oftice, March 1988); U.S. Congress, Office of Technology Assessment Holding the Edge: Maintainingthe Defense Technology Base, OTA-ISC-420 (Washingto~ DC: U.S. Government Printing Office, April 1989).
Chapter l-introduction and Summary ● 17
a supplier might withhold critical componentsdue to pressure from adversaries; or an adver-sary might gain access to critical technologiesmore easily if they are foreign- rather thanU. S.-sourced. For example, the Soviet Unionwas able to purchase a sophisticated millingmachine from Toshiba (Japan) and Kongsberg(Norway) in 1987. This technology enabled theSoviets to construct much quieter propellers fortheir submarines and has thus greatly reducedtheir detectability.
Foreign suppliers might also judge theircommercial interests as more important thanU.S. security interests. For example, they mightwithhold state-of-the-art technologies that theUnited States needs for defense applications inorder to gain a commercial edge. Such withhold-ing may already occur in the commercial sectorfor semiconductor manufacturing equipmentand certain computer chips, among others.
On the other hand, the performance attainableby U.S. weapon systems maybe reduced in theabsence of the best available technology, someof which is commercially available from ourallies. It may also be more expensive to procuresystems solely from domestic sources ratherthan from lower cost foreign sources.
The defense market is no longer large enoughto drive the development of many electronicstechnologies. In 1987, the Defense ScienceBoard Task Force on Semiconductor Depend-ency found that, in contrast to the 1950s and1960s, DoD procurement of semiconductorswas too small compared to civilian markets to beof much importance to the overall semiconduc-tor industry: It concluded, however, that ahealthy semiconductor industry was critical tonational defense. Defense applications may alsolag far behind the state-of-the-art due to the longprocurement times. The same point could bemade about many other electronics technolo-gies, from components to computers, and islikely to be the case for FIRS technologies aswell.
The Defense Advanced Research ProjectsAgency (DARI?A) is supporting generic R&Din HRS displays and display electronics (ch. 2),but it will not be able to leverage more than asmall fraction of the R&D that would beconducted by a viable civilian HDTV/HRSindustry. Further, DoD and DARPA do not havethe legislative authority to directly promote acivilian HDTV/HRS industry.
A strong civilian HRS technology base isnecessary if many HRS technologies are to beavailable for defense needs at all. The low costsrealized for HRS technologies in the commer-cial sector, however, will not be automaticallytranslated into low-cost HRS for defense appli-cations. The complexity and specialized natureof defense systems results in long productcycles, high R&D and engineering costs, andstringent performance and reliability criteriathat may have little relationship to commercialneeds—such as for electronics that can with-stand high levels of radiation. Further engineer-ing development of commercially ‘availablecomponents is often necessary; and even whencommercial components can be used, they areoften just a small fraction of the total systemcost. Byzantine procurement practices also keepcosts high.22
THE COMMUNICATIONSINFRASTRUCTURE
HDTV could involve much of the U.S.communications infrastructure-terrestrial, cable,and satellite broadcasting; mobile communica-tions; and potentially even the telephone compa-nies. The current terrestrial broadcast spectrumallocation and transmission standards have beenin place for nearly 40 years. Accommodating thelarger information content of an HDTV-qualitypicture could force changes in the frequencyallocation and more efficient use of the spec-trum. These changes would also have conver-sion costs and create competitive tensionsamong the media. HDTV opens new opportuni-
zz’’spec~ Issue: The Price for M@,” IEEE Spectrum, November 1988; U.S. Congress, OffIce of Technology Assessmen4 Holding the Edge:Maintaining the Defense Technology Base, op. cit., footnote 21.
18 ● The Big picture: HDTV and High-Resolution Systems
ties to develop standards and systems that allowfor an easy and flexible transition to futurecommunications systems such as interactivehigh-resolution video on optical fiber.
Recorded Media
HDTV might be first introduced into the U.S.market through recorded media such as HD-VCR tapes, which are not subject to FCCregulation. The sale of HDTV-quality video-tapes and associated consumer electronics equip-ment might create a market that would thendefine a de facto standard for all U.S. media,whether optimal or not.
HD-VCRS may not require data compressionto the extent needed for terrestrial broadcasting.This would allow HD-VCR producers to sethigher quality standards than might be practicalfor terrestrial broadcasters or cable operators. Inthe longer term, wider bandwidths might con-tinue to provide HD-VCRS or other recordedmedia a competitive advantage over broadcastmedia+ able, terrestrial, satellite—for certaintypes of programming.
Terrestrial Broadcasters
HDTV could dramatically impact terrestrialbroadcasting. A 1988 FCC study found that withthe current geographical limits and channelseparation requirements,23 increasing the TVchannel width 50 percent (several of the HDTVproposals would require greater bandwidth thanthis) might force a quarter of today’s TV stationsoff the air.”
Policymakers could use the introduction ofHDTV as an opportunity to reexamine the entirequestion of spectrum allocation for the first timesince the current system was defined in 1952.25
The standards and frequency allocations madein the early days of TV broadcasting wereintended to keep the cost of receivers withinreach of the mass market by using then currenttechnology. There was little need then to con-serve the spectrum.
Technological advances since 1952 allowmore efficient use of the spectrum at littleadditional cost, permitting more channels ofhigher quality to be packed into less space. Thespectrum saved can be used for other services,such as mobile communications.2G
The FCC could choose an augmentationpolicy that would minimize the impact HDTVtechnologies have on spectrum use. Existingbroadcasters would be granted a 3- or 6-MHzchunk of spectrum—most likely one of the‘‘taboo channels’ (one left vacant by regulationto reduce interference between local stations)—in addition to their existing 6-MHz NTSCchannel in order to transmit the added informa-tion that HDTV needs to create a high-qualitypicture. If a taboo channel was unavailable, thena noncontiguous charnel would be used. Thewider the frequency separation between themain NTSC channel and the augmentationchannel, the more likely they would sufferdifferent types and amounts of distortion—making it difficult to meld the two signalsseamlessly into one picture.
Systems that augment NTSC broadcasts wouldtend to lock in the same inefficiencies of theNTSC technology that currently hamper theindustry (ch. 4). This might prevent the develop-ment of additional broadcast TV channels orprevent other uses of this spectrum.
23~wwhic lfits ~q ~ou~y 50 mile sep~ation if the S@OIIS are transmitting over adjacent ChtUllEk, lso tO zoo miles ~p~tion iftransmitting over the same channel. UHF charnel separation is typically one blank channel between active stations in the same geographic are% andfive blank channels between active stations in the same geographic area for VHF. Channel separation requirements are given in more detail in 47 CFR73.609; 47 CFR 73.610.
‘“A High-Tec~ High Stakes HDTV Gamble,” Editorial Research Reports, vol. 1, No. 7, 1989.~S&th Rqofi and &der, Television Mlocations, 41 FCC 167 (1952)+ The NTfA ~ be~ a ~jor project to reassess the tdlOCzUiOIl Of r~O
fkquency spectrum. Kathleen Killette, “New NTIA Chief Tackles Agend&” iUultichunnel News, Aug. 28, 1989, p. 34.26D~eN. ~&leldmdGene Ge~, ~~’rheOpp~~CoS& of spec~~o~t~ to Hi@ Deffition Televisio~” papapresent~atthe 16thanrNud
Arlie House Telecommunications Policy Research Conference, Arlie, VA, Oct. 30, 1988.
Chapter LMroduction and Summary ● 19
A sindcastpolicy would have a much greaterlong-term impact than augmentation, dependingon how it was implemented. The simulcastsignals would not have to be compatible withexisting NTSC standards and would require fewif any taboo channels; a new set of standardscould be adopted for them that would permitcloser spacing of the new stations’ channelassignments. As the penetration of HDTV setsincreased, the NTSC stations might be phasedout and the freed spectrum used for: 1) a nextgeneration of even higher quality video broad-casting technology; 2) additional new TV sta-tions; or 3) mobile communications or otherservices.
Cable TelWision
The possibly wider bandwidth of HDTVbroadcasts might require cable operators to usemore fiber-optic technology or perhaps leaseportions of the telephone companies’ fiber-opticnetworks. Cable operators may also considerusing DBS to supplement cable systems.
A system using a cable company’s coaxialcable and a telephone company’s twisted copperpairs might be able to provide a reasonable levelof interactive video services in the midterm (ch.3). Existing coaxial systems can provide HDTV-quality programming, particularly when strength-ened with a fiber optic backbone. Existingtwisted copper pairs of the telephone networkwill be able to provide a data rate sufficient formost information services-but not moderate-tohigh-definition real-time video-within theplanned upgrades to the N-ISDN level ofservice.
Direct Broadcast Satellites
Direct Broadcast Satellites use frequenciestoo high to be of practical use to earthboundbroadcasters; there is a relatively broad range offrequencies available; and satellite broadcastershave not yet developed strong vested interests ina particular allocation of these radio frequency
bands although competition for geosynchronousorbital space is keen. It is therefore easier toadjust the satellite transmission system to meetthe demands of HDTV. A partnership wasrecently formed in the United States to establisha DBS system by as early as 1993.27 Someanalysts believe that a DBS system could‘‘cherry-pick’ the most lucrative HDTV oppor-tunities and poses a formidable threat to cable-based or other delivery systems.28
Mobile Communic@”ons
The potential benefits of using additionalspectrum for HDTV broadcasting must bebalanced against the benefits of using portionsof the TV spectrum for other purposes such asmobile communications. Due to the physics ofradio wave propagation in the atmosphere, themost desirable frequencies for mobile commu-nications are the same as those now used for TV.When the current TV broadcasting system wasput into place and the channels allocated in the1940s and early 1950s, there was no competitionfor these frequencies from alternative uses; nordid alternative distribution systems for TV—such as cable or satellite+xist. Today, thereare many alternate means for distributing TV;the choices for mobile communications aremore limited.
HDTV could have an enormous impact onmobile communications such as cellular tele-phone, paging services, and related systems.More spectrum might be made available forthese services during the transition to HDTV ifsimulcast standards are used and charnels arerepacked. In the longer term, additional spec-trum might also become available if terrestrialbroadcasters were unable to provide as high-quality pictures as competitors-which couldcause their viewing audience to shift to altern-ative services and allow these frequencies to bereallocated to other uses.
Additional spectrum for land mobile servicescould help reduce future congestion on mobile
27Jolm Burgess, “Satellite Partnership Plans Pay-TV Systeu” Washington Post, Feb. 22, 1990, p. El.
2sDavid Rose@ “The Market for Broadband Services via Fiber to the Home,” Telemutics, vol. 6, No. 5, h%y 1989.
20 ● The Big picture: HDTV and High-Resolution Systems
frequencies, particularly during rush hours incities like Ims Angeles, New York, and Wash-ington, D.C. Spectrum and digital radio technol-ogies might allow economical and portablepersonal communications to be made availableto most individuals. With sufficient spectrum,cellular systems might someday compete effec-tively with local telephone companies wire-linesystems.zg Rates for telephone services mightthen be left to the competitive market the waythat cellular rates are now.
To develop awidely available personal communi-cations network, however, will require morespectrum than is currently available. Althoughthe United States is still a leader in mobilecommunications technologies, it could falterunless the industry can gain similar access tospectrum and, correspondingly, achieve similarmarket scales as its competitors in foreignmarkets.
Teikphone Companies
Twisted copper-pair will continue to be thepredominant medium in the local loop for thenext 10 to 20 years. With the transition toN-ISDN (ch. 3), the existing copper-pair net-work will be able to handle most informationneeds, including voice, data, and even somelow- to moderate-quality video. Over the mid-term, mixed telephone/copper pair and cableTV/coaxial cable networks might provide amedium level of interactive information serv-ices: cable could provide a high flow of informa-tion (hundreds of Mbps), including high-qualityvideo, from a central location to the home; andcopper pairs could transmit data at a rate of 1.5Mbps (two pair) from the home to any otherpoint desired through the switched network ofthe public phone system.
Fiber is likely to be the medium of choicethroughout the cable and telephone networks inthe longer term. Inserting fiber in the cablebackbone as a first step significantly improves
cable capacity and performance for a relativelysmall investment. In contrast, although tele-phone companies can replace their backbonesfor roughly the same cost as cable companies,the copper pairs in the local telephone loop to thehome do not have sufficient capacity to delivera high-definition signal.
Cable systems camot easily be adapted toprovide high-capacity switched two-way com-munications such as the public phone systemdoes. Regulatory and financial structures mayhamper a move to such a system. Xl An impor~ntquestion confronting policymakers is whether amixed cable/telephone network is an importantintermediate step toward a national two-waybroadband network or an evolutionary dead-end. If the former, changes in the regulatoryenvironment to aid this transition would benecessary; if the latter, means of encouraging adirect transition to a national broadband networkmust be considered.
Market Share
Terrestrial broadcasters (as do all spectrumusers) have limited spectrum available to them;in turn, this limits the quality of the pictures theyare able to deliver to viewers. If terrestrialbroadcasters cannot deliver pictures of as highquality as cable or DBS broadcasters, they maylose market share. Because of this, many U.S.terrestrial broadcasters want a single uniformtransmission standard applied to all broadcastmedia—terrestrial, cable, DBS, VCRs—to limitall media to the same technical capability asterrestrial broadcasters.
Conversely, many of the competing medialook to HDTV as an opportunity for them tocapture market share from terrestrial broadcast-ers by use of their greater technical potentiaI andflexibility to transmit high-quality HDTV to theconsumer. Terrestrial broadcasters have little togain and a lot to lose in such a contest: theycurrently have between 54 and 59 percent of the
z~~leld and AX, op. cit., footnote 26.
3~mceL. Egmand Do~g~ A. Com ‘ ‘capi@BudgefigAlte~tiv~ ForResi&nti~ Broadband Nemorh,” Center for Telecommunications adInformation Studies, Columbia University, 1989.
Chapter LMroduction and Summary ● 21
television audience, and TV households watchprograms more than 7 hours/day .31
There is more at stake in the contest formarket share among the various media thansimply profits: it could overturn the traditionalroles that these media have played in serving thepublic. Broadcast TV is a central cultural focusin American life, providing a shared experienceand information for all. Over-the-air broadcast-ers are the only ones to be required by a statutoryobligation to serve local audiences—providingnews, local election coverage, public amounce-ments, and community affairs. Similarly, publicbroadcasting stations are major providers ofeducational programming. Because of theseroles and the lack of similar regulatory demandson alternative media, there is concern that ifbroadcasters are unable to provide the samequality service as alternative media, these serv-ices could be lost. While market forces willensure the provision of acceptable TV program-ming, Congress may need to take steps to ensurecomparable services— news, community affairs,etc.—will be fully and equitably provided to thepublic by each medium.
costs #
The costs of upgrading to HDTV for programproducers, broadcasters, and consumers couldbe substantial, depending on the standard cho-sen. Program producers will need to convertexisting studio equipment to video HDTVproduction equipment, but this may reduceproduction costs by eliminating delays in proc-essing film and by facilitating editing andincorporation of special effects in the productionprocess.
Estimates of the cost of upgrading terrestrialbroadcaster’s NTSC equipment to HDTV capa-
bility range from $7.4 million to $40 million.32A simulcast system might cost less than oneusing augmentation channels, because much ofthe existing transmission equipment could beused (except for the digital coding, and it wouldnot necessarily require new wide-band equip-ment) and the transmission power requirementswould be much less.33
Costs for cable companies may be less. TheFCC Advisory Committee has estimated that theextra cost of introducing 12 channels of HDTVprogramming on a sample cable system serving100,000 subscribers would be about $1.9 mil-lion.34
Consumers, too, may find HDTV receiversmore expensive than the set they use now. Byupgrading to HDTV at the time they wouldnormally replace their old set, these costs couldbe blunted somewhat.
Finally, if the HDTV system is more spec-trum efficient, these costs must be weighedagainst the benefits of freeing valuable spectrumfor other uses. Consumers camot make thistrade-off, however; policymakers must.
The chicken-and-egg problem of who investsfirst might only be resolved by all acting inconcert. Color TV was successfully introducedonly through the patience and enormous invest-ment—some $3 billion in 1988 dollars-ofRCA.35 A similar risk will have to be borne tolaunch HDTV. On the other hand, UHF broad-casting was made possible by governmentaction through the “All Channel ReceiverLaw” which requires all TV sets sold in the
SIA.C. Nie~on CO,, personal communication October 1989.
s2FcCA’TVAdvisory Committee, Systems Subcommittee, Working Party 3, Terrestrial Sp=idist Group 1, “InterimRepofi” table 1, Apr. 10,1989.The Boston Consulting Group, “Development of a U.S. Based Am Industry,’ Preliminary Report prepared for the American Electronics Association1989, p. 22.
ssHowmd ~fler, “me Simulcmt Strategy for HDTV,” PBS, 1989.
%tFCC Am AdvisoV Cotifiee, op. cit., f~tnote 32, a~c~ent E, ~ble 2; ~d Boston Consdfig Group, op. cit., fOO&lOte 32.
35u.s. Conwess, House su~o~tt= onTcl&omm~cations ad F~ce, (Jo~ttee on Energy and Commercq c~nsorfia andthe Development
of High Dqiinition Systems, testimony presented by Barry Whale@ Sept. 13, 1988.
22 ● T~ Big Picwe: HDTVand High-Resolution Systems
United States to have both VHF and UHFtuners.3G This prevented manufacturers fromcutting costs by eliminating the UHF tuner—thus raising costs to’ consumers slightly-butover time allowed the development of a suffi-cient market so that UHF broadcasting could besuccessfully launched.
THE STANDARDS-SETTINGPROCESS
Standards can reduce or prevent confusion inthe marketplace. Standards allow manufacturersto increase production efficiency by producingin large volumes for a uniform market; they canstimulate competition; and they reassure con-sumers that whichever brand of HDTV they buyor wherever they use it in the United States, itwill be able to receive and display the localbroadcasts.
These lessons have been hard learned. Afterunsuccessful attempts to set AM stereo radiostandards, for example, the FCC left the deci-sion to the ‘‘marketplace’ in 1982.37 Severalincompatible systems then began to be adopted;this increased consumer, broadcaster, and man-ufacturer confusion. As a result, AM stereobroadcasting is growing very slowly, and AMradio generally is losing market share to FMradio.
Standards sometimes have drawbacks aswell. When a technology is rapidly changing,standards can lock in an obsolete technology;standards can limit choice; and if poorly de-signed yet widely used, standards can slow
innovation. 38 A good example might be theQWERTY typewriter. Designed in the late1800s, the QWERTY layout was intended tolimit typing speeds; the mechanical systemsthen available could not otherwise keep up witha fast typist and tended to jam. Since then,however, this keyboard has proven impossibleto dislodge despite its widely acknowledgedshortcomings.
Standards have also been used to promotenational political and/or commercial interests.Rather than using the U.S. NTSC system,France developed and adopted its own color TVstandard, SECAM, in the early 1960s in order toprotect its color TV industry during the develop-mental stages. By 1976, the French TV industryaccounted for roughly 0.5 percent of the FrenchGNP.39 Similarly, patents on the German colorTV system, PAL, were used to help excludenon-European manufacturers from the Euro-pean market.a
In the United States, technical standards fordomestic communication technologies are cur-rently handled almost exclusively by the FCC.Not only has Congress granted the agencyvirtually sole jurisdiction over broadcastingstandards,41 but the courts have also recognizedthat its legislative authority permits the agencyto preempt conflicting State or local regulationsof technical standards in telephone42 or cabletelevision.43 Where the courts have found thatnational uniformity is important to foster inter-state commerce, they have prevented Statesfrom establishing differing standards.
3647 CFR Section 15.65.
37staIIlq M. Bc~n@d Leland L. Johnsoq Compatibility Standards, Competition, andInnovation in the Broadcasting 1ndusQ (SW@ Moni~ CA:Rand Corp., November 1986). Note that the FCC faced seveml difficulties in setting this standard, including incomplete information on theperfonnaneeof different systems, conflicting test data or data that was gathered under differing conditions, and fierce opposition from the f- that would have losthad the tentative FCC findings been formalized.
sgDavid WC~ Librq of Congress, Congressional Research semi% “Telecommunications and Information Systems Standardization-Is AmericaReady,” 87=$58 SPR, May 21, 1987.
s~on~ J. ~ane, The politics of International Stan&rds: France and the Color TV War (Norwood, NJ: Ablex fiblishing COW., 1979).
~rade policy may have played a more important role in keeping imports out.
dlsee, e.g., &g~iar~o v. United States, 366 F.2d 720, 723 (9th Cir. 1966).dz~orth Carolina u~”litie$com-s~ion v,FC’, 537 F2d 787 (4th Cir. 1976) upholding theFcc’s s~~ds forcustornaprankes Ct@p3Jl@ (CPE).
dsciO of N@ York v. FCC, 108 S.Ct. 163’7 (1988).
Chapter I--introduction and Summary ● 23
The manner in which the FCC administersthis power can have a significant impact onmany areas of telecommunications; thus thenature of the ATV standards-setting processcould strongly influence who benefits and wholoses from those decisions. Four aspects of theprocess appear to have particularly importantimpacts within the United States itself: 1) howmuch discretion is delegated to the marketplace;2) how fast the process is pushed; 3) whether allserious proposals are fully considered; and ~)whether the process permits the selection of astandard that combines aspects of differentproposals. An important related issue is U.S.participation in international standards fora; thiscomplex issue will be discussed elsewhere.
Standards Setting and MarketplaceParticipants
Ideally, the participants in a market wouldreach a consensus on the best standards to adoptfor a particular product, that maximizes theirprofits at the same time that they maximize theattractiveness of the product to the public.Private sector firms have the technical expertiseand the best knowledge of the markets; and theyare financially accountable for their errors. Themarketplace, however, provides many incen-tives for firms to establish standards unrelated tosocial benefit.
The FCC Advisory Committee on AdvancedTelevision Services provides a mechanism forthe direct input of private firm expertise. Thisadvisory committee is voluntary and open to allwho wish to participate. Most of those whocurrently participate represent manufacturing ormedia interests—they can’t afford not to partici-pate. In contrast, representatives of labor andconsumer interests face difficulties in participat-ing.
Private sector firms directly influence thestandards-setting process by providing regularand extensive technical staff participation to thecommittee; by conducting independent and/orsupporting studies; and by widely distributingtechnical documentation. These are potentiallyvaluable inputs and are incorporated in FCC
30-368 - 90 - 2 : QL 3
decisionmaking. Yet there is also the potentialfor abuse. Large firms may be able to fund morestaff participation and technical inputs thanthose with limited resources. This can bias theprocess.
Foreign TV manufacturers, for example, maybe able to channel much greater resources intothe standards-setting process than U.S. TVmanufacturers, since they now dominate theU.S. television market. The standards promotedby foreign manufacturers will not necessarilyrepresent U.S. national interests in developingdomestic communications infrastructure; thesestandards almost certainly do not represent U.S.national interests in encouraging additional U.S.firms to reenter ATV manufacturing. Foreigngovernments often do not allow reciprocalaccess by U.S. firms to their standards-settingprocesses for similar reasons.
IJ.S. broadcasters might oppose standardsthat would make it easier for the FCC toreallocate portions of terrestrial broadcastingspectrum to other purposes in the future, regard-less of the long-term interests of the public inmobile communications or other services.
Finally, if market participants are unable toreach a consensus on a single choice, there isdanger that multiple and incompatible standardscould result. This could raise uncertainty amongmanufacturers and consumers and hinder theintroduction of ATV.
The Timetable for Establishing Standards
In responding to the array of issues raised byHDTV there are two conflicting time pressures:1) taking the necessary time to establish aflexible standard that can support the rapidlychanging technologies and needs for the nextseveral decades; and 2) setting standards quicklyenough that the U.S. market can grow in parallelwith those in Europe and Japan, thus providingU.S. producers similar opportunities for realiz-ing economies of scale and learning in produc-tion.
If the United States acts precipitously andestablishes standards prematurely, the ATV
24 ● The Big Picture: HDT’V arid High-Resolution Systems
equipment produced under those standards mightquickly become obsolete, the Nation might haveto endure with inferior technology, or haveexcessive difficulties in making the transition tofuture generations of equipment.
Alternatively, if potential U.S. entrants waitfor a national fiber system, for example, beforeentering the ATV market, they could wait 20years or more. During that time, foreign compe-titors would have the opportunity to furtherstrengthen their manufacturing capabilities anddistribution systems, and would receive enormousrevenues for conducting R&D into new technol-ogies. If additional U.S.-owned or U.S.-basedfirms are to enter these markets, there is nosubstitute for getting in quickly and gainingintense day-to-day manufacturing experience.
Providing Full Opportunities for AllSerious Proponents
Without some minimum threshold for thoseseeking to establish ATV standards, the FCCcould be subject to numerous quack proposalssubmitted in the misguided hope of winning astandards ‘‘lottery. ” Indeed, at least one of theproposals submitted on paper was believed to‘‘challenge the known laws of informationtheory. ‘“ The FCC Advisory Committee cur-rently requires standards proponents to submit afully developed set of broadcasting and recep-tion hardware to the Advanced TV Test Center(A~C) for testing; 45 and the A~C in ‘m
requires the proponent to post a $200,000 bondto hold a time slot for testing their system (thisbond can be waived). On its face, this makessense. The FCC should not spend governmentmoney to develop and test a private group’ssystem, particularly if all the royalties go to thatprivate group.
On the other hand, designing and building acomplete set of hardware can cost severalmillion dollars; buying test equipment to debugthe hardware before presenting it to the A~Ccan cost millions more. Even fairly large firms,such as Zenith, are straining to find the man-power and financial resources to meet thesedemands. It is not surprising, then, that otherU.S. proponents, who are generally small,entrepreneurial or university-based efforts, arehaving even greater problems. Of the more than20 standards proposals submitted to the FCC,only 5 or 6 appear likely to be developed intohardware at this time due in part to the lack offinancial resources. Several of these are ofEDTV-rather than of HDTV-quality and all butone or two are sponsored by foreign-basedmanufacturers. Even the FCC Advisory Com-mittee has obliquely noted this problem.ti
It is not entirely clear why the U.S. capitalmarkets have failed to provide the necessaryfinancing, but they have not. If financial supportis not forthcoming there is the risk that aforeign-owned standard will be selected andpotentially hundreds of millions of dollars inlicense royalties will flow to that entity, despitethe possibility of a superior U.S.-owned systemthat could not be considered for lack of a fewmillion dollars in timely funding.
This early focus on hardware may also becounterproductive for other reasons. Today,complex systems are always simulated on acomputer before they are produced. For HDTVthis is particularly important because many ofthe improvements in hardware that can beexpected in coming years should be assessed,but the technologies themselves are not and willnot be available for years. For example, it maybe useful to develop a broader set of standardsthat allows for the gradual upgrading of ATV to
44XC Advisory Chmnittee On AW, Systems Subcommittee, “~terim Repo@ “ Apr. 10, 1989, SS/WPI Interim Progress Report, Feb. 12, 1989.ds~c~d E. Wiley, ~<smnd ~te~ Repofl of tie KC Advisory Committee on Advanced Television Service,” Apr. 26, 1989.
@’FCC Advisory Committee on Advanced Television Service,’ Systems Subcommittee,“InterimRepO%” Apr. 10, 1989; “AnAssessment of theATV Systems and Technologies Presented at the Nov. 14-18 Meeting of SWWPI,” p. 3. “MIThas provided outstanding technical inforrnationrangingffom technical papers on psychophysical aspects of television to ATV system proposals. At the “marathon” meeting MIT presented very interestingresults from computer simulation studies on A’IW transmission in low quality analog channels . . . . l?mtl MIT. . . k dectied to submit mw~e fortesting.”
Chapter 14ntroduction and flummary ● 25
a fiber system, or that can handle the differingrequirements of commercial or industrial users.
The standards process is already fallingbehind schedule. This is due, in part, to the rapidevolution of the technology and the difficulty ofdeveloping a broadcast standard for HDTV—which many groups are beginning to more fullyappreciate. Ultimately, of course, any proposedstandard must be proven in full-scale tests withreal hardware before it can be formally adopted.Full-scale manufacturing and marketing willrequire a firm with enormous financial resourcesand skilled manpower.
Hybrid Standards
The current requirement that proponents pro-vide their own hardware combined with theFCC’s limited technical and financial resourcesfor designing and testing ATV systems mayhinder the synthesis of a standard ffom the bestfeatures of several proposals. Proponents ofparticular standards might also object to such asynthesis, gambling instead for their proposal tobe chosen exclusively.
In Europe and Japan the governments havemaintained extensive staff experience throughtheir national telecommunications services andbroadcasting organizations. Raising the level oftechnical manpower and financial resourceswithin the FCC and possibly utilizing theexpertise within universities and National Labo-
ratories, including the U.S. Department ofCommerce’s National Institute of Science andTechnology (NIST), might enable the gover-nment to play a greater role in protecting U.S.national interests and reduce U.S. reliance onforeign industry groups with potentially con-flicting agendas.
Alternatively, a small elite group of industryand university technical experts might beformed and funded to work together, or inparallel, to synthesize the best possible standardfrom the numerous competing proposals. Acorresponding patent pool might be formed withappropriate safeguards for the interests of theindividuals and companies involved and to payback the government its investment from royal-ties. Some portion of these patent pool royaltiesmight also be used to fund R&D in videoentertainment technologies, ranging from cam-era to broadcasting to receiver display technolo-gies.
Royalties to RCA similarly supported asignificant fraction of the R&D in consumerelectronics done in the United States. RCA wasestablished by GE, Westinghouse, and AT&Tin1919 at the request of high U.S. Governmentofficials (including the Acting Secretary of theNavy, Franklin D. Roosevelt) who wished toprevent foreign domination of the growingtransatlantic communications services.47
ATUnde~cableswere a,lreadyundesforeign (if friendly) control. Incontras4 the ‘wireless” stations of British owned American Marcoti Compmyhad been held by the U.S. Government during WWI for wartime purposes. Rather than return these stations to foreign control, RCA was formed andthese stations were transferred to it in 1920. RCA Corp. “RCA: Unhistorical Perspective,” 1978. For a discussion of the development and fate of RCASee: “The U.S. Television Set Market, Prewar to 1970”; Donald ChristianseU “A Stirring Gian4° IEEE Spectrum, February 1986; NhoraCortes-Comerer, “A Venerable Giant Sharpens its Claws,” IEEE Spectrum, February 1986; Herb Brody, “Picking Up The Pieces of RCA,” HighTechnology Business, May 1988.
Chapter 2
The Historical Development of HDTV
INTRODUCTIONJapan is currently the world leader in HDTV
development. Sony began selling HDTV studioproduction equipment in 1984 and HDTV receiversare now being sold commercially. Regular satellitebroadcasts in Japan began in June 1989 and, by1991, NHK (the state-owned Japan BroadcastingCorporation) plans to broadcast 6 to 7 hours ofHDTV prograrnming daily.
Driven by the threat of Japanese domination ofEuropean consumer electronics markets, industry in19 European countries initiated Eureka95 in 1986-a collaborative crash HDTV development effort—that has made rapid progress in developing aEuropean HDTV system. They have since demon-strated numerous pieces of production, transmis-sion, and reception equipment and they plan to beginExtended Definition ED-MAC broadcasts in 1991.
U.S. electronics manufacturers have a signtilcantset of handicaps as they move to meet the competi-tion. U.S. broadcast standards are not scheduled tobe established until 1992. U. S.-manufactured HDTVconsumer equipment probably could not be com-mercially available until 1993 or 1994. U.S. manu-facturers are running a distant third in a three-wayrace.
This chapter addresses the history of HDTVdevelopment. By far the most detail is provided on
Japan’s efforts because of theti magnitude andcomplexity. The HDTV work in the United Statesand Europe will be briefly examined here.
JAPAN1NHK researchers began investigating HDTV
following the 1964 TolgTo Olympics. A formalresearch program was begun in 1970. In 1972, NHKproposed an HDTV research program to the CCIR(the ConsultativeCommitteeon InternationalRadio).2Development of techniques for program productionand to make broadcasting practical followed. In1979, NHK began experimental transmission tests.3
Sony began marketing HDTV studio productionequipment in 1984 and introduced a second genera-tion of HDTV studio equipment in 1989.4 NHKexperimentally broadcast HDTV pictures of theSeoul Olympic Games in 1988 via optical fibers andsatellite, and began daily l-hour experimentalHDTVbroadcasts in June 1989 using satellite BS-2. Lim-ited HDTV “HiVision’ satellite broadcasts arescheduled to begin in 1990 using tie BS-3Asatellite; regular broadcasts will begin in 1991 withBS-3B.5
At the same time, Japanese TV manufacturershave developed Improved Definition TVs (IDTV)which use digital electronics to improve the pictureattainable with current NTSC broadcasts. In August1989, private Japanese Broadcasters began terres-
l~e p~cip~ somes for this section me: James G. Parker and E. Jan VardamaQ “HDTV Developments in Japau” T~hs~ch~t~tiO~, ‘C”*Austir.L TX, 1989; MarkEatoq MCC, personal communications, Aust@ w May 11,1989, Oct. 12,17, and 18, 1989; and Pamela Doughmanand MarkEato~ “HDS Projects and Developments in Europe and Japan,”unpublished manuscrip~ MCC, Aust@ ~ August 1988; See also: Takao Shirnizu,“High Defiition Television: Comparison of Research and Development Strategies,” M.SC Thesis, Sloan School, Massachusetts Institute ofTechnology, May 1989; Chalmers Johnson, “MITI, ~ and the Telecom Wars: How Japan Makes Policy for High Technology,” in Politics andProductivity: The Real Story of Why Japan Works, Chalmers Johnson, Laura D’Andrea Tysoq JohrI Zysman (Cambridge, MA: Ba.llinger Publishing,1989); Robert B. Coheu “A Profile of Japan’s HDTV Industry, Its Links to Other Sectors, and Policies Adopted to Promote HDTV Development”for the JTEC Panel Repo@ National Academy of Sciences, draft.
2~e cc~ is an orga~tion ~volved with ~temtio~ telecoH@cation5 s~n~ds se~g. It i s ~sociated with t h e h.WI’MtiOndTelecommunications Union and chartered under the United Nations. The ITU was founded in 1865 and has been a United Nations organization since1947. Currently 160 members strong and headquartered in Genev% Switzerland, it is the supreme international organiza tion dealing with issues oftelecommunications and standards setting. For more information see: George A. Codding, The International Telecommunications Union in a ChangingWorld (Dedham, MA: Artech House, 1982).
sKe~ethR. Donow, “HDTV: ~arming for ActiorL” National Association of Broadcasters, WashingtorL DC, April 1988. Note that these early testswerewideband-Birney Dayto@ NVisiou personal communication Oct. 12, 1989; William Schreiber, Massachusetts Institute of Technology, personalcommunication Oct. 12, 1989.
A“Super Television “ Business Week, Jan. 30, 1989; “Development of MUSE Family Systems,” NHK, Nov. 15, 1988, mimeo; David Sanger,“Japanese Test Illustrates Big had in TV of Future,” New York Times, Mar. 21, 1989; and, Larry Thorpe, Sony COT., personal communication Oct.12, 1989,
sp~ker and v~m~ op. cit., footnote 1; Eato~ OP. cit.. foo~ote 1.
–27–
28 ● The Big picture: HDTVand High-Resolution Systems
trial broadcasts of an NTSC compatible ExtendedDeftition TV (EDTV) System known as “ClearVi-sion” in the Tokyo and Osaka area. Later phases ofthis project will provide a wider screen picture andadd digital soundtracks.b
The Japanese may not have begun their quest forHDTV in the 1960s with anything more in mind thanhigh-quality pictures and the prospects offered bylarge consumer electronics markets. Beginning in1983 with the emergence of the “teletopia”7 con-cept, HDTV became something more than that-anintegral part of their information society of thefuture. The Japanese envision not a service economybut an information economy with more and highervalue-added manufacturing than today. By the year2000, the Japanese expect fidly one-third of theirindustrial investment to be in manufacturing relatedto information technologies and products.8
In developing such an economy, several princi-ples are shaping their policies, including: “technol-ogy fusion, ”9 or the potential of a technology toimpact a broad range of other technologies andindustries; sharing risk between companies and thegovernment; the promotion of consortia designed tostimulate competitiveness; and the importance ofdeveloping mutually supporting markets.
Within this framework, the Japanese Governmentis providing an intricate web of direct and indirectsupports for the development of HDTV. In addition,the government has worked with industry to estab-lish standards and has aggressively promoted thesestandards and Japanese commercial interests world-wide.
Research and Development
The Japanese Government performed the initialhigh-risk R&D for an HDTV system under theumbrella of the state-owned NHK. Once the majorsystem parameters were developed, NHK ‘ ‘encour-aged” equipment suppliers to participate in the
development of the system components. Companieswere assigned spectlc research areas and sometimeseven specific production tasks, the results of whichwere shared with NHK and the other pmticipatingcompanies for at most a nominal licensing fee.
NHK divided the development of specializedintegrated circuits (ASICS) for the MUSE HDTVreceiver, for example, among six different compa-nies to avoid the duplication of effort and cost if eachfm had to design a complete set of ASICS itself.Thus, Toshiba developed motion compensationchips; NEC developed color signal processing chips;Matsushita developed audio processing chips; ands. On.lo s~wly, MUSE to NTSC convefler cfiPset development was divided between Sanyo, Mit-subishi, and Matsushita. These designs were thenshared by all participants.
Numerous fms are involved in other joint (orperhaps more accurately described as partitioned)R&D efforts on HDTV-related technologies that arecoordinated by NHK, MITI, MPT, or the KeyTechnology Center. 11 These include the develop-ment of: camera sensors (CCDS) and related cameraequipment; CRT, projection, and LCD displays;analog and digital HD-VCRS; optical recordingtechnologies; bandwidth compression equipment;and graphics software. Examples of the HDTV-related R&D and product development of variousJapanese companies are shown in table 2-1.
Some types of production are being partitioned aswell. For example, to minimize duplication andensure economies of scale, NHX assigned the highlycapital-intensive development and production of a35-inch color picture tube for HDTV to Mitsubishi.In turn Mitsubishi agreed to provide competitorspicture tubes off the same production line on anequally shared basis. They continued to shareproduction even when the capacity proved to beinsufficient for the unexpectedly large market.12
GBob whis~ BIS Mackintos4 personal communicatio~ Mar. 29, 1989.7Note tit “teletopi=” we one set of approaches to local and regional information system development effOfis. There ~e ~nY other approaches.
More broadly, one should speak of these many efforts under the rubric of “information society.’gp~ker and Va@~ op. cit., footnote 1.
9F3atou op. cit., footnote 1.l~~hseach ~temtio~, kc., ~SE Decoder B1ock Die fiorn Nikkei Elec@ofics, Aug. ‘7, 1989, pp.126-127; md “Report Of the JTEC
Evaluation Team on HDTV Developments in Jap~” National Science Foundatio~ July 26, 1989.1 Ipmka md Vmda op. cit., footnote 1.
lzHow~d Wer, PBS, personal communicatio~ IWiy & 1989; Oct. 12, 19*9.
Chapter 2—The Historical Development of HDTV ● 29
Table 2-l—Selected HDTV Technology Development Efforts by Japanese Companies
Displays Total corporateCameras Recording Transmitter/receiver Projection sales (est. 1989)
Firm Tube CCD Optical Digital Analog Encoder/demder CRT CRT SS FPD ($ billions)
Canon . . . . . . . - -Fujitsu . . . . . . . - RHitachi . . . . . . R -Ikegami . . . . . . C RMatsushita . . . R RMitsubishi . . . . - -NEC . . . . . . . . - RPioneer . . . . . . - -Sanyo . . . . . . . R RSharp . . . . . . . R RSony . . . . . . . . C RToshiba . . . . . . - RJVC . . . . . . . . . R R
R
R
RRR
R
RR
RcR
RRRRR
cR
c
R
R
c
cRc
R
:RcRRcR
C R - RRC-Rc c - -C C - RC C - RCCRRC R - -CCRRCCRRC C - RC C - RC R - -
$ 8 . 917.846.0
0.440.019.222.4
1.510.710.015.529.2
5.9NOTES: R—research; ~ommercializing. Camera tubes are similar to those used today; CCDS (charg~oupled devices) are a type of solid-state sensing
element. Optical recording systems are typically optical disks; digital recording systems are usually for studio use; and analog reeording systems areprimarily for home use, although some using l-inch tape are for use in the studio. CRT (cathode ray tube) displays area large, highquality versionof the picture tubes used at home today. SS (solid-state) projection displays are typically either LCDS (liquid crystal displays-like those used on manylaptop computers) or deformable membranes from which light is refleeted. FPD (flat panel displays) are usually LCDS but plasma display panels andother types of thin displays (ch. 5) are also under development by some firms. Whether or not the technology is under research or beingcommercialized is a subjective judgment because the status of these technologies is changing rapidly and because these same firms are oftenmarketing similar technologies in less demanding applications today.
SOURCE: James G. Parker and E. Jan Vardaman, “HDTV Developments in Japan,” TechSearch International, Inc., Austin, TX, 1989, citing Nikkei Electronics,Oct. 3,1988, p.1 14; and Mark Eaton, personal communications, MCC, Austin, TX. Please note that many other developments, such as equipmentfor studio production or satellite transmission, or subcomponents of some of the above listed systems, are not included here. Development effortsby Fujitsu General are shown as part of Fujitsu Ltd. even though it is only 33 percent owned by Fujitsu Ltd.
Through this mechanism, the competitive effortsof the participating firms are being redirected fromearly, high-risk basic system design to secondaryfeatures such as overall product quality, ease ofoperation, and manufacturing cost.13
The development of large area LCD displayspresents a particularly interesting example of thenumerous linkages-’ ’technology fusion’ ’-that isin part guiding Japanese policy. The most ambitiousLCD effort, known as the Giant Electronics Project,is planned by MITI to be a 7-year approximately$100 million project to develop a 40-inch diagonalflat panel display by 1996.
MITI foresees a variety of technological linkagesand potential spinoffs to result from this effort.Participants will have to: produce an alkali-freeglass substrate with less than O.001-inch (20-microns) variation in thickness over the entire
40-inch diagonal area; deposit a high-precisionthin-fti on this substrate; develop the manufactur-ing skills to etch circuitry into this fti to a precisionof roughly O.0001-inch (3 microns) over this entirearea; develop precision techniques for automaticallyattaching leads and assembling the display; andinvent new technologies to test it. It is relatively easyto achieve such precision in the confines of today’stiny integrated circuits, but doing so across largeareas poses formidable technological challenges.14
MITI expects these capabilities to be applied insuch diverse products as: ultra-high-density opticalrecording systems; ultra-thin photocopying systems;solar cells; optical engraving; large flat-light sources;high-precision electronic components; and “chip-on-glass’ electronic packaging and assembly. The‘‘chip-on-glass’ packaging and assembly technolo-gies in particular offer the potential for significant
13u.s. Conpess, House Committe onlhergyand Commerce, Subcommittee on Telecommunications and Finance, Report of the PublicBroadcastingService and Comments of the Association of Maximum Service Telecasters, Inc. on High-definition Television, Hearings March 1989, Print No. 101-E;This approach has been used quite generally in Japa~ and in several of the exceptions where basic system design was done independently at difkrentcompanies; there have been serious conflicts and losses as a result. One of the best known such cases was the Sony development of Betarnax and theMatsushita development of VHS Video Cassette Recorders. Although many believe that the Betarnax was technically superior, the marketing muscleof Matsushita successfully made VHS the industry standard.
14WsW Ofktemtioti Trade ~d ~dusq, “Development of Basic Technology for Large Surface Circuit Elements, ’ Tokyo, September 1988;English translationprovided by Parker and Vardama~ op. cit., footuote 1; Gary Stix, ‘ ‘Manufacturing Hurdles Challenge Large-LCD Developers,’ IEEESpectrum, September 1989.
30 ● The Big picture: HDTV and High-Resolution Systems
improvements in the production and the perform-ance of advanced electronic systems (ch. 5).15
Since the mid-1960s, the total investment of NHKin HDTV R&D and related activities has been about$150 million. The projects funded through MITI,MPT, and the Key Technology Center, as well asinvestments by private fm greatly increases thetotal R&D expenditure. Including the costs ofsetting up production lines, private sector invest-ment is estimated to range horn $700 million to asmuch as $1.3 billion. Japanese Government fina-ncial policies have provided an enormous pool oflow-cost capital that greatly aids such investmentsby individual fins.
The Japanese Government is supporting muchmore than just HDTV. They are, for example,funding both R&D and construction of networking,database, and other information infh.structure atlevels 10 to 100 times greater than their support ofHDTV.lG Some argue that this indicates the greaterimportance of these activities; others argue that thetechnology and manufacturing infrastructure forproducing the terminals (HDTV and related technol-ogies) for these information networks is now in-handand that the Japanese Government has simplymoved on to the next step in developing theirinformation society.
Financial Supports
A variety of direct and indirect governmentfinancial mechanisms are supporting the devel-opment of HDTV in Japan. NHK performed much ofthe initial research on HDTV. Funding for NHK isfrom a mandatory household color TV subscriptionfee of Y107O ($8)17 per month from televisionhouseholds. In August 1989, this fee was increasedto Y2000 ($14) for homes with satellite broadcastreception.
The Key Technology Center (KTC) was foundedby MITI and MPT in 1985. It is funded in part bydividends horn government-owned shares of NITand Japan Tobacco, Inc., and direct contributionshorn government financial institutions and theprivate sector. 18 By fi&1989, the KTC had ‘“P-
ported some 141 company projects and 75 consortia,of which several are HDTV-related.19 Loans of up to70 percent of the total outlay are available with nopayments due on the principal or interest for up to 5years. The principal is then repaid within 10 years ofcompletion of the project. There is a ceiling on theinterest of about 5 percent which is paid tithe projectis successful, but is waived entirely if the R&Dproject fails. An advisory board to the KTC passesjudgment on the degree of success of a project andthe amount of interest to be repaid.m
A web of financial supports is also being providedfor HiVision development and promotion efforts. Inthe MITI HiVision Communities program, specialloans are available from the Japan DevelopmentBank and the Hokkaido Tohoku Development Fundfor HiVision equipment, software, or promotion.The loans will cover up to 40 percent of the costs at5 percent interest with up to 3-year grace periods and15-year repayment times. No-interest loans forfacility construction are available with similar graceperiods and repayment times.21
Small and medium-sized businesses are includedin the MITI plan. Performing R&D, initiating abusiness, or promoting a market related to HiVision,or purchasing HiVision-related products can makeone eligible to receive a 15-year loan of up to Y1OOmillion ($700,000) at 3.5 percent interest the frost 3years and 5.4 percent after. Even more favorableloan rates are available to small businesses insmaller towns and rural areas.22
15MITI, “Development of Basic TechrIoIogy for Large Surface circuit Elements, ” Op. Cit., fOOtnOte 14.
16Cohe~ op. cit., fw~ote 1; DemM Kogyo Geppo, Mar. 1989, pp. 2-11; Kenneth Flamm, Brookings hti~tio~ Persod comm~catiom; ‘t” 13and Oct. 25, 1989.
17Bm~ on ylLI&$I.00 and m~ded off.
lsp~ti and Vardm op. cit., footnote 1.ls~kfitoq MCC, pmso~ comm~catio% &t. 17, 1989. Note tit&e KTC fwwtions much like an investment bank. It creates new COlllpallieS
in the form of consortia and provides up to 70 percent of the capital. If the effort of the consotia looks like it will be successm the companies involvedwill then fund anincreasingportion of the work until, after 5-7 years, it is a stand-alone company. Other activities of the KTC include: lending to projectswithin companies, mediating (especially technology transfer) between the mtional laboratories and private companies; supporting contract research;creating and promoting databases; and handling trust funds to sponsor foreign researchers in Japanese labs.
20p~m and Viid- op. cit., footnote 1.
zl~ “HiVision Application Guide, HiVision Community Concept,” July 31, 1989, p. 31; Translation provided by Eato% op. cit., footnote 1.2z~ op. cit., fOOtnOte 21-
Chapter 2—The Historical Development of HDTV ● 31
Potentially significant tax benefits are provided inthe MITI plan as well. For example, expenditures onHiVision promotion in MITI HiVision Communi-ties can be taken as a loss for tax purposes.23 Taxadvantages can occur in a wide range of othersettings. For example, small and medium-sizedbusinesses that buy or lease satellite communicationfacilities can receive accelerated depreciation orother tax considerations.x
The government has also directly reduced risk tofms by acting as a holding company-purchasingexpensive equipment and then leasing it back tofirms. The MPT is now pl arming to establish a $740million holding company for the BS-4 satellite dueto be launched in 1997. Companies will then be ableto lease HDTV channels on that satellite accordingto market conditions.
Using satellites for HDTV is an example of theJapanese effort to develop mutually supportingmarkets. Although the satellite systems for HDTVcould have been purchased from the United States,the Japanese Government has instead pursued aprogram of developing a domestic space industry.This effort began with the February 1983 launch ofJapan’s frost commercial communications satellite,which cost three times more than similar or bettersystems that could have been purchased from theUnited States.25 Japanese efforts to create a domesticHDTV industry are thus channeled into supportingthe domestic space industry as well.
Fin~y, ~WS will not be purchased if HDTV-compatible programrning is not available. Thus,both MPT and MITI have established leasingcompanies. Nippon HiVision was formed in April1989 by the Mm, NHK, and 40 private companiesto purchase and then lease HDTV equipment tobroadcasters. MITI is establishing a leasing com-pany with about 30 private companies to leaseHDTV equipment and software to movie producers,electronic publishers, and non-profit groups such as
schools and museums. These leasing companiesreduce risks for both manufacturers and users ofHDTV equipment—by guaranteeing sales for pro-ducers, providing equipment to users at low cost,and ensuring that movies and TV programs ofHDTV quality are produced for viewers to watch.Sony’s recent purchase of Columbia Pictures for$5.6 billion is seen by some as a move to ensure thatthere will be programrnin g available for Sonymanufactured HDTVS.26
Market Promotion
R&D and financial supports can assist in gettingthe products to market, but will not sell them. Tostimulate market demand, the MPT and MIT’I haveestablished a number of promotional committeesand councils. These groups are staging public eventswith HDTV and have begun extensive procurementefforts.
The MPT HiVision Promotion Council, for examp-le, broadcast the Seoul Olympics to 205 HiVisionsets at 81 public sites across Japan. Other promo-tional efforts organized by NHK, MITI, and MPTinclude: The World Fashion Fair held in OsakaKobe, and Kyoto during HiVision week (Nov.18-26, 1989); the HiVision Gallery for the GifuMuseum of Art-in which nearly half of their workshave been put into an electronic ffle system forviewing on a large screen; the Exposition of Flowersand Greenery in Osaka (April to September 1990);and many others.27 MITI’s HiVision PromotionCenter also helps set standards for industrial HDTVequipment and surveys new uses for industrial use ofHDTV,28
Private fms and consortia are also investingheavily. Recently, Japan Victor Company (JVC)contracted with a Hollywood producer for $100million to produce ftis.29 It had earlier consideredpurchasing a Hollywood studio but backed off for
~~id.
up~ker md V~d_ op. cit., footnote 1.
fiJo~o~ Op. Cit., fOOtnOte 1.
“’WhySony is Plugging Into Columbia,” Business Week, Oct. 16,1989, p. 56. There have been other notable deals forU.S. entertainme ntcompaniesrecently, including Australian Qintex Entertainen~ Inc. purchase of MGM/UA Communications, Inc. for $1.5 billion. Such deals are not related tothe development of HDTV nor the penetration of the potential U.S. HDTV marke~ however. “Invasion of the Studio Snatchers,” Business Week, Oct.16, 1989, p. 52.
27p@w ~d Vmd- op. cit., footnote 1.
uJ~es G. p~a, T~hsemch International, Inc. personal communicatio~ Wt. 6, 1989.
%lichard W. Stevenson, “Japanese Put Up $100 Million To Back Fihns in Hollywoo&” New York Times, Aug. 21, 1989.
32 ● The Big picture: HDTV and High-Resolution Systems
fear of a consumer backlash to the large number ofrecent foreign purchases of Hollywood companies.30
The parallel MPT HiVision Cities and MITIHiVision Communities programs are among themore ambitious of these market development ef-forts.31 These projects are intended to demonstrateHiVision hardware and software, stimulate thepurchase and use of HiVision equipment, andprovide a test market for HiVision applications.32
The MPT HiVision Cities program, for example,selected 14 cities from 71 proposals in March 1989to receive support in purchasing and using HiVisionequipment. These cities are pilot projects to studyuses for HDTV, including developing video andgraphics databases for museums, schools, and otherpublic uses, and to use leading-edge informationtechnologies and services to stimulate the localeconomy .33
By American standards the Japanese seem to bemarching in lock step, but there are inevitabletensions. The development of HDTV has createdconflicts between NHK and private broadcasters.Similarly, the numerous parallel programs describedabove indicate the intensity of the competitionbetween MITI and MPT to take the lead in HDTV—which both see as the gateway to Japan’s infor-mation society of the future.~ For MITI, this maybea particularly important turf battle as their traditionalrole of ‘‘protecting and nurturing Japanese indus-tries until they could compete in any market in theworld’ ’35 is ending and they must find new roles toplay.36
It is too early to tell whether this competitionbetween MITI-MPT will slow or speed Japaneseefforts in HDTV and telecommunications. Forexample, MJ.TI proposed its “New Media Conmm-nity’ ‘ and MPT countered with its ‘‘teletopia’program in the summer of 1983. The combinedcosts, however, proved too much for the Diet (theJapanese parliament), and both proposals weredropped in 1984. Not until 1986 was a compromisefound that allowed them to move forward.37 In othercases, competition might heighten efforts.
Given their lead in developing HDTV technologyand their desire to penetrate foreign markets, theJapanese have worked intensively to ensure thattheir standard was adopted worldwide. This effortstalled at the May 1986 meeting of the CCIR inDubrovnik, Yugoslavia, when European countriesrefused to accept the Japanese standard. Theirrefhsal was based on: the lack of compatibility38 ofthis standard with existing European systems and thecost of converting between these formats; thecompetitive threat they foresaw to European domes-tic television manufacturers posed by the Japanese;and the haste with which the single standard wasbeing pushed on them.39
EUROPEEuropeans formed a joint venture HDTV develop-
ment effort in June 1986 known as ‘‘Eureka Project95.” The project is charged with development ofEuropean production, transmission, display, record-ing standards, and equipment. Proposals are ex-pected to be ready forpresentationto the 1990 and/or1994 Plenary Assembly of the CCIR for acceptance
“’Invasion of the Studio Smtchers,” Business Week, Oct. 14, 1989, p. 54.
qlNote that ~e~ are no reliable estimat~ of how much the MITI Communities and MPT Cities plans are going to C@ it all depen~ on how WYprojects apply and how much is requested by each. The MPT program reportedly has roughly $100 million pending in requests; there are no figuresavailable for MITI. So-called “third sector” enterprises are being created to actually run the trial cities and communities programs once they have beenformed. Mark Eato~ MCC, personal communicatio~ Oct. 18, 1989.
3ZU.S. Congrws, House Committee on Energy and Commerce, House Subcommittee on Telecommunications and Finance, Consom”a ad theDevelopment of High Definition Systems, testimony presented by Barry H. w’hale~ Sept. 5, 1989.
qspmk= and Vard+ op. cit., footnote 1.
34Jo~o% op. cit., footnote 1.
ssJodo~ Op. cit., footnote 1, p. 183.
36David E. Sanger, “Mighty MITI Imses Its Grip,” New York Times, July 9, 1989, p. El.37Jo~ou 0p. Cit., footnote 1.
38~s lack of ~mpatibili~ is ~liev~by some to ~ve~n a s~tegic choice by Japanese manufac~e~ to force~ej-g of existing ~Uipmen4to increase their companies sales, and to squeeze competitors-who generally cannot afford to risk the long payback times that an incompatible systemwould require due to its slow acceptance. At the same time, the lack of compatibility with present TV sets was the fundamental flaw of the Japanesesystem which allowed it to be successfully attacked. Adam Watson Brow “Towards the Triumph of the Matte Black Box,” vol. 16, No. l, Zntennedia,January 1988.
3qJoseph Roizeu “Dubrovnik Impasse Puts High-Definition TV On Hold,” IEEE Spectrum, September 1986.
Chapter 2—The Historical Development of HDTV ● 33
as a co-world standard. NV Philips Co. (Nether-lands) is President of Eureka Project 95 and Thom-son SA (France) is Vice President. Robert BoschGmbH (Federal Republic of Germany), Them/EMI(United Kingdom), and more than 200therorgan.iza-tions from 9 countries are participating in the effortas well. Approximately 1,700 engineers, techni-cians, and support personnel have been involved fulltime in the project.a
The first phase of the Eureka Project 95 ended inDecember 1989 and cost an estimated $318 mil-lion.41 Sixty percent of these funds were private.
Europeans plan to begin public broadcasts ofdigital sound and provide a slightly improvedpicture to viewers through their MAC (MultiplexedAnalog Components) system in 1990. Broadcastsusing the Extended Deftition wide-screen ED-MAC system are scheduled for late 1991, and ftdlHDTV-MAC broadcasts are plamed for 1994.42 By1995, European broadcasters plan to offer fullHDTV satellite and cable services.43 Like theJapanese, the major focus of European HDTVefforts has been Direct Broadcast Satellite sys-tems.~
HDTV promotional efforts currently scheduledinclude broadcasts of the 1990 World Cup SoccerChampionships horn Italy; and broadcasts of the1992 S ummer Olympics from Barcelona, Spain.One thousand public receivers will be scatteredthroughout Europe to receive these broadcasts.45 Asin Japan, there have also been efforts in Europe to
extend the performance of existing TV systemsrather than proceeding to new and completelyincompatible systems. There is some concern thatsuch a move could slow the market acceptance oftheir MAC system.4G
Market Protection
The Europeans formed the Eureka Project 95 inpart to support their domestic consumer electronicsindustries. In addition to stimulating R&D to ensurecompetitiveness, the Europeans have aggressivelyenforced antidumping statutes and have imposeddomestic content requirements in those sectorswhere dumping has occurred. In electronics, forexample, there are now domestic content requirementsfor both systems and components. Regulationsrequire the skilI-intensive integrated circuit fabrica-tion steps, for example, to be done in Europe.Minimum prices have been set on DRAMs, as in theUnited States!7 Regulations have also been imple-mented that prevent products from simply beingtransshipped through other countries (including theUnited States) with minimal additional manufactur-ing.~
Similarly, Europe is concerned with protecting itsdomestic movie and TV-program producers as wellas its cultural identity. As Direct Broadcast Satellites(with conventional TV broadcasts) begin operationsover new charnels in Europe they are ffig air timewith U.S. TV programs and movies.49 Faced by adeluge of U.S. programs, the European Community
‘%ureka $ecretaria~ project documents.41 Peter De Selding, “Europeans Feeling Cocky When It Comes to HDTV,” New Technology Week, Sept. 11, 1989; Eureka Secretariat project
information. The budget for 1988-90 is 270 million ECU.dzNote that tie Europea.11 system is compatible with existing standards only by using electronic conv~ers in their satellite rtiivtig Wipment to
change the MAC transmissions to PAL or SECAM. William Schreiber, Massachusetts Institute of Technology, personal communication Oct. 12, 1989.Aswmley R. Iverso~ “U.S. (3ropes for Unity on HDTV,” Electronics, March 1989; Bob Whisk@ BIS Mackintosh, Pmsoti com.mtication, ~.
28, 1989; Hugh Carter Donahue, “Choosing the TV of the Future, ” Technology Review, April 1989.
‘Paul Gregg, “Satellite TV Flies European Flag, “ Electronics Weekly, Sept. 27, 1989.AsRoger wroo~ougk “World.ClaSs HDTV,” Electronic Engineering Times, Sept. 11, 1989, P. lg.
“’German EDTV Tests,” TV Digest, vol. 29, No. 32, Aug. 7, 1989, p. 15.LtT~~w pollack, “E~pe Sets Prices for Japan’s chips, ” New York Times, Jan. 24,1990. Note that EC regulations have recently been successfully
challenged under GATT See, William Dullforce, “Japan Scores Victory Over EC on Duties,” Financial Times, Lxmdon, Mar. 29, 1990; and PeterMontagnon and Lucy Kellaway, ‘‘EC Refuses To Adopt GATT Report on Dumping, “ Financial Times, Imndou Apr. 4, 1990.
~~e new EEC ~les rq~e mat tie most substitial process in producing a semiconductor be done in Europe. European ~ar~ef%esf, vol. m No.6, June 1989, p. 9.
49Jacques Neher, “A Revolution Brews In European Television@” Washington Post, Mar. 5, 1989.
34 ● The Big picture: HDTV and High-Resolution Systems
recently set guidelines to devote the majority of airtime “where practical” to European programs.50
Eureka
The Eureka Project 95 to develop HDTV is part ofthe much larger Eureka program begun in July 1985by 19 European nations to support collaborativemultinational efforts in high-technology R&D, man-ufacturing, and services. Eureka was founded on therecognition that many of the European nationalmarkets are too small for a “national champion”fii to achieve a level of sales capable of recoveringdevelopment and manufacturing costs for manyhigh-technology items. For example, the R&D costsfor a new generation of public phone switchingequipment have been estimated at $700 million to $1billion and it would require sales of $14 billion torecoup these costs.51 As a consequence, numerousEuropean collaborative R&D efforts have beenlaunched in recent years, including the $1.5 billionRACE project in telecommunications, the $5 billionJESSI project in semiconductors, and several oth-ers.52
Procedures for establishing a project under Eu-reka are both fast and flexible. Industry groupsdefine and propose their project to their respectivenational governments. If they do not already havepartners to work with, the Eureka Secretariat tries toidentify such groups. With the agreement of therespective national governments, the proposal be-gins a 45-day period of circulation and discussion.At the end of this period a gentleman’s agreement or,sometimes, an early formal agreement is reached onwhether or not to move the project forward.
European nations are also moving rapidly toupgrade their communications systems through theimplementation of ISDN (Integrated Services Digi-tal Network). ISDN allows voice, video, and data tobe carried over telephone networks in a digital
format (ch. 3) and is the next important step intelecommunications systems. European efforts inHDTV and telecommunications promise tothem important competitors in these markets.
UNITED STATES
make
After black and white TV was introduced in the1940s and color TV in the 1950s, American re-searchers looked for ways to fi.uther improve resolu-tion, color, and picture quality during the late 1950sand 1960s. They achieved resolutions in experi-mental systems roughly comparable to those nowbeing demonstrated by the Japanese.53 These ad-vanced capabilities, however, were not pursued: thecosts of these systems would have been far too highfor consumer use; the newness of the medium hadnot yet whetted consumers’ appetites for dramaticimprovements in picture resolution; and practicaltransmission technologies (e.g., bandwidth com-pression) were not available. By the late 1970s whenelectronics technologies had matured sufficiently tomake consumer HDTV systems a reasonable goal,U.S. fms were rapidly ceding TV manufacturing toforeign fms (see table 1-2). These factors limitedfurther development of HDTV in the United Statesuntil recently.
RCA’s Sarnoff labs began research in HDTV inthe late 1970s. By mid-1987, some $40 million hadbeen spent overall on its advanced television work,with an estimated $35 million required to finishdevelopment of the system they are proposing to theFCC.54 CBS has also worked on HDTV. h 1981,CBS requested permission from the FCC for use ofthe 12 GHz spectrum and began terrestrial broadcastexperiments in 1982.55 Some foreign firms havebeen active in HDTV research in the United Statesas well, including Philips (Netherlands) and, with itsrecent purchase of RCA, Thomson (France).
~Steve L,Ohr, “EurOp~nTvS vast Growth: Cultural Effect Stirs Concerq’ New York Times, Mar. 16, 1989, p. Al; U.S. Congress, House COmmittwon Energy and Commerce, Subcommittee on ‘Telecommunications and Finance, Television Broadcasting and the European Conznzuw”ty, Hearings July26, 1989, Serial No. 101-84; Edward Cody, “EC Adopts European TV Program Quot&” Washington Post, Oct. 4, 1989, p. Fl; Paul Farti “U.S. toFight EC Directive Limiting Foreign TV Shows,’ Washington Post, Oct. 11, 1989, p. F1.
51 G~efioy Dmg N@Ye% ~~Empem R&D PolicY for Tele~OmmuniCatiOns3* K~MS~ einer s~die im A~trag des ~, Bad Homef, Apd1989, Nr. 49.
52Steven Greenhouse, “Europeans United To Compete With Japan and U.S.,” New York Times, Aug. 21, 1989.53wiI~ Glenn, FIori& Atl~tic University, Boca Raton, Feb. 10, 1989; Schreiber, op. cit., fOO@te 42.54u.s. Con=ess, House co~tt= on Energy ~d Commerce, Subcommittee on Telecommunications and Fimmce, Testi~ny o~Steven Bo~”ca!
Serial No. 100-188, Oct. 8, 1987.Sscorey P. C@O~, “History of High Defiinitio~” paper presented at the “HDTV: The Second Annual conference and Exhibition on High
Deftition Television,” Arlingto~ VA, Feb. 12, 1990.
Chapter 2—The Historical Development of HDTV ● 35
Other HDTV efforts that began in the late 1970sand early 1980s include those by: Zenith; WilliamGlenn, New York Institute of Technology5b; Wil-liam Schreiber, Massachusetts Institute of Tech-nology; Yves Faroudja of Faroudja Laboratories;Richard Iredale, Del Rey Group; and others. Thesmall size of most of these efforts, however, haslimited the extent and rate of progress in HDTV andrelated technologies in the United States. Totalprivate expenditures on HDTV R&D in the UnitedStates were roughly $70 million57 as of early 1988.
There has been little public-private cooperation inthe U.S. HDTV effort compared to Japan andEurope. Significant amounts of money have beenspent, however, on related technologies such asdigital signal processing for military radar and sonarapplications.58
Production Standards
HDTV requires new standards for producingvideo material. These standards might include suchfactors as the number of lines (resolution), the framerate (50 in Europe v. 59.94 in the United States), andthe type of scarming (interlaced or progressive). Ifthese production standards differ horn one region toanother, expensive transcoding might be requiredfor programs produced in the United States to beshown in Europe and vice versa. This could dampeninternational trade in movies and TV programs.
Many U.S. movie and TV-program producershave promoted a single international productionstandard to avoid this possibility. A single world-wide production standard would aid the exchange ofvideo material between countries and help U.S.program producers maintain their current $2.5 bil-lion annual export market.59 Everyone is interestedin improving international communication and un-derstanding: uniform world standards might helpachieve this ideal.
The Society of Motion Picture and TV Engineers(SMITE) formed a Working Group on High Defti-tion Electronic Production in 1977 and began workin 1981 on a production standard. One concern ofthis group was to document and standardize HDTV-production equipment development. A second con-cern of the SMPTE was that differing European,American, and Asian production standards wouldlimit the U.S. producers global marketing of theirprograms.
The work done by the SMFTE resulted in a set ofparameters which differed slightly horn the NHK1125/60 system, then the only system available inthe world and one of the focuses of the SMFTEwork. These parameters were adopted by the Japa-nese.m The U.S. Advanced Television SystemsCommitteebl (ATSC) recommended the resulting1125/60 standard to the U.S. State Department andit was supported by the official U.S. delegation at theDubrovnik meeting of the CCIR in May 1986. Achronology of all these various activities within theUnited States is listed in table 2-2 and many of thegroups actively involved with Advanced TV de-velopment in the United States are listed in table 2-3.
A number of domestic groups, however, opposedthe 1125/60 standard developed by the SMPTE andothers. NBC and ABC, for example, have opposedapproval of the 1125/60 format because they expectthat conversion from 1125/60 to whatever standardis chosen for the United States, such as 1050/59.94,will be costly and will introduce unacceptable flawsin the picture-thus hurting their domestic broad-casting market.b2 Instead, they believe that a trans-mission standard should be chosen first that meetsU.S. needs. Following that, a family of internationalproduction standaxds can be established to best fitthe national needs of each country. Americanbroadcasters are also concerned with the potentialcosts of converting their equipment to HDTV and
fire. Gl~ is c~nfly with Florida Atlantic University, Boca RiXOn.57~FQ Dwby, “~onomic potenti~ of Advanced Television prOdUctS, “ National Telecommunications and Information Administration+ Apr. 7,
1988.5sFor res~h directly related to HDTV, relatively little has been spent, with NASA a particularly important supporter of HDn-like ~D.5~s is the netuos. tie for 1988 fimotion pic~s and ~ progr~s. Motion picture &sOc~tionof Americ@-fotion piCluI’C %port Association
of Americ% “Fact Sheet: Film and Video Piracy.”
@The SMPTE began their work assuming no parameters and spent several years looking at a variety of factors such as chromaticity before settingline and fmrne rates. Birney Dayton, NVisiom personal communicatio~ Oct. 12, 1989.
61~ ad hoc SOUP fom~ in 1984 by v~ous groups (s= Box 3.’3) to develop VOl~~ stamkds for ~Tv.
62 C$HDTV pr~uction: The Future IS Almost Now,” Broudcusting, Oct. 17, 1988, P. 39.
36 ● The Big picture: HDTV and High-Resolution Systems
Table 2-2-Chronology of N development
1928
1936193919411953
1960s19641964
19721977
19791981
1982
19821983
1984
1984
Radio Manufacturers Association begins planning forB&WTVRCA demonstrates prototype B&W TVFirst public TV broadcast done at New York World’s FairNTSC standards are set for B&W TVNTSC standards are set for color TV that is compatiblewith B&W setsU.S. firms develop prototype high-resolution TV systemsAll-Channel Receiver Law enactedJapanese state-owned broadcasting system NHK beginsHDTV developmentNHK submits draft HDTV study program to CCIRSociety of Motion Picture and TV Engineers (SMPTE)begins study of HDTVNHK conducts first satellite transmission tests of HDTVFirst demonstration of HDTV in US; SMPTE recommendsaction on HDTV in USCBS conducts transmission tests in satellite broadcast-ing frequenciesCBS announces two-channel HDTV systemThe Center for Advanced Television Studies (CATS) isformedThe Advanoed Television Systems Committee (ATSC) isformedJapanese NHK establishes the MUSE transmissionstandard for DBS
19841985
1986
1986
1986
1987
1987
1988
1988
19881988
1988
1989
FCC allows the introduction of stereo sound broadcastsATSC recommends the NHK production standard toDepartment of StateNational Association of Broadcasters forms task Forceon HDTVU.S. Department of State and Japanese propose NHKproduction standard to CCIREuropeans begin Eureka projeot to develop their ownHDTV systemNational Cable Television Association forms Comnitteeon HDTVFCC creates Advisory Committee on Advanced Televi-sion Services (ACATS)Electronic industries Association forms an Advanced TVCommitteeAmerican Electronics Association forms a task force onHDTVAdvanced Television Test Center establishedBell South/Belicore transmits HDTV by fiber at Demo-cratic National ConventionFCC tentatively decides HDTV must be compatible/simulcast with NTSC; no additional spectrum will bemade available outside existing bandsDARPA announces $30 million of funding for HDTV R&D
SOURCE: Adapted from: “A Chronology of High-Definition Television Development,” VarjetV, Oct. 12, 1988, m. 74-75: Tonv Uvttendaele, ABC, wmonalcommunication, Sept. 7, 19&; and Electronic Industries Association, “Chronology-of TV and” AT/ D6~elopment,” N&v. ~0, 1988, Washin~on, DC.
are considering intermediate EDTV quality systemsas a substitute.
Ofllcial U.S. support for the Japanese/ATSC1125/60 standard was reversed in May 1989 due tothe unsuitability of its use within the United States;the recognition that 1125/60 would not be adoptedas a world standard due to European opposition; andpolitical opposition to handing the Japanese acompetitive advantage in manufacturing HDTVS.63
Program production standards that can easily betranscoded from one system to the other (which theoriginal Japanese NHK system cannot) are, how-ever, still possible and desirable.~ Work on such a“Common Image Format” was proposed by theU.S. delegation and endorsed for further study at theMay 1989 CCIR Extraordinary Meeting for HDTV.
Transmission Standards
The FCC formed the Advisory Committee onAdvanced Television Service in 1987. The Advisory
committee formed three subcornmittee8- Planning,Systems, and Implementation-to review proposedterrestrial broadcasting standards for the UnitedStates.
In September 1988, the FCC issued a tentativedecision that whatever new transmission standardwas chosen, HDTV broadcasts should also beavailable to those with today’s NTSC sets. Further,additional spectrum outside of today’s UHF andVHF bands for TV broadcasting would not be madeavailable for HDTV, although individual stationsmight get as much as 6 MHz additional spectrumwithin these bands.
New Initiatives
U.S. Government efforts in support of HDTVR&D have increased during 1989. In December1988, Defense Advanced Research Projects Agencyannounced a $30 million, 3-year R&D program forhigh-resolution displays and supporting electronics.
63John Burgess, “U.S. Withdraws Support for Studio HDTV Standard,” Washington Post, May 6, 1989; Lucy Reilly, “State Department ChangesStance On HDTV Standards,” New Technology Week, May 1, 1989.
@w. F. Schreikr, “A)?riendly Family of Transmission Standards For All Media and All Frame Rates,’ Draf4 Feb. 12,1989, Massachusetts Instituteof Technology, Cambridge, MA.
Chapter 2—The Historical Development of HDTV ● 37
Table 2-3-Private U.S. Groups Involved With HDTVa
Group Purpose Members. . . . . . .
Ad Hoc Group on An
Advanced TV Research Pro-gram (ATRP)
Advanced TV Systems Comm-ittee (ATSC)
AdvancdlVTestCenter (A~C)
American Electronics Associa-tion (AEA)
American Nationai StandardsInstitute (ANSI)
Association of Maximum Serv-ice Telecasters (MST)
Broadcast Technology CenterCabie TV Labs. Inc.
Center for Advanced TV Stud-ies (CATS)
Electronic Industries Associa-tion (EIA)
HDTV 1125/60 Group
Institute of Electrical and Elec-tronicsEnginsers (lEEE)Td-nology Activities Council
Motion Picture Association ofAmerica (MPAA)
National Cable TV Association(NCTA)
Nationat Association of Broad-casters
Society of Motion Picture andTV Engineers (SMPTE)
Telecommunications IndustryAssociation (TIA)
To reoommend policy to aid U.S. HDTV efforts
R&D for advanced TV systems and AssociatedRegulatory Policies
Develop voluntary standards for HDTV
Testing of ATV systems
To assist U.S. firms in electronics manufacturing
A national standard setting organization that setsvoluntary standards for the U.S. and helpsdevelop U.S. position internationaily
To find means for 100al stations to provide HDservice
R&D for broadcasters into HDTVR&D in HDTV for the cable TV industry
Funding for independent TV R&D
To recommend policies on Advanced TV
To promote the 1125/60 (modified version oforiginal NHK) production standard
To recommend policy to promote U.S. interests inelectronics
Trade Association for U.S. motion picture industry
To assess HDTV systems for cable applications
Trade Association for U.S. radio and TV broad-casters
Tdnicai society deveiophg voiuntary recommendedstandards and practices
To present position of telecommunications supplycompanies on HDTV
Various; organized by researchers at Ml I
ABC; NBC; HBO; PBS; Ampex Tektronix; RCA;Zenith; Kodak; Generai Instruments; etc.
Electronic Industries Association; IEEE; NationaiAssociation of Broadcasters; Nationai CabieTV Association; Society of Motion Picture andTV Engineers; etc.
Nationai Association of Broadcasters; Associationof Maximum Service Telecasters; Associationof Independent TV Stations; Capitai Cities/ABC; CBS; NBC; PBS; etc.
U.S. hi-tech electronics firms inciuding: AT&T;iBM; Hewiett-Packard; Motoroia; intei; ilTDEC; National Semiconductor; Tektronix; Ti;Zenith; etc.
Various
270 iooai stations: independent ornetworkaffiiiate
Nationai Association of Broadcasters; CBS; etc.63 cable TV companies with over 75% of U.S.
cabie subscribersABC; Ampex; CBS; Kodak; Zenith; NBC; PBS;
RCA; HBO; 3M Co.; Harris Corp.Foreign and U.S. firms: Foreign consumer elec-
tronics firms inciude: Mitsubishi, NEC, Sony,Phiiips, Panasonic, Thomson, Hitachi, etc.
Sony; NEC; Panasonic; JVC; Toshiba; Chyron;Cinema Produots; Compression Labs; DynairElectronics; etc.
Represents 240,000 engineers in the U.S.
Various
Teiecabie; Scientific Atianta; Generai Instrument;HBO; ESPN United Artists Cabie; Fox Cabie;Warner Cabie; etc.
Various
Various end users, manufacturers, and individualengineers from the TV and fiim industry
Aicatei, NA; AT&T; Corning Giass; GTE; Hughes;IBM; Motorola; Scientific-Atianta
~his list is not inclusive.
SOURCE: Electronic Industries Association, “Non-Governmental Groups Involved With HOW,” HDTV Information Packet, Washington, DC, Nov. 30, 1988;Greg Depriest, Association of Maximum Serviee Telecasters, personal communication, Jan. 18, 1990.
DANA received 87 proposals requesting roughly U.S. commercial interests have also attempted to$200 million. By November 1989, 11 contractors forge a basis for public-private cooperation. In Mayhad been named.G5 Congress appropriated $20 rnil- 1989, the American Electronics Association (AEA)lion for this R&D in the fiscal 1990 budget, but Advanced Television Task Force, representing 36stipulated that the money could not be spent until the U.S.-owned firms, proposed a $1.35 billion (govern-Administration had developed a comprehensive ment share) 6-year program of R&D, low-cost loans,program for HDTV. and loan guarantees to assist U.S. manufacturers in
tiBfi Rob~oq ‘‘DARPA Names DTV ContractorS,” Electronic Engineering Times, June 19, 1989, p. 11; Richard Doherty, “Six Firms GetDARPA HDTV Nod,” Electronic Engineering Times, Oct. 30, 1989, p. 1.
38 ● The Big picture: HDTVand High-Resolution Systems
designing, developing, and producing HDTVs.tiThis effort, however, met considerable skepticism.A report fi-om the Congressional Budget ~lce(CBO) in July 1989 questioned the premise-thatHDTV was an important market that U.S. fmshould enter.G7
Lacking support at home, the tentative unity ofU.S. industry groups began to hay. In November1989, an initial agreement was reached between theU.S. Semiconductor Industry Association and theElectronic Industry Association of Japan for U.S.fms to supply Japanese HDTV manufacturers withchips.~ Some observers cited this action as a goodfaith effort on the part of Japan to increase use ofAmerican chips.
Skeptics, however, noted that U.S. producersreceived no guarantees as to what proportion of the
semiconductors in Japansese HDTVS would besourced from them. The general failure of theSemiconductor Trade Agreement to substantiallyincrease U.S. chip sales in the Japanese market wasoffered as a basis for questioning the value of thisHDTV accord. Skeptics suggested that the Japanesemight use this agreement to influence the debate inthe United States over which broadcast standards arechosen. If Japanese HDTVS contain a si@lcanttiaction of U. S.-sourced semiconductors, it wouldbe much more dilllcult to restrict imports in favor ofU.S.-developed or U.S.-produced HDTVS. In thatcase, it might also be more d~lcult to just@supporting the development of a domestic HDTVindustry.
fi~en=n Elec@onics Association, “Development of a U.S.-Based ATV Industry,” May 9, 1989.
cTCongressio@ Budget OffIce, “The Scope of the High-Definition Television Market and Its Implications for Competitiveness,” Staff WorkingPaper, July 1989.
‘David Sanger, “Industries in U.S. and Japan Form Alliances on New TV Technology,” New York Times, Nov. 9, 1989, p. Al.
Chapter 3
Communications Technologies
INTRODUCTIONThe U.S. electronic communication infrastructure
is a mixture of five media: 1) terrestrial radiofle-quency transmissions; 2) satellite radiofrequencytransmissions; 3) paired copper wires; 4) coaxialcables; and 5) optical fibers.l Each plays a role inproviding efficient communication services for theUnited States (table 3-l). Rapid technological andeconomic changes are reshaping the roles thesemedia play in our communication system.
Terrestrial broadcasting and coaxial cables dis-tribute the major share of television to consumerstoday. Satellite transmission has ffled the need of
those not served by cable TV (CATV) or who are toofar from broadcast stations to receive a reliablesignal. HDTV could change the competitive balancebetween these various media and may stimulate thedelivery of television services by other means,including direct broadcast satellite and, in thelong-term, switched optical fibers through the tele-phone system. HDTV may also allow significantimprovements in the efficiency of spectrum use.This might enable services now constrained by alack of spectrum to be expanded.
The shift towards a digital operating environmentis making the media perform more alike. Digitiza-tion can provide a common format that allows
Table 3-l—The U.S. Communications Infrastructure
ServiceYear Approximate Available industry
Technology Businesses began Current capacity number of nodes audience revenues
Copper pairs Telephone, datacommunication(LANs, WANS,MANs)
Coaxial cable Cable IV
Optical fiber Voice, data,and videocommunication
Terrestrial AM radiocommunication FM radio
VHF TVUHFTVITFS & OFS TVMMDS TVCellular telephone
Satellitecommunication Voice, data, video
1876
1948
1978
1920s1920s1946195219631962/831983
1962
L band Radio RDSS paging 1989C band BroadcasVcable 1976Ku band Medium-powerDBS 1983Ku Band High-oower DBS 1987-.
From 3 kHz for single analog voicechannel 144 kbps for basic ISDN
Up to 550 MHz or 80 video channels
1.76 Gbps common and higher ratesrapidly becoming available
Total’ Spectrum Channelsspectrum; /channel; /market1070 kHz 10 kHz; 2720 MHz; 200 kHz; 5072 MHz; 6 MHz; 7
330 MHz; 6 MHz; 10138 MHz; 6 MHz; N/A
60 MHz; 6 MHz; N/A50 MHz; 30 kHz; varies
16 MHz;1 GHz; 40 MHz;1 GHz; 40 MHz;1 GHz; 15 MHz;
23,000 exchanges240,000 PBXS
11,000 headends
Overlaps with cop-per pair and coaxialcable systems
5,000 AM stations5,600 FM stations703 [i300 ~ SfnsSame800 stations24 stations700 systems
530 Civ. transpon-ders1 sat/7 slots19** sats/35 slots14 sats/35 slotsO sats/8 slots
93Y0 wlphones $140 billion20% WIPCS
8Y0 w/modems
81Y0 of homes $14 billionpassed;56% subscribe
98% population $2 billion98% population $4.5 billion98% population $24 billionSameN/AO.l YO population1% population $3 billion
1‘Yo direct
* Note that the total spectrum available and the spectrum per channel are set by regulation. The channels usable per market are also set by regulation in orderto control interference between stations in the same geographic location.
● * Four satellites serve both C-band and medium-power DBS Ku-band; slots are orbital slots where satellites might be placed, with varying numbers oftransponders per satellite.
SOURCE: Adapted from U.S. Department of Commerce, National Telecommunications and Information Administration, Te/scom 2000, NTIA SpecialPublication 88-21, October 1988.
Iucs, c~nge~~, offi~ of Tm~ol~~ &sessmen~ C~’tiCal connections: co~nicafion for the Fu~re, Ow-~-407 (%bh@toll, ~; U.S.Government Printing OffIce, January 1990). More detail of the technologies and analyses of communication issues maybe found in this report.
-39–
40 ● The Bi~ Picture: HDTV and High-Resolution Systems.- — u
signals to be easily moved among media-terrestrialbroadcast, microwave, satellite, cable, optical fiber—without distortion.
TERRESTRIAL RADIOCOMMUNICATIONS
Terrestrial radio frequency communication in-cludes mass media such as AM and FM radio andtelevision; mobile radio links such as cellulartelephone and police radio; and special services—amateur radio and aeronautical and marine naviga-tion. Ground-based microwave links are also usedfor long-distance telephone, data and video trans-missions. (Fundamentals of communication tech-nologies are listed in box 3-l.)
Each service using the radio frequency spectrumis assigned a range of frequencies that matches thetechnical needs of the user and reduces interferencefrom others sharing the spectrum (figure 3-4).2
The ikquencies from about 30 MHz to 1 GHz arewell suited for terrestrial, short-distance communi-cation. As the frequencies increase above 30 MHz,radio transmissions become limited to direct line-of-sight (reflections from the ionosphere decrease),with less interference between stations. Below 1GHz, the strength of radio waves are not signiflcantly reduced by rainfall or other atmosphericeffects. Above 10 GHz, attenuation from rainfall andother atmospheric factors increases rapidly and limitthe use of these higher frequencies to short-haullinks or for satellite communications where thesignal must pass through only a few miles ofatmosphere to the receiver.3
Competition for spectrum space is often keen. Astechnology develops new and better communica-tions systems, user demand can cause crowding andoveruse of some bands. Cellular telephone and othermobile services are assigned frequencies adjacent tothe UHF TV bands. These services have experiencedphenomenal growth in several urban areas, wherethey now suffer horn congestion.
The Federal Communications Commission wasconsidering the reallocation of the upper parts of theunderutilized UHF TV spectrum to these otherservices when the potential needs of broadcasters for
additional frequencies to broadcast HDTV led theregulatory agency to stay its action.4
Microwave systems are important links in thenational telephone network and are also used inprivate nets. Modern digital microwave systems canachieve data rates of hundreds of megabits persecond. Microwave provides line-of-sight commu-nication, and repeaters are spaced at differentintervals (commonly 20 to 30 miles) depending onterrain, weather conditions, obstructions, and fre-quencies.
Microwave systems operating in the 2-, 6-, and11-GHz bands are typically used for long-distancetransmission, while 18- and 23-GHz systems arebetter suited to short hauls due to the greaterattenuation by rain at higher frequencies.
Sufficient bandwidth is available in the micro-wave bands to carry high-resolution, fill-motiontelevision. Microwave-based mtitichannel multipointdistribution systems (MMDS) have been licensed todeliver video services in some metropolitan areas. Ingeneral, these ventures have had difficulty compet-ing with conventional CATV service.
Cellular telephone is perhaps the best example ofthe tremendous changes taking place in the use of theradio spectrum as a result of technological advances.Conventional radio telephone service-the prede-cessor of cellular-has a capacity of about 25channels, each able to carry one call. Radio tele-phone required a high-power transmitter and areceiver capable of communication up to 50 miles.
Cellular telephone technology dramatically in-creased the capacity for mobile telephone communi-cation by breaking the service area into many‘‘cells, ’ each served by its own transmitter andreceiver coupled to a computer-controlled system(figure 3-5). Each cell operates at low power andshort range. Neighboring cells use different frequen-cies to avoid interference, but cells far enough apartto avoid interference may share the same frequency.If a customer moves into the range of another cellduring a call, the fkequency is automaticallyswitched to that cell without the caller knowing.These techniques greatly increase the carryingcapacity of the limited spectrum available for
2For details of these services, see FCC Rules, 47 CFR 90 to 100.3whme tie ~tellite is n= the hofion, the dis~~ through the lower atmosph~e cotid be much longer, perhaps 50 mileS Or mOre.
dF~m S- of the ~ ~d Mobile fidio seti~, Ordti, 2 FCC 6441 (1987).
Chupter 3-Communications Technologies ● 41
Box 3-1—Fundamentals
Information can be electronically communicated in ananalog or digital format. Other characteristics that must beconsidered include: how frequently the signal must beamplified to make up for losses in the medium; how tocarry several independent signals in a medium withoutinterference; and how much information can be carried.
Analog—The analog format sends a continuouslyvarying electronic signal proportional to the information.For example, a microphone generates a signal with avoltage proportiona.I to the loudness of the speaker’svoice. For applications such as radio and TV broadcast-ing, this signal modulates a carrier wave, most commonlyby varying the amplitude (AM) or frequency (FM), figure3-1.
Digital—For digital communications, the continu-ously varying signals of the real world—sound, light,
are first converted to numbers (usually in a binaryetc.—format, that is, as a string of “0s” and “ls”l ) propor-tional to their loudness or brightness. (See box 4-2.) Thesebinary digits, orbits, can then be sent as a series of “on’or ‘‘off pulses (figure 3-2) or can be modulated onto acarrier wave. Typical modulation techniques vary thefrequency, amplitude, and/or phase of the carrier waveaccording to whether the signal is” 1“ or “O.” Examplesof these techniques are shown in figure 3-3.2
Amplifiers (Analog Repeaters)—Transmitted overlong distances, signals are inevitably attenuated. To boosttheir strength, analog signals are electronically amplifiedby repeaters at points along the way. With each amplifica-tion, there is a tiny amount of distortion because theelectronics are not perfect. In turn, because the analogsignal is continuously varying, it is extremely difficult todetect or correct the distortions that creep in.
Digital or Regenerative Repeaters-digital signalsare also attenuated over long distances, but digital signalsare either “l” or “O.” Therefore, instead of simplyamplifying the digital sigmil along the way and introduc-ing more distortion each time, the data in the signal—‘‘1s” or “Os”-are recovered and anew digital signal isgenerated. As long as the signal is not so severelyattenuated or distorted that a “1“ appears like a “O” orvice versa, there is no degradation of the data. Of course,digital signals can also be transmitted over analog systemsand their signal simply amplified. Digital repeaters arealso known as “regenerative repeaters. ”
Figure 3-l—Amplitude and FrequencyModulation of a Carrier
Carrier
WuModulating sine-wave signal
Amplitude-modulated wave
Frequency-modulated wave
SOURCE: William Stallings, Data and Computer Ccxnmunieations(New York, NY: MacMillan Publishing Co., 1985). Usedwith permission.
Figure 3-2—A Digital Signal With ItsCorresponding Numerical Equivalent
DataIransmllec [1 1 c 1 1 0 0 1 1 0 0 1 0 1 0
SOURCE: William Stallings, Data and Computer Communications(New York, NY: MacMillan Publishing Co., 1985). Usedwith permission.
I
l~us, to count from () m 15 in binary gives the following series: 0000; 0001; 0010; 0011; 0100; 0101; 0110; 0111; Im, lal; 101Q1011; 1100; 1101; 1110; 1111. Each position from the right corresponds to a power of 2 just as in the decimal system we commonly use, eachposition from the right corresponds to a power of 10.
2L. Bre~ Glms, C ‘M~e~ M~~ Me~ods,” Byte, June 1989; William s~~gs, Data a~~COmp~ter Co?rl?nunicarions (NeW YOr~ NY:Machfillanl%blishing, 1985).
Continued on next page
42 ● The Big pic~re: HDTVand High-Resolution Systems
Box 3-1—Fundamentals<ontin uedRepeater Spacing—High frequencies are attenu-
ated more rapidly than lower frequencies-whether overcopper pair, coaxial cable, optical fiber, or terrestrial orsatellite broadcasting through the atmosphe~and thislimits the range of carrier frequencies that can be used orelse requires analog or digital repeaters to be spaced moreclosely.
Multiplexing—To carry more than one stream ofinformation in a medium at a time, signals are multi-plexed. Frequency division multiplexing (FDM) usescarrier waves of different frequencies so that a number ofsignals can share the same medium without interferingwith one another. FDM is used in radio and TVbroadcasting, and the frequency of the carrier wavecorresponds to the channel. Time division multiplexingbreaks the data into srnallerpackets and intersperses them.(See bOX 4-2.)
Bandwidth—Modulating a signal onto a carrierwave-whether by AM or FM-causes the frequency ofthe signal to vary from that of the carrier and creates‘‘sidebands’ with the signal occupying a range offrequencies about the carrier frequency. This can be mosteasily seen in figure 4-3 for flequency modulation. Toprevent interference between radio or TV stations, therange of frequencies of a signal are not allowed to overlap.This range is known as the bandwidth, and is about 3 kHz3for a telephone conversation; 10 kHz for an AM radiostation; 200 kHz for an FM radio station; and 6 MHz fora TV station.4
The maximum potential information content orcarrying capacity of an analog signal is given by its
Figure 3-3-Techniques for Modulating anAnalog Carrier in Order To Send Data in
a Digital Format
0 0 1 1 0 1 0 0 0 1 0
Amplitude-shift keying
0 0 1 1 0 1 0 0 0 1 0
IFrequency-shift keying
0 0 1 1 0 1 0 0 0 1 0
Phase-shift keying
SOURCE: William Stallings, Data and Computer Communications(New York, NY: MacMillan Publishing Co., 1985). Usedwith permission.
bandwidth (but is limited by the noise present). The higher information content of HD~ requires either widerbandwidths than today’s allotted 6 MHz or much more efficient use of that available.
Data Rates-The information content of a digital signal, or data rate, is measured by the number of bits persecond (bps) that are transmitted. When digital signals modulate a carrie~, the data rate is determined by howefficient the encoding process is and the available bandwidth of the carrier. Ttiy, data rates of as high as 7.5 bpsper Hz of bandwidth are achievable with electronic signals, although at these very high rates transmission errorsare more likely.6 Due to technical limitations in the semiconductor lasers that drive light through optical fibers andthe sensors that detect this light, encoding efficiencies (bps/Hz) for optical transmissions are not as great as forelectronic signals. Optical fibers are nevertheless able to carry much more information than twisted pairs or coaxialcables because of the greater range of frequencies they can carry without excessive attenuation.
s~e ICHZ is 1,000 cycles p saond; 1 MHz is 1 million cycles per second; 1 GHz is 1 billion CyChX pm -rid.
4For AM and FM tiese include guardbsnds; for TV, adjacent channels cannot be assigned in the same g~gr$ptic region $nd ~ ~o~as taboo channels.
5Note that all cfiers are analog.6D~e ~~eld, ~CRep@ on tie pot~~ for Extreme Band~d~ Compression of Digi~ized HDTV Signrds,” Hhgs bSfOre he
Committee on Science, Space, and Technology, U.S. House of Representatives, Mar. 22, 1989.
Chapter 3-Communications Technologies ● 43
Figure 3-4-Selected Allocations of the Radio Frequency SpectrumFrequency
Electromagnetic (hertz)Spectrum
~“”
t 10 2’
X-raysand
t
10’ 0
Gamma rays
t1 0 ”
I
4’0Ultraviolet t
10 ‘
Infrared
t
10 ‘
– lo”
Radio frequency – 1 0 ’bands
Frequency
3oGH/
/10GHI
EHF
SHFt
l o ”
UHF
I10’
VHF IHF
[
10
(shortwave)10‘
l’- 10LF
}
104
VLF
1- 10
I 10’
J,o
3 GHI
1 GH,
300 MH/
100 MH/
MF
30 MH
10 MH7
\
300 KH
U S FrequencyAllocations
Amateur
4UHF
TV broadcastchannels 14-69
4
VHF TV broadcastchannels 7-13
4-
FM broadcast
VHF TV broadcastchannels 2-6
Land mobile(lncludl~g police)
Amateur and CB
4
Aeronautical
Pohce
Amateur
AM broadcast
—
> C Band
- ITFS
- MMDS
Cellular- t e l ephone
– Cit izens’ radio— L a n d m o b i l e
(including police)
Land mobile~lncluding police)
- Mar ine
SOURCE: Adapted from: Encyclopedia Americana, vol. 23, 1986, p. 160; and FCC Rules, 47 CFR 90 to 100.
44 . The Big Picture: HDTV and High-Resolution Systems
Figure 3-5—Map of a Cellular Radio System
w Mobile telecommunications❑ switching office(MTSO)
o Cell siteCellular radio layout
SOURCE: William Stallings, Data and Computer Communications (NewYork, NY: MacMillan Publishing Co., 1985). Used with permis-sion.
cellular phones. Digital cellular, now coming on thescene, promises to improve spectrum utilizationeven more.
A cellular system can grow to meet demand.Initial service to a metropolitan area might require afew large cells. As demand grows, cells can besubdivided and distant cells can use the samefrequency. The system is quite flexible, but there arelimitations.5
The FCC has allocated 50 MHz in the UHFregionof the spectrum to the mobile telephone service. Thisspectrum is subdivided into 832 pairs of 30 kHzbands (two-way conversations require one pair).Overall, this is sufficient capacity to handle theroughly 3 million users in the United States today,but there is significant congestion in some areas. InIms Angeles, New York, and Washington, DC, thecellular channels are overloaded during peak tirnes.6Faced with an explosion in demand for cellularservice and little prospect for acquiring additionalfrequencies, the industry is focusing on technologiesthat will use the assigned frequencies more effi-
ciently. The most promising is end-to-end digitaltransmission, which is being adopted as an industrystandard.
Digitization will increase the capacity of theassigned spectrum by as much as three to eight timesor 9 to 24 million calls. The cost of conversion couldbe more than $4 billion.7 Trends toward miniaturiza-tion leading to pocket-size cellular phones couldboost demand even further.8 Some industry analystsbelieve that cordless, mobile telephones may be thetrend of the future, and that if provided sufficientspectrum for expansion, they could be substantialcompetition for the local telephone companies.9
With the prospect of advanced TV (ATV)serviceslooming in the future, competition between Amand cellular for the UHF TV bands will heighten.Both services may suffer spectrum compaction atthat point unless technology can improve the effi-ciency of spectrum use or new transmission schemesare devised.
SATELLITE COMMUNICATIONSSatellite systems are long-distance links for tele-
phone, data, and television transmissions to distantground-based stations. Although currently usedprimarily for feeds to broadcast stations and cablesystems in the United States, another generation ofsatellites called Direct Broadcast Satellites (DBS)could be used to beam HDTV signals directly tohome viewers in the future. Japan and Europeancountries are planning to use DBS to deliver HDTV.The Japanese are introducing HDTV via DBS as anew service, and not as a replacement for NTSC,which will continue to be broadcast terrestrially.
Satellites are ‘‘parked’ in geostationary orbits22,000 miles above the Earth-about one-tenth thedistance to the Moon. By staying in the sameposition, their “footprint” (area of coverage) re-mains the same. Time delays for signals to and fromEarth at that altitude disturb voice communications,but have no practical effect on video signals.
sDaleN. Ha&le]d and GeneG.@‘‘The opportunity CoStS of Spectrum Allocated to High Deftition Television,” paper presented at The SixtemthAnnual Telecommunications Policy Research Conference, Airlie House, Airlie, VA, oct. 30 to Nov. 1, 1988; William Stallings, Data & ComputerCommunications (New York, NY: MacMillarL 1985).
6~dusv a~ysts for-t tit by 1995 there w~ be abut 20 million cellular phones compared to 2.6 million iII SemiCe today. See “Cell~m phoneIndustry’s Numbers,” Washington Post, June 8, 1989, p. El, E6.
TC~v~ SiLUS, C’Meeting Mobile Phone Demand,” New York Times, July 19, 1989, p. D1.gFr@ James, “Bat~e of Miniature Cellular Telephones Heat Up with Launch of Motorola Model,” Wall Street Journal, May 5, 1989, p. B1.%Iatfield and Ax, op. cit., footnote 5.
Chapter 3-Communications Technologies ● 45
Fixed Satellite Services (FSS)1° are assignedorbital positions in space with separation betweensatellites arranged to minimize interference accord-ing to power levels and frequencies.11 Spacingrequirements limit the number of satellites in orbitfor each band to about 40 for FSS or 8 for DBS.12
Satellite transponders receive signals at onefrequency and relay them to Earth on a secondfrequency. The FCC has allocated 1 GHz of band-width in the spectrum (500 MHz for the uplink(ground to satellite) and 500 MHz for the downlink(satellite to ground). C-band satellites were the firstdomestic satellites in operation and carry the major-ity of satellite voice, data, and video communica-tion. The Ku-band and the Ka-band (when opera-tional) will carry increasingly more traffic.
The FCC regulates the broadcast power of C-bandtransponders to reduce the potential of interferencewith terrestrial communications, e.g., point-to-pointmicrowave links.13 Imw power requires the use oflarge receiving dishes (C-band antennas are 10 to 12feet in diameter). Antenna eftlciency is improvingthrough new technology, and 4-foot diameter anten-nas are now used for TV terminals.14
Ku-band transponders operate at higher frequen-cies than are commordy used for terrestrial commu-nication, so the FCC allows them to operate at higherpower. Higher power allows the use of smallerreceiving dishes (3 feet in diameter), but the higherfrequencies are subject to interference horn rain andother atmospheric conditions.
There are currently no high-powered DBS TVservices in the United States. The FCC promulgated
regulations for DBS in 1982, but the medium hasfaced economic uncertainties that have sloweddevelopment. M ~s si~tion dramatically changedrecently with the announcement of a partnership toestablish a DBS system for the United States by asearly as 1993.16
DBS can be received on antemas of 1 foot indiameter. This scale-down could overcome some ofthe objections and limitations imposed by somemunicipalities on Television Receive-Only (TVRO)home satellite dishes that receive low-power sigmdshorn the C-band satellites. Others argue that cablecompanies installed plant and equipment in urbanareas minimizes the incremental cost of adding newcustomers or HDTV progr amming and limits DBSSadvantage to largely rural areas.]7
COAXIAL CABLESCoaxial cables are widely used for distributing
television to subscribers.18 About 80 percent ofAmerican homes have access to cable TV, and 55 to60 percent currently subscribe.
Cable TV systems generally use a cable trunk thatcarries up to 80 frequency division multiplexedanalog video signals. Taps from the coaxial trunk arefed to each subscriber. Signal strength in the trunkdiminishes with distance. Therefore, analog amplifi-ers are placed at regular intervals within the systemto boost the signal strength; but each amplifier addsnoise to the signal. To avoid excessive loss of picturequality, cable operators limit the length of the trunksystem. Improvements in amplifiers and their closerspacing might increase coaxial cable bandwidths upto 1 GHz.19
l~re ~ SeV~ chses of~s mtel.lit= OWrat@ at different frequencies and different power levels, e.g., C-band (4/6 GH.z), Ku-bnd (1 1-12/14GH.z), and in the future, Ka-band (20/30 GHz).
ll~siment of ~bi~ Locations to Space StatiOnS in the Domestic Fixed-Satellite Service, 50 Fed. Reg. 35228 (1985).
12TJIis rePesents tie n-r of ~tellites that can cover between about 60 to 140 degrees longitude across the United States. Fewer Mtellites wotidfit into the range of 85 to 130 degrees longitude (hat provides full useful coverage.
13Timo@ fia~ ~d tiles W. Bos@ satellite co~nications (New York ~: Job Wiley & SC)IIS, 1986); ~d ChdeS W. Bost@ VirginiaPolytechnic Institute, personal communication, Sept. 29, 1989.
14Harry Jessell and Peter Lamb@ “The Uncertain Future of DBS,” Broadcasting, Mar. 13, 1989, p. 42.lssee f~DBs vi~ility Questioned,” TV Digest, July 31, 1989, p. 8; David Rose~ “The Market for Broadband Services via Fiber to the Home,”
Telematics, May 1989, vol. 6, No. 5.
lbJobn Burgess, ‘‘Satellite Partnership Plans Pay-TV System,”Washington Post, Feb. 22, 1990, p. El.ITJolM Sic, Telecommunications, hlC., pCrSOIlrd commmUnicatiow Oct. 12, 1989.18cofi c~1e8 &ve a ~en~ COPF me ~o~d~ by ~ fi~tor wi~ a fiely br~ded COpper stield around the outside of the insulator. The
unit is encased in plastic or rubber armor. The concentric geometry allows higher frequencies to be carried W can pairs of copper wire. Most coaxialcable systems now inusehave bandwidths of about 300 to 400MHz, and some go as high as 550 MHz. Tbis is sufficient to carry 35 to 80video cbannels.See, Trudy E, Bell, “The New Television: Looking Behind the ‘IWc,” IEEE Spectrum, September 1984, p. 53.
l~~e ~~lel~ Hatlleld Associates, he., personal Communication% oct. 12, 1989.
46 ● The Big picture: HDTV and High-Resolution Systems
Many cable operators are considering replacingthe “backbone” of their cable system with opticalfiber.20 This would greatly reduce the number ofexisting arnp~lers between the cable headend andany customer and correspondingly improve signalquality and channel capacity .21 This could, by someestimates, be achieved for as little as $50 to $75 persubscriberazz This comparatively inexpensive up-grade of the system by the addition of an optical fiberbackbone might provide today’s cable TV operatorsa significant performance and cost advantage overpotential competitors (such as the telephone compa-nies) in distributing HDTV in the near- to mid-term.23 Such a system might not easily evolve intoa two-way broadband system supporting interactiveservices as a fully fiber-baspd telephone systemwould.
A broadband telephone network is not likely to bewidely available for 20 or more years, however, andsome believe that an integrated cable TV-telephonenetwork may provide some of the advantages of aninteractive broadband system in the interim.wTtiay’stelephone network using N-ISDN can provide suffi-cient capacity for many information services (but nothigh-quality video) and the coaxial cable networkcan provide one-way distribution of high-qualityvideo entertainent, possibly even called up overthe telephone network. Whether such a cable/telcosystem is an evolutionary dead-end that slows orprevents the development of a fully switched broad-band network by skimmm“ g off the most lucrativeservices or is an important interim step in thedevelopment of such a broadband network must becarefully considered.
COPPER TELEPHONE LINESOften referred to as “twisted pairs,” the tele-
phone lines leading into nearly every residence and
business are the backbone of the telephone system.A number of twisted pairs are combined into a“cable system,” and these cables may be bundledinto several rope-like “binder groups” (screenedcables). These are currently being replaced byoptical fibers between large central telephone switchesand in high-traffic sections of the long-distancenetwork. The economies resulting horn the immensecarrying capacity of optical fiber are driving thesubstitution of fiber for copper. Fiber is on thethreshold of being cheap enough to replace copper tothe home in new, high-density residential develop-ments. As many as 250,000 homes are projected tobe comected by fiber by the year 1992.X Itspenetration into existing residences and low-densityareas will be slower because of its higher cost inthese applications.
The information carrying capacity of a twistedpair is limited to typically just a few hundred kHz toa few MHz, depending on the distance betweenanalog or digital repeaters along the line. Thebandwidth for an analog (voice) signal on a tele-phone is much more limited still. It is normallyspecified to range from about 0.3 to 3.1 kHz. Thisrange is restricted so as to limit the amount ofinformation transmitted by the many calls beingcarried at any one time in certain portions of thetelephone network, such as the microwave linksused for long-distance calls. This increases thenumber of calls that can be handled.
The capacity of a twisted pair for transmittingdigital data varies considerably according to theefficiency of the encoding, modulation, and multi-plexing processes, and to the amount of noise andinterference on the line. Telephone line modems fortoday’s analog system with transmission rates of 9.6thousand bits per second (kbps) are now common
mA r~ent poll of cable television MSO and System managers found tbat two-thirds expect fiber to replace coaxial cable in their networks. “FiberOptics Handboo~” Cablevision, Apr. 24,1989.
ZlRo~ti PepPm, ~~~ough the ~o~g G~s: ~tegrat~ Brotitid Networks, Re@atory Policies ~d ~StihItiOXld -e,” 4 FCC Rcd 13M~1307 (November 1988); U.S. Department of Commerce, National Teleummmrdcations and Information Administration, Telecom 2000, NTIA SpecialPublication 88-21, October 1988, p. 268.
=tfiel& op. cit., footnote 19; David P. Reed and Marvin A. Sirbu, “Integrated Broadband Networks: The Role of the Cable Companies,” 1989Telecommunications Policy Research Conference, Airlie House, Airlie, VA, Oct. 1-3, 1989; “Fiber Optic: Backbone for CA~,” IEEE Spectrum,September 1988, p. 26. Note that the various estimates of cost correspond only to within a factor of 2.
~BmceL. Egan~d Dough A. co- “capi~ Budgeting ~te~tives fOrR~iden~ Bro~b~d Networb,’ Center for Telecommunications andInformation Studies, Graduate School of Business, Columbia University, October 1989.
~John Sic, Telecommunications, Inc., “Proposal for a Telco/Cable Interface,” paper presented at the Telecommunications Reports’ Conference onTelco-Cable, Washingto~ DC, Sept. 22, 1989.
~John ~koff, “Here Comes the Fiber-Optic Home,” New York Times, NOV. 5, 1989, sec. 3.
Chapter 3-Communications Technologies ● 47
and rates twice this are becoming available.xAnalog transmission over twisted pairs will continueto be the principal technology for the midterm,although Integrated Services Digital Network (ISDN)are gradually being installed in some areas, primar-ily for business service.
ISDN can carry voice, data, and video informationover switched telephone networks in a digitalformat. It is a set of transmission, switching, andsignaling technologies that support advanced digitalsemicesozT ISDN was designed to use existingsubscriber loops consisting of copper pairs fordigital transmission. The Basic Rate Interface (BRI)of 144 kbps spec~les two 64 kbps channels forsubscriber use and a 16 kbps channel for signalingand some slow-speed digital subscriber service.
Most homes today are wired with two pairs oftwisted copper lines. When using the full capacity ofN-ISDN, these two pairs will be able to carry datarates of 1.544 Mbps.28 This rate is sufficient for avariety of interactive digital computer informationservices, but is not able to deliver real-time, high-resolution video services. By “compressing’ thedigital signal (see box 4-2), it is possible to transmita modest quality video signal for teleconferencing.29For conventional TV, ATV, or HDTV, highercapacity transmission media are required, e.g.,coaxial cable or optical cable.
The public telephone network could bean impor-tant medium for delivering HDTV to the home. Thegoal of “universal service’ set forth in the Commu-nications Act of 1934 is nearly a reality. It permitspoint-to-point routing for voice services through aswitched network30 to nearly all residences andbusinesses in the United States. It provides two-waycommunication, which could bean important attrib-ute should HDTV evolve into an information
‘‘appliance’ for the interactive exchange of infor-mation.
Large investments will be needed if the telephonenetwork is to provide the bandwidth and switchingnecessary for carrying real-time, high-resolution,full-motion video. If this is to happen, optical fibertechnology and optoelectronic switching deviceswill hold the key.
OPTICAL FIBER CABLESFiber optics technology enables inforrnation—
voice, data, and vide~to be transmitted as lightpulses through glass fibers. Optical fibershave manyadvantages over cables and wires. These includewider bandwidth and longer spacings betweenrepeaters, lower weight, immunity to electricalinterference, higher reliability, and lower mainte-nance costs.31 Because of this, shielding is notrequired for fiber to meet FCC requirements. Glassfibers do not conduct electricity, thus a lightningstrike cannot send a pulse down a fiber to crippleequipment on the network.
Because optical fibers do not conduct electricity,providing power to the system on the customer’spremises may require using the customers’ powersource. This would likely not be as reliable astoday’s conventional system where backup power isprovided by the central telephone exchange. Usingrechargeable batteries on the customer premisescould provide backup power in the event of anoutage, but might lead to significant maintenanceproblems. Fiber systems are still more expensivethan copper pair systems,32 but costs are comingdown rapidly.
An optical fiber transmission system includes oneor more optical fibers housed in a protective cable,and devices for converting electrical signals intolight pulses and back again. Repeaters are used to
XL. Brett GIass, “Modern Modem Methods, ” Byte, June 1989; Eli No-, “The Political Economy of ISDN: European Network Integration vs.Anerican System Fragmentation%” Apr. 23, 1986, p.5; ‘‘V.32 Modems,’ High Technology Business, November-December 1989, p. 45.
27~ere ~ W. ~en~c fom of ISDN: Narrow band ISDN (N-ISDN) that can c~ voice or data wiw its bandwidth, and broadband ISDN(B-ISDN) with suftlcient bandwidth to carry high-resolution full-motion video but requires coaxial cable or optical fiber in addition to a wide-bandswitching network. See David Hack+ Telecommunications and information-Systernx Standardizatior+Js America Ready 87-458 @%shingto~ DC:Congressional Research Service, 1987).
X’rhe T1 rate of 1,544 ~p~ was chosen based on the typi~ distance betw~n @ole cover%where atiog m digital ~ptXtterS $113 Or Ca13 beplaced-in the current telephone system. Bob Mercer, Hattleld Associates, Inc., personal communication Nov. 13, 1989.
29Ric~d Dohe~, “system PUtS Real-time Squeeze on Color Video, “ Electronic Engineering Times, Feb. 27, 1989.m~y c~c~t-s~tch~ system win k discussed here. Packet-switched systems and otiers are also possible.
slpep~r, op. cit., footnote 21.32J+E. D~er, ~~when DWs Fib M&e Sense?” Business Co-nications Review, February 1989.
48 ● The Big Picture: HDTV and High-Resolution Systems
regenerate the signal over long distances, andmultiplexing can be used to increase the carryingcapacity of fiber links (see box 4-2). Data rates of565 Mbps (8,064 voice channels) on optical fibersare common; rates of 1.76 Gbps are available;33 andexperimental systems have reached 16 Gbps.”Technologies to achieve much higher rates are beingresearched.
The versatility and technological capacity ofoptical fiber makes it superior to copper in manyways. Its bandwidth suits it to the transmission ofHDTV signals as well as high-volume telephonetrunk circuits. Many long-distance lines have beenconverted to optical fibers, but little fiber has beeninstalled in the local telephone loops. A number ofoptical fiber demonstrations to residential subscrib-
ers have been undertaken in selected municipalitiesby the Bell Operating Companies and others,including: Heathrow, Florida (BellSouth); Perry-opolis, Pemsylvania (Bell Atlantic); and Cerritos,California (GTE) .35
The cost of broadband switching equipment andrewiring the local loop to existing buildings willdelay widespread installation of fiber to residences.It may take 20 years or more before fiber becomescommon in households. There me currently fewservices that demand broadband telephone service toresidential subscribers. Telephone companies viewthe delivery of video dial-tone services and thepotential delivery of HDTV to the home as animportant driving force for the installation of fiberover the “last mile. ”
33JoM~n fiaush~, “Fiber Deployment Update, ” FCC, Common Carrier Bureau, Industry ha.lysis DivisioU Feb. 17, 1989, p. 3.34Miyoko Wkarai, “semi ‘Trailblamr ,‘ “ Electronic Engineering Times, Feb. 20, 1989.sspa~ W. Shumate, Jr., “Optical Fibers Reach Into Homes,” IEEE Spectrum, February 1989; Herb Brody, “The Rewiring of Americ%” High
Technology Business, February 1988; Michael Warr, “Southern Bell Switched-video Plans on Traclq” Telephony, May 15, 1989, p. 11.
Chapter 4
TV and HRS Technologies
INTRODUCTIONAs an entertainment medium, HDTV is not
revolutionary. It is simply another step in theongoing evolution of television that began withblack and white (B&W) TV in the 1940s and willcontinue into the future with as yet undreamed oftechnologies. Each successive generation of TVtechnology attempts to provide a more realisticpicture and sound within the constraints of low-cost,easy-to-use consumer technology.
Here we describe conventional NTSC1 television,Advanced Television (ATV) systems, and some oftheir underlying technologies (box 4-l). The conver-gence of ATVS, computer and telecommunicationsequipment toward High Resolution Systems (HRS)is discussed later.
CONVENTIONAL TELEVISION:PRODUCTION, TRANSMISSION,
AND RECEPTIONTelevision systems involve three distinct activi-
ties: l)production, 2) transmission, and 3) display ofthe TV program.
TV cameras begin the process by creating, foreach of the three primary colors (red, green, blue), anelectronic signal proportional to the brightness oflight at each point of a scene. This signal istransmitted immediately or recorded on magnetictape for later use. 2 The quality of the transmittedpicture is determined by the range of fiequencies—‘‘bandwidth’ —that can be used, and by the effi-ciency with which the video signal is encoded andthen modulated onto this bandwidth.3 At the re-ceiver, the transmitted signal is reconverted to a
picture on the TV screen by scanning electron beams(one for each primary color) across the picture tubeand varying their intensity in exact synchronismwith the original picture signal.
The 1950s technologies used today to bring colorTV pictures to the home have a variety of shortcom-ings and imperfections that modern systems cancorrect. TV production formats, established in the1930s and 40s, were originally based on 35-mmmotion picture film. This gave today’s TV picture itsnearly square shape (or aspect ratio) of 4 units wideby 3 high (4:3). Research has found, however, astrong viewer preference for screens 5 to 6 unitswide by 3 units high—as seen in today’s theatres—that correspond to the human field of vision.4
The original motion picture standard was 16pictures per second—manually cranked camerascould go no faster.5 At that rate the viewer saw asignificant ‘flicker’ in the picture displayed (hencethe term, the ‘flicks’ ‘).6 In developing TV transmis-sion standards, engineers sought a system that sentpictures often enough that viewers did not see themflickering, yet did not send so many that the neededbandwidth (information camying capacity) was ex-cessive. The effective picture resolution (or ability toshow detail), its color fidelity, and its range ofbrightness were all limited in order to meet thesebandwidth constraints.
Other shortcomings of today’s NTSC systeminclude: the blurring of bright colors around theedges; the generation of rainbow colors wherebrightness varies rapidly, such as striped shirts; andsusceptibility to “ghosts,” snow, interference fromother stations, and image distortions.
INatio~ Television Systems Committebthe group that formulated the B&W and color TV standards in use kl the United Smtes todaY.2fiowW p~duction ficluda, of come, editing, the insertion of special effects, audio, etc.
sFor tems~~ broadcasters, the available bandwid~ is set by the Federal COmmtiQtiOIIS commission (FCC), which brokers comPe~g c-for the limited radio- fkquency speetrum.
4*,ue Mc.@~ w. Russe~ Ne~~, ~k Repolds, sham 0’Dormell, and Steve Schneider, “me Stipe of ~gs ‘0 come: A S~dY ‘f
Subjective Responses to Aspect Ratio and Screen Size,’ ATRP-T-87, MIT Media Lab, May 17, 1988.5R,ic~d J. Solomon, “Broadband Communications as a Development Probleu” International Seminar on Science, Technology and Economic
Grow@ OECD, June 6, 1989, Paris. This was later increased to 24 frames per second to give better sound qurdity.
%lu-y Lu, “High Deftition TV Comes At A High Cost,” High Technology, July 1983, p. 45.
-49-
50 ● The Big pic~re: HDWand High-Resolution systems
Box 4-l—Inside Today’s NTSC Analog Television
The production of a TV show begins with a TV camera. TV cameras separate the primary colors-red,green, and blu=f the scene with optics and filters and then focus the scene for each color on separate sheetsof light-sensitive material. This material generates an electric charge that varies in size by the amount of lightthat falls on it. The charge on this sheet is then used to control an electron beam that scans across the sheetfkom left to right and top to bottom-just as we read text. By this means, three “electronic” pictures areformed, one for each color, every 1/30th of a second.
The three pictures thus formed consist of 483 active lines vertically. Each line has the equivalent of about427 horizontal picture elements.1 An additional 42 vertical lines and the equivalent of 82 horizontal pictureelements could, in principle, be displayed. Instea& the time is allotted to generating vertical and horizontalsynchronization, or sync, signals. These sync signals note the beginning of each new picture and of eachhorizontal line within the picture.
This electronic picture is encoded into a composite signal, including brightness, color, and sound. Thecomposite then modulates the amplitude of a carrier (corresponding to a particular TV channel) which istransmitted over-the-air, through cable, or by other means. To be compatible with B&W TV, the brightness(luminance) signal transmitted is a complicated mix of the Red, Green, and Blue signals and corresponds tothe relative sensitivity of the eye to these different hues. The “color” signal (chrominance) contains the restof the information needed for the receiver to decode these combined signals back into the primary colors.
At the home, the transmitted signal induces a tiny voltage in the TV antenna or provides it directly viacable. The electronic picture, or video signal, is separated from the carrier and amplified. The same syncsignals used by the camera now tell the receiver when to start scanning the electron beam across the insidesurface of the picture tube for each new picture, as well as for each horizontal line composing the picture. Asthe electron-beam smoothly scans across the picture tube, its intensity is varied according to the brightnessof the image as originally measured by the TV camera, and is thus reproduced on the phosphors of the picturetube. Color TVs do this simultaneously with three different electron-beams, one for each color of phosphor.This type of electronics, where the information content is represented by a continuously varying voltage orcurrent, is known as analog electronics.
This system is much like writing on a blackboard at a distance using a set of strings to pull the pen backand forth. What is written onto the camera is reproduced on the picture tube. If there are errors in thetransmission process, they are likewise written onto the TV screen at the other end. With the pictureinformation carried by a continuously varying voltage, it is very difficult to tell if an error has crept in, let alonecorrect it.
Such a system was all that 1950s electronics technology was capable of. This approach carries with ita variety of impairments, including the use of interlaced scan, errors in displaying colors, an excessivesusceptibility to transmission degradations, and inefficient use of the broadcast spectrum.
In developing TV transmission standards, engineers had to send pictures often enough that viewers didnot see them flickering, yet did not send so many that the needed bandwidth, or information carrying capacity,was excessive. Both of these were accomplished by showing each picture twice using an “interlaced scan’technique. First, the odd horizontal lines (numbered consecutively from top to bottom) of the picture are sent;next, the even lines are sent. In this way, the 525 lines of today’s TV (NTSC) set are displayed half (or a field)at a time, l/60th of a second apart. An entire picture, or frame, is then seen every l/30th of asecond-minimizing bandwidth and flicker simultaneously.
Although the vertical resolution of such a picture could be 483 lines, it is normally much lower than this.As one example, when an interlaced scan system displays an odd horizontal line as black and the even linenext to it as white, there can be an annoying flicker to the picture every l/30th of a second-easily perceivableby viewers. Consequently, the vertical resolution of the studio camera was reduced to, at most, 330 lines (by
ltiog TV d~s not really have discrete horizontal picture elements, but rather a continuous horizontal scan. me tX@v~encehere is based on the scanning times and picture resolution.
Chapter kTVand HRS Technologies ● 51
changing the spot size of the scanning e-beam within the camera) and the horizontal resolution was reducedby roughly the same amount. Progressive scan systems, in which every line is sent, odd or even, in a singlepass down the picture do not suffer this problem of flicker and can thus give much higher effective resolutionsfor a given line count.
The transition in the 1950s horn B&W to color was achieved by reducing the resolution of the brightnesssignal and inserting a color signal as described above. This led to a slightly lower resolution for B&W TVs,among other problems. For color TVs, the manner in which the color and brightness signals were intermixedcould cause bright colors, especially red, to blur at the edges; and places where the brightness varied rapidly,such as striped shirts and checked jackets, to degenerate into rainbow colors.
In the past, the interference between the brightness and color signals was minimized by further reducingthe resolution of the receiver’s circuits-to about 250 lines horizontal resolution. In 1978-79, special circuits(comb filters) were introduced which allowed more of the entire range of brightness to be displayed; in 1984,other special circuits were introduced that allowed the entire range of colors to be displayed with somewhatless susceptibility to the above degradations.
Other errors, or artifacts, in today’s TV picture are also widely noted. Viewed up close, today’s NTSCTV looks like an ant’s nest with the dots and lines “crawling” around the picture. These are again due to themanner in which the color and brightness signals are intermixed and subsequently decoded by the receiver.Under certain conditions, NTSC sets have a tendency to switch between different colors. For this reason, theNTSC system has sometimes been disparaged as Never-Twice-the-Same-Color.
The NTSC system is also highly sensitive to transmission degradation. When the broadcast signalreflects off obstacles, viewers may see ‘ghosts’ and if reception is weak, viewers may see ‘snow’ or noise.
Finally, the NTSC system uses the broadcast spectrum inefficiently. These old electronics technologiesand the use of lowquality home antennas required very strong sync signals to ensure proper operation.Further, the signals transmitted by broadcasters are all in the same format.2 Because of this, TV stations cannotbe operated on adjacent channels within the same geographic region without visibly interfering with eachother. Thus, in the VHF band (channels 2-13) stations are separated by one blank-’ ‘taboo’ ‘-channel; in theUHF band, stations are separated by as many as five taboo channels. These channels have not been usablefor other purposes. Improved ante~as, and P&haps other changes, might enable significant reductions in thechannel and geographic separation requirements at a very small cost. This would enable closer packing of TVchannels and might save spectrum for other purposes, but has not been required by the FCC.
Even within the normal 6-MHz bandwidth of a single channel the NTSC signal makes inefficient useof the available spectrum. Much of the available spectrum corresponding to extreme brightness or the mostvivid colors, for example, is rarely used. It was such a “hole” in the spectrum that allowed the original B&WNTSC standard to be turned into a compatible color system. By slightly reducing the maximum B&Wresolution, all of the color information could be squeezed in. Similarly, the EDTV proposals today offeralternative means of further exploiting NTSC’S poor bandwidth use3 to provide higher resolution and a widerscreen without requiring a wider channel bandwidth.
Although the NTSC system has many shortcomings, it was a remarkable achievement in its day. Theengineers that developed NTSC were working at the very limits of analog technologies in the 40s and 50s,with an overriding constraint to make the receivers as low cost as possible. Their remarkable achievementsresulted in a system that has served us well for 40 years. With today’ ~ technologies, however, we can do better.
References: “TelevisionEngineering Handbeo~” K. Blair Benson (cd.); Trudy E. Bell, “The New Television: Looking Behindthe ‘lWe,” ZEEE Spectrum, August 1984; Ronald K. Jurgen, “High-Deftition Television Update.” IEEE Specfrum, April 1988;personal communications: Andrew Lipp~ MIT Aug. 16, 1989; William Schreiber, MIT Oct. 12, 1989; John Henderson, SarnoffLabs, Oct. 12, 1989.
2T’he sips me tr~fitted in the same raster scanned format as generated by the camera. Consequently, they are cohermt.Sother ~efflciencies h spectrum use by NTSC include: the vertical and horizontal retrace intervals, d- wtich additio~
picture information could be sent the manner in which the carrier is modulated, including the use of a separate sound carrier; and thetransmission of the entire picture in every frame, including redundant information within and between frames.
S2 ● The Big Picture: HDTV and High-Resolution Systems
ADVANCED TELEVISIONSYSTEMS: IDTV, EDTV, HDTV
Modern computer-like digital electronics canreduce many of the defects found in today’s NTSCsystem. Three levels of improvement in televisiontechnology are either now under advanced develop-ment or are entering commercial markets.
The fiist level is Improved Definition TV(IDTV). IDTVreceivers take the standard broadcastand convert the continuously varying, analog volt-age of the NTSC transmission into a digital signal.This digital signal represents the information con-tent of the picture as a series of numbers spec@ngthe color and brightness of each point in the image.A simple example of analog-to-digital conversion isshown in box 4-2.
The advantage of converting the analog broadcastto a digital signal within the receiver is that thepicture information can then be manipulated (modi-fied, adjusted, corrected, etc.) by digital signalprocessing techniques and/or stored in memory.This is similar to the way computers manipulate andstore data. Special techniques can be used to reduceflicker, ghosts7, snow, and other picture imperfec-tions, and to improve apparent resolution and color.NA Philips (Netherlands) began marketing IDTVSin the United States in the spring of 1989.
Extended Definition TV (EDTV) is tie next stepin the evolution of TV technology. It requiresmodest improvements in today’s TV transmissionsand small changes in the broadcast standards. Theimproved TV signal will still be compatible withtoday’s TVs, and it will not carry so much additionalinformation as to require greater bandwidth than isavailable in today’s 6-MHz TV channel.
EDTVS will use digital memory and signalprocessing techniques. In most proposals EDTVprovides a widescreen picture and somewhat higherresolution than is now possible. Japan was sched-uled to begin EDTV broadcasts in 1989, and Europeplans to do so in 1990. Some believe that theimprovements in TV picture possible from IDTVand EDTV will significantly reduce the futuremarket for HDTV (ch. 6).
High Definition TV (HDTV) is the third majorstep up from today’s NTSC system. HDTV istypically portrayed as having roughly twice thevertical and horizontal resolution as is theoreticallypossible for current TVs; a widescreen picture(aspect ratio of at least 5:3) displayed on a largescreen; better color; and compact disk (CD)-qualitydigital sound. The much higher quality pictures ofHDTV, however, will require the transmission ofsubstantially more information than current TVsystems. The original Japanese NHK productionsystem, forexample, required 30 MHzofbandwidth-the equivalent of five TV channels. To reduce this tomanageable proportions, a number of tricks are used,beginning with an analysis of what the human eyecan actually perceive.
VISUAL RESPONSE ANDHIGH DEFINITION TV
Human vision responds rapidly to the grossdetails of bright or moving objects, particularly inperipheral vision, but does not perceive their finedetail or color well. Conversely, to perceive finedetails or color requires the object to be stationaryfor longer periods+r for the eye to track a slowlymoving object and ‘‘freeze” it on the retina. Thus,efforts to avoid flicker by transmitting all of thecolor and detail of both moving and stationaryimages at the same maximurn rate, i.e., 60 picturesper second (as an analog TV would), are unneces-sary; the eye cannot perceive all of it.
The groups developing HDTV systems have usedknowledge gathered in experimental work on visualperception to reduce the picture information thatmust be transmitted. Digital processing techniquesare used to send the picture in low resolution at arapid rate so that rapidly moving portions of thepicture are seen quickly; and to also send the picturein high resolution and full color at slower rates sothat the still portions can be seen in detail.g The TVreceiver then uses digital signal processing tech-niques to reconstruct the picture in memory, fromwhich it is flashed onto the screen 60 times a second.
These techniques reduce the bandwidth requiredto transmit the TV picture to a fraction of that of the
7Ghost c~ce~tion usually requires adding a speciid test sigmd to the ~“tted signal.8T0 ~mfit this ~omtion my r-e a Combfition of water b~dwidth, much mom efficient enc- and modulation tahlliques, VWiOUS
digital compression techniques, and good signal to noise ratios, among others.
~illiam F. Schreiber and Anchew B. Lippma~ “Bandwidth-Efficient Advanced Television Systems,” MIT Media Lab, no date.
Chapter 6TV and HRS Technologies ● 53
Box 4-2—Inside Analog/Digital Television
The various proposals for advanced TV systemsuse combinations of analog and digital techniques toimprove the picture seen on the receiver. The processof generating a picture is described in box 4-1. Thisprocess results in an analog signal coming from thecamera-that is, with a continuously varying voltageor current proportional to the brightness of each pointof the scene.
Selected portions or all, as desired, of the analogsignal from the camera are then converted into a digitalformat for storage and/or transmission. In some cases,such as IDTV, the analog to digital conversion is onlydone in the receiver. A particularly simple example ofanalog to digital (A/D) conversion is Pulse CodeModulation (PCM).
The PCM process begins by sampling the contin-uously varying analog signal, as shown in figure 4-1.The sampling rate is usually at least twice the highestfrequency of the analog signal. The result is a series ofpulses whose amplitudes correspond to the analogsignal at regular, discrete times.
These pulse amplitudes are next converted intonumerical values by comparing them to a set ofdiscrete amplitude steps. This does introduce errors,depending on how many discrete steps are used andhow well the pulse amplitude matches the closest step.This error can be reduced by increasing the number ofsteps, but at the same time increases the circuitcomplexity and the amount of numerical informationthat must be transmitted.
The precision of this quantization is normallygiven by the number of steps used and is commonlycited in terms of the “bits” of binary information thatresult, each bit representing an additional power of 2
Figure 4-1—AnaIog to Digital Conversion Using PulseCode Modulation
B 1 [— —I I 1
II I 1’
1Amphtude
modulated 1I I I I 1— p u l s e s _
II 1 I I 1II I— .-—
001 001 111 000 101
c -J.... _nnl__.._.JMlL____nculA. The analog signal is sampled at regular intervals.B. The amplitude of these samples is then compared to a set ofdiscrete amplitude steps and converted into numerical vatues.These values are shown written in binary.C. The numerical values can then be transmitted in a digitalformat. Information from other signals can be sent between thesesamples in a time division multiplexed format.SOURCE: Kamilo Feher, Advanced Digita/Communic40ns: Systems and
signa/ Processing Techniques (Englewood Cliffs, NJ: Prentice-Hall, Inc., 1987). Used with permission.
in the number of steps. Thus, 2 bits equals 4 steps; 3 bits equals 8 steps; 4 bits equals 16 steps; and so on. Some256 steps, or 8 bits, are usually sufficient to encode the brightness information for each primary color—red, green,and blue-of a studio TV signal. This is a total of 24 bits.
The binary information representing the picture can now be stored in a computer-like memory, manipulated,or transmitted with little degradation. Each “bit” of information, for example, can be transmitted as either “on”or “off.” At the receiver, even if the amplitude of “on” was greatly reduced or distorted due to transmissionproblems, the receiver can usually still detect that it was ‘‘on’ and correct its value to the nominal level. In contrast,an analog receiver normally has no way of knowing whether or not a signal is distorted. It can only reproduce whatit receives, errors and all.
In two-way transmissions of data, such as by a modem, a digital signal can also include information that tellsthe receiver how to check for errors in the transmission. For example, eight bits of information could be sent witha‘ ‘parity” bit. The “parity” bit is “on” if the number of 1s” sent is even; “off” if the number of “1s ” sent isodd. The receiver can check to see if the number of 1s’ received and the parity bit correspond. In this way, at leastsome of the transmission errors so serious as to cause “on” to appear “off” and vice versa can be detected. Morecomplex techniques will be used to reduce the number of errors in transmitting HDTV pictures and to ensure thecorrect transmission of critical components of the signal.
(continued on nexfpage)
54 ● The Big Picture: HDTV and High-Resolution Systems
Box 4-2—Inside AnaloglDigital Television-Continued
The basic digital signal for an HDTV picture carries an enormous amount of information-as much as 1.2billion bits per second (1.2 Gbps). In comparison, a phone conversation uses less than 64 thousand bits per second(64 kbps). A personal computer uses just 16,000 bits of information internally to represent a screenful of text and,on a standard monochrome monitor, just 250,000 bits are needed to display this text—which often remainsunchanged for long periods while one ponders what to write next. 1 To transmit the total 1.2 Gbps of the HDTV signalrequires far more bandwidth than is available in all but the most advanced fiber or short-haul cable systems. Thus,digital signal compression techniques are used to make transmission to households practical.
Digital signal compression eliminates the redundant information in a signal. For example, a large expanse ofblue sky in a picture carries little or no new information from one point to the next. Special “intrahrne”compression techniques effectively tell the receiver that these picture elements are the same, and allow therepetitious “blue sky,” “blue sky,” . . . information to be eliminated.
Similarly, the stationary portions of a picture do not change from one frame to the next. There is no need totransmit such redundant information 60 times per second in full detail. In fact, if nothing moves no new informationneed be transmitted at all. In this case, one frame is compared to the next to do “interframe compression’ and onlythe differences are transmitted. Interframeprocessing, however, may result in other types of errors such as inaccuratedepiction of moving objects.
As already noted, the eye is least sensitive to the color and detail of those parts of the picture which are moving.This information can be reduced simply by sending the rapidly moving portions at lower resolution.
Together, these techniques greatly reduce the information content of the transmitted signal and, in fact, makethe transmission of HDTV pictures practicable. The Japanese MUSE HDTV system, for example, reduces the signalbandwidth horn 60 to 8 MHz using such techniques.2 Digital systems might reduce the information content fromsome 1.2 Gbps to 70-80 Mbps. Subsequently, the transmitted signal is decoded and turned back into a viewablepicture in the receiver.
In some cases, it may also be desirable to fit additional information into the HDTV signal. This is illustratedin figure 4-1. This technique is known as “time division multiplexing” and is especially valuable in cases wherethe transmission media is filled close to capacity or is being used inefficiently.
TVs that use analog electronics alone cannot do this type of image manipulation. Because analog signals varycontinuously, it is impossible to do such signal compression before transmission, to catch and correct transmissionerrors, to recreate the picture in memory at the receiver, to multiplex the signal, or to do a host of other advancedsignal manipulations, Instead, purely analog systems are constrained to simply write the picture as transmitteddirectly on the screen, like chalk guided by a string.
References: Kamilo Feher, “AdvancedDigital Communications”; “ Television Engineering Handbook” K. Blair Benson (cd.); WilliamSchreiber, various papers, op. cit.; Submissions to the FCC, various. Dale N. Hattleld, “Report on the Potential for Extreme BandwidthCompression of Digitized HDTV Signals,” Mar. 20, 1989; “Television Technology Today, ” Theodore S. Rzeszewski (cd.). Other referencesas in box 4-1.
1A screefil of text has 25 rows by 80 characten per row and 8 bits per character when handled as ASCII text intetily. A s@ndardgraphics based monochrome monitor has 720 by 348 pixels, each either “on’ or ‘off.’ Assuming that the text on the screen remains unchanged,this requires 250,560 bits to represent. If the text was being changed at 60 times per second, the data rate would be 15 Mbits per second.
Zllis total is for the component signals-30 MHi for lumina rice, 15 MHz for Red, and 15 MHz for blue. Larry Thorpe, Sony Corp.,personal communicatio~ Oct. 12, 1989.
basic unprocessed analog signal. Only the use of decode the signal. In large-scale production, today’sthese and other (box 4-2) digital signal processing semiconductor technologies can minimize thesetechniques makes HDTV practical. Such digital added costs.signal processing, however, comes at a price. In It is uncertain whether or not such visual tricksgeneral, the greater signal compression required to will work as well on the large screen that will likelyreduce the bandwidth, the greater the complexity be used for HDTV. Some note that, especially onand cost of the receiver’s electronics needed to large screens, the eye is able to track relatively
Chapter 4+TVand HRS Technologies ● 55
slow-moving objects and freeze them on the retina.In such cases, the digital signal processing mightremove visual informah“on that the eye does Pemeivbreducing the visual quality of the picture.
The most important attribute of an HDTV may bethe much larger field of view it provides. The regionof greatest visual acuity of the human eye is withinthe central one degree of the visual field. Viewerstend to move away horn the screen to view movingobjects in order to maximize the ii-action of theimage within this central one degree field, and stopwhen the details of the image begin to be lost. Theideal viewing position for an NTSC set (525 lines)is experimentally found to be roughly 7 screenheights fkom the receiver, providing the viewer an 8by 11 degree field of view. For the typical HDTVsystem (1125 line Japanese system), the preferredviewing distance is found to be about 3.3 screenheights away, providing the viewer a 17 by 28degree field of view. These are compared in figure4-2.
Thus, viewers position themselves for roughly thesame resolution on the eye, but they have a muchlarger viewing angle with HDTV, contributing to asense of “telepresence.”lo Although the higherresolution and picture quality of HDTV would bevisible on a smaller screen TV, there would be lessincentive to purchase such an HDTV. It would notprovide the same sense of telepresence, and manypeople would find it difi3cult to focus on orexperience discomfort if close enough to benefitflom the higher resolution.
“Telepresence’ or the sense of “being there” isa potentially powerful market inducement for largescreen, high resolution displays. An example of thissense might be the feeling, when viewing a veryhigh-quality motion picture screen up close, ofmoving with the airplane or roller coaster when itmakes a fast turn, or of unconsciously putting yourfoot on the brake to prevent an accident. In watchingsports, it allows one to have a greater sense of beingin the stadium itself-being able to watch clearlyand in detail the entire scene of a quarterbackcompleting a pass to a receiver, rather than relying
Figure 4-2-Comparison of the Field of View foran HDTV and Today’s NTSC TV To Have the
Same Resolution for the Viewer
I ‘N’. \/
\\
SOURCE: William E. Glenn and Karen G. Glenn, “High Definition Trans-mission, Signal Processing, and Display,” paper presented atthe SMPTE Conference, San Francisco, CA, January 1989.Used wfth permission.
on instant replays of the separate tight closeups ofthe quarterback and receiver that are used today.
HDTV SYSTEMS: PRODUCTION,TRANSMISSION, RECEPTION
The key attributes of HDTV-twice the resolu-tion of today’s television, a wider screen, truer color,and digital sound-will require significantly moreinformation to be transmitted than is currently thecase. This cannot be done while simultaneouslyremaining compatible with today’s TVreceivers andcharnel bandwidths. Significant changes are neces-sary in production, transmission, and receptionstandards and equipment to achieve such highdefiition pictures.
lowil~ E. G]e~ “H@ Defiition Television,” Society for Information Display, 1988.
30-368 - 90 - 3 : QL 3
56 ● The Big pic~re: HDW and High-Resolution Systems
Production Standards
Today’s production standards are based on 24hrne per second motion picture f~l or on videotape for television. Although the different worldstandards for color TV—NTSC (USA, Japan),SECAM (France, Eastern Europe), and PAL (West-ern Europe)-vary in their frame rate and number oflines scannedll, equipment to convert horn onestandard to another (transcoding) with little loss inquality has been technically quite successful. Themost dfilcult aspect of transcoding is to convertbetween formats with different ftame rates. Thisrequires interpolation between pictures in order toshow motion smoothly. The conversion from inter-laced scanning to progressive scanning has also beendifficult.
Lower-cost and more flexible electronic videoproduction systems will become increasingly impor-tant in coming years. Such systems will competewith fti for many applications and complementcinema photography in cases such as special effects.
Each medium-NTSC, HDTV, and fti-has itsown artistic “look and feel” according to itstechnical characteristics and format. NTSC videotape, for example, can generate the afternoon soapopera’ look under various conditions. This is due tothe sensitivity and response characteristics of theelectronic camera used. The low resolution of NTSCsystems also forces the program producer to empha-size closeups or tight shots of the actors.
In contrast, the greater sensitivity, color, andbrightness range of film and the better match offilm’s response to that of the human eye gives ftia much more natural look for scenery; its higherresolution allows the producer to step back and usemore wide-angle views. These characteristics givedifferent artistic feels to the media, which are likelyto persist well into the HDTV era.
In the longer term, however, electronic capabili-ties are likely to bring video media closer to thecapabilities of film. In addition, electronic mediaoffer advantages over fti by eliminating processingand allowing, in the filly digital case, an endless
number of overlays of special effects or addition ofother material. These potential new applications andcross-technology interactions highlight the impor-tance of production standards.
Although it is now unlikely that a single worldstandard for HDTV production will be adopted (ch.2), international program exchange standards mayyet be developed that allow easy transcoding hornone system to another.12
Transmission Systems
HDTV transmission systems must take into ac-count a variety of media as well as different degreesof compatibility with existing receivers and channelrequirements. Although the U.S. communicationsinfrastructure includes cable, satellite, VCRs, andwill someday encompass fiber to the home, terres-trial broadcasting has been the major focus for U.S.HDTV development. This is due to its marketimportance and the constraints it faces in spectrumallocation. The technical requirements for terrestrialbroadcasting are also significantly different than forthe Direct Broadcast Satellite systems approachfollowed by Japan and Europe.
The FCC has received nearly 20 proposals forterrestrial transmission standards, although only ahandful of these appear likely to be developed.These proposals encompass such issues as: thedegree of compatibility with existing NTSC receiv-ers and the current channel bandwidth; the means ofadding additional information needed to improvepicture resolution, including the video encodingtechnique; methods of widening the picture; and themanner in which the picture is scanned-interlacedor progressive. 13 ~ general, the proposals Ca k
grouped in the following four categories: NTSCReceiver-Compatible HDTV; Channel-Compatibleor Simulcast HDTV; Alternative Media HDTV; andNon-Compatible Recorded Media HDTV.
In September 1988, the FCC tentatively ruled thatbroadcasters must continue to transmit signals thatcan be received on today’s TVs; and that noadditional radio spectrum would be allocated forHDTV broadcasting outside the existing spectrum
ll~sc shows 59$94 fields ~r s~ond in a 525 he in~r~c~ scam or @v~~fly, about ’30 fr~es pa ~nd. SECAM (Squentkd Color withMemory) and PAL (Phase Alternation by Line) show 50 fields or 2S frames per second at 625 lines per frame.
lw.F. sc~i~, ‘CA Friendly Family of Transmission Standards for All Media ~d All Frame wta” (~‘dge, MA: Massachusetts Institute ofTechnology, Feb. 12, 1989), draft.
13See, for example: “FCC Advisory Committee on Advanced Television Service: Systems Subcommittee, Interim Repor4° Apr. 10, 1989.
Chapter &TVandHRS Technologies ● 57
allocations for TV.14 Together, these effectively ruleout noncompatible HDTVS such as the originalJapanese MUSE system, but would allow either thereceiver-compatible or simulcasting approaches forterrestrial broadcasting (box 4-3).15
Today’s TV audience views a picture far worsethan what is theoretically possible, and might easilymistake a studio quality NTSC picture for HDTV.NTSC is susceptible to a variety of transmissionproblems-ghosts, snow (noise), interference fromother stations, etc. (box 4-l). Further, NTSC doesnot use the available spectrum efficiently: moreinformation could be packed into the existingbandwidth; and large amounts of spectrum areunusabl~very other channel in VHF and typi-cally five out of six channels in UHF are not usedbecause of interference problems. To the extent thatEDTV and HDTV require compatibility with theexisting NTSC system, they could lock in thesetechnical flaws for many years in the future. Betterantemas and modem electronics can reduce some ofthese problems. Development of a new system,however, might achieve far more.
NTSC Receiver-Compatible HDTV
The transition from B&W TV to Color TV wasdone by altering the NTSC signal to include colorwhile degrading the performance of B&W setsreceiving color broadcasts only slightly-thus maint-aining compatibility. This is also the most frequentproposal for making the transition to HDTV. Theinefficiency of the current NTSC signal, however,means that only modest improvements can beachieved in the picture while staying within thecurrent 6-MHz channel bandwidth. (The 6-MHz,NTSC-compatible HDTV proposals are here termedEDTV.)
HDTV quality and fill receiver-compatibilityrequire additional spectrum. The full 6-MHz chan-nel (with some motilcations) plus an additionalone-half or full channel (3- or 6-MHz) someplaceelse in the spectrum must be used to augment thesignal and to provide the additional detail needed fora high-quality picture. Standards proposals of thistype include those from: FaroudjaLabs (U. S.), JapanBroadcasting Corporation (NHK),MIT(U.S.), North
Box 4-3—Levels of Receiver Compatibili~
o.
1.
2.
3.
The current receiver is unable to display, in anyform, HDTV transmissions; the HDTV receiveris tiable to display NTSC transmissions.An adapter box can be purchased separately sothat existing NTSC receivers can display HDTVtransmissions at the NTSC level of quality;HDTV receivers can display either HDTVtmnsrnissionsat high definition orNTSCtransmis-sions at NTSC quality.An NTSC system can display HDTV trans-missions at somewhat reduced quality comparedto conventional transmissions. This is whathappened in the conversion from B&W to colorbroadcasts. HDTV receivers, as above.An NTSC can display HDTV transmissions atthe same quality as a normal NTSC transmis-sion; HDTV receivers, as above.
American Philips (Netherlands), Samoff Labs forThomson (France) and NBC.lG
Channel-Compatible or Simulcast HDTV
An HDTV-quality picture could be broadcastwithin the current 6-MHz channel bandwidth, butonly if the constraint of NTSC-compatibility isremoved so that the bandwidth can be used moreefficiently. In this case, a standard NTSC signalwould be transmitted with no changes on onechannel; an HDTV signal would be simulcastindependently on another channel. With properdesign of the signal, now unused “taboo channels”could likely provide the needed broadcast spectrumfor such simulcasts. Over a long period of time, theconventional NTSC channels could be graduallyphased out and replaced with the simulcast signalalone. The portion of the spectrum vacated couldthen be used for the next generation of televisiontechnologies or for other uses such as mobilecommunications.
Zenith and MIT are the principal proponents ofsuch systems, with NHK (Japan) also recentlyoffering a simulcast system and North AmericanPhilips announcing that they are developing one.
58 ● The Big Picture: HDTVandHigh-Resolution Systems
These latter proposals indicate the increasing atten-, tion and interest in the simulcast approach.
Alternative Media HDTV
These systems are designed for use with mediaother than temestrial broadcasting. They are primar-ily oriented toward satellite broadcasts. NHK (Japan),North American Philips (Netherlands), and Scien-tific Atlanta are among the proponents of suchsystems.
Noncompatible Recorded Media
HD-VCRS or HD-Video disks in the noncompati-ble Japanese MUSE format might be introduced inthe United States as early as 1991. These systems donot require FCC approval, but are a concern tobroadcasters because they might accelerate theongoing erosion of broadcasters market share. As aresult, these media have been an important drivingforce behind the current FCC study of advancedtelevision.
Receivers for HDTV
The various proposals for HDTV receivers re-spond to the current debate over transmissionstandards-speci.flea.lly, whether the transmissionsystem will be receiver-or channel-compatible, andwhether there will be one standard or many for thedifferent media. Receivers now envisioned can beclassified as “closed,” “multiport,’ ‘‘open,’ or‘‘smart.
Closed Receivers
Closed receivers are similar to those used today.They are possible only if a single standard is setindustry-wide for all media. Such a system has noflexibility to allow future changes in broadcaststandards or to allow later options to be added to theHDTV without substantial modifications to thereceiver.17
Over the near-term, these systems might costsomewhat less than more flexible designs discussedbelow. The current rapid pace of technologicalchange might, however, make closed systems quicklyobsolete-increasing the costs to consumers overthe longer term.
Multiport Receivers
Multiport receivers will have multiple jacks orinputs that could accept several incompatible sig-nals: one for terrestrial broadcasts, a second forcable, a third for DBS, and so on. Such HDTVSwould have less flexibility to adapt to fiture changesor to allow the addition of various options than thosedescribed below. The costs of multiport (and other)receivers may be increased if they must accommo-date several radically different standards.
Open Architecture Receivers
These receivers would be designed to accept avariety of standard plug-in circuit boards. The setscould be moMled over time (at a price) to accept awide variety of signals and standards, yet properlydisplay the picture. These systems migM also beadapted to provide other services-such as homecomputing and telecommunications-just as today’spersonal computers can have circuit boards added toincrease their versatility and power. Open Architec-ture Receivers might also create new businessopportunities for third-party vendors of equipmentand services.
Opponents of the Open Architecture Receiverargue that this approach would increase costs andcomplexity for users. Proponents point out that therapid changes in technology demand flexible opensystems and that such systems may ultimately loweroverall costs to users. Nor are such systems neces-sarily complex. A simple channel selector andvolume control like today’s can be provided forthose who want only to watch TV.
“SmartJJ Receivers
Smart receivers are the most technologicallyadvanced receivers currently conceived. They wouldadjust to a wide variety of transmission standards byautomatically decoding the transmission format.Such sets could even adjust to a format that variedaccording to the type of material displayed—scenery with little motion could be shown in veryhigh resolution, whereas rapid action sports wouIdemphasize the display of motion.18 Smart receiverswould probably be more expensive for the near term
ITTheorig~~SC stidmd~~nupgraded overtime-with color, stereo, closed captioning, etc. However, the rapid changes incommuni@tionstechnologies and electronics capabilities suggest that much more dramatic changes may be possible in the near future.
lgwil~ F. sc~ei~r ad ~&ew B. Lipp~ “single.kel ~~ systems, compatible and NonCorn@ible,’ hhs~ch~tts hlstihlte OfTechnology, Mar. 11, 1988.
Chapter &TVand HRS Technologies ● 59
due to the high cost of the electronics required forsuch advanced data manipulation.
HIGH-RESOLUTION SYSTEMSAND TECHNOLOGIES
The evolution of television technology towardsdigital electronics is converging with the evolutionof computing towards multi-media presentations(including text, data, graphics, and full-motionvideo) .19 The underlying storage, processing, trans-mission, and display technologies of high-resolutionentert ainment video and multi-media computing arebecoming largely indistinguishable.20
All of these technologies which emphasize high-quality imaging or displays can generically betermed High Resolution Systems (HRSS). HRSSencompass an enormous range of markets-ATVsgenerally and HDTV in particular; multi-media PCs;engineering workstations; scanning and imagingequipment for electronic document storage; digitalphotocopiers and facsimile machines; and manyothers.
The applications of HRSS are equally diverse.These could include: video entertainment (HDTV),interactive video systems, electronic imaging for
document storage21, desktop publishing, graphicsfor engineering workstations, and many others.Many foresee these markets merging. For example,an HDTV/PC might be able to receive or record anHDTV program, search a video database, or includea video clip in a report along with text and graphics.Such systems are potentially interactive, allowingthe HDTV/PC viewer to request additional informa-tion of a news clip, or even to mod@ the picturebeing watched.
HDTV plays a particularly important role in thedevelopment of High Resolution Systems. HDTVSmust handle enormous information flows and sorequire immense processing power. These demandsmake HDTV a sign~lcant driver of certain technolo-gies (ch. 5). The high demand for video entertain-ment may aid the penetration of these powerfulsystems into the residential market. This mightallow the HDTV/PC of the future to become thehome terminal-an ‘information appliance’ ‘-on anational fiber network. It is this vision of theinformation society of the future that is beingpromoted by MITI and MPT (ch. 2). Whether or notHDTV can fulfill such a role in the United States isa matter of public debate (ch. 6).
lg’’Moving Beyond VGA,” Electronics, July 1989.
2oThere are, of course, differences in terms of the signal transmission formats, receiver processing, and the brightness and size of the display, ete.
Z1’’Imaging,” Electronics, July 1989.
Chapter 5
Linkages Between HDTV, HRS, and Other Industries
INTRODUCTIONHDTV is possible only through the intensive use
of digital electronics; signiilcantly higher qualitypictures than today’s NTSC cannot be delivered tothe home by any other means because of bandwidthconstraints. As a result, the core technologies ofHDTV-production, storage, transmission, process-ing, and display of information-are the same asthose used in computer and telecommunicationsdevices.
It is often overlooked in the current debate thatHDTV is a development vehicle for “High-Resolution Systems” (HRS) generic to all informa-tion systems. HDTV proponents, for example, arguethat the ability to produce high-performance dis-plays and other technologies gained from a presencein the HDTV market will give manufacturerssignificant advantages in producing related compo-nents and systems for the computer and telecommu-nications markets. Skeptics resist the linkage argu-ment as an unproven hypothesis and insist that thesetechnologies can be developed by the computer andtelecommunications industries independently.
OTA found evidence that HDTV developmentsare driving the state-of-the-art in several of thesetechnologies more rapidly than are developments incomputer or telecommunication systems. The enormousamount of information in a real-time, full-colorHDTV signal-some 1.2 billion bits per secondl(1.2 Gbps) in the uncompressed signal-placessevere demands on today’s technologies. This con-trasts sharply with the conventional stereotype ofconsumer electronics as low-technology productslagging far behind the leading edge of computers andtelecommunications.
HDTVS must handle huge information flows andrequire special hardware to provide high computa-tional speeds to convert a signal compressed fortransmission back into a viewable picture. DigitalSignal Processors (IMP) tailored to this specific task
are meeting that requirement. In contrast, engineer-ing workstations, for example, must flexibly per-form a broader range of calculations than an HDTV;therefore, they are software programmable.2
The major technological bottleneck for the work-station today is the computational speed and theflexibility of its microprocessor and graphics displaychips. Workstations put less stress on communica-tions, storage, and certain aspects of display technol-ogies than HDTV because they do not yet approachthe information flows or the (specialized) computa-tional speeds demanded of HDTVS.
Other High-Resolution Systems (HRS), such asdesktop publishing and medical imaging, placedifferent and often lower demands on DSP, storage,and communications technologies than does HDTV.They typically do not operate in real-time andaccordingly have lower rates of information flow.Many of these other applications also place lowerdemands on display technologies than does HDTV.They work well enough with slower response times,limited colors, lower brightness, or smaller displayareas. Providing sharp images of stationary objectsas is usually the case with computer applications is,in many respects, technologically easier than provid-ing high-resolution, real-time, fill-motion video.These other HRSS often require, however, higherresolutions than currently planned for HDTV dis-plays.3
While many of the linkages between these tech-nologies are obvious, they are not easily measurable.Nevertheless, these linkages can have an enormousimpact on widely scattered technologies and mar-kets. Simple analyses in which the projected fhturevalue of an industry is discounted to the presentcannot account for the new and unforeseen opportu-nities that might be created by being in a market.Sony’s and Philips’ development of the compactdisk player, for example, has opened huge marketsin computer data storage. Similarly, flat panel
1’IMS is for the NHK system. Other systems might have somewhat higher or lower data rates.
?Note that “smart” HDTVS would have the capability to be somewhat software programmable, but are not incIuded here as they are not likely tobe sufficiently low in cost for consum er use for some time. Their development will push the state-of-the-art significantly in programmable DSP.
sB~use ~omputer ~sp~ys me nom~ly viewed closeup and fiere are eye fati~e issues, some of the design criteria are different.
-61–
62 ● The Big Picture: HDTVand High-Resolution Systems
plasma and electroluminescent displays, amongothers,4 drove the initial development of PowerIntegrated Circuits (Power ICS) which, in turn, arerevolutionizing the distribution and control of elec-tric power in equipment ranging from aircraft toair-conditioners.
The linkage argument does have limitations. Forexample, unlike leading-edge PCs or workstationswhere performance is everything and price a second-ary consideration, HDTVS must be produced andmarketed at a price within reach of consumers. Thisdemands exacting design and manufacturing disci-pline that is often lacking in narrower or morespecialized markets, such as the military or medicalimaging.
The large potential size of the HDTV marketcould enable signiilcant improvements in manufac-turing technology as firms seek to lower productioncosts. In some cases, low-cost manufacturing willrequire pushing the state-of-the-art in componenttechnologies; in others, it will mean that HDTV willlet the computer or telecom markets push thestate-of-the-art and will then use those results.Above all, HDTV-asfor allconsumerelectronics—will require pushing the limits of cost-effectivemanufacturing of sophisticated electronic systems.This might be one of the most important impacts ofHDTV.
Whether or not a consumer HDTV market devel-ops, the expectation that there will be a large marketis forcing manufacturers who wish to participate topush the state-of-the-art in the various HDTV-related technologies. If the market does develop,then large-volume production might give the pro-ducer economies of scale in a number of othercomponents and products.
Having a technology matters little if markets areclosed to innovators or the entry barriers areeffectively insurmountable. As a result, it is alsoimportant for those that develop the technology tocapture a si~lcant share of the market. In the past,the United States has assumed that if the technology
was developed, markets would follow. Faced withlarge, usually vertically (and horizontally) inte-grated, aggressive foreign competitors; and con-fronted with increasingly skill- and capital-intensiveR&D and manufacturing to produce high-technologygoods, this assumption is no longer valid.
Neither linkages nor market share nor volumeproduction of computers were su.flicient to save theU.S. DRAM business. U.S. firms produce about 70percent of the world’ spersonal computers today andlead the world in PC design. Nevertheless, domesticfms have lost the market for DRAMs to Japanesefins. A combination of factors, including lessefficient manufacturing by some U.S. fiis on onehand and aggressive foreign trade practices on theother, forced most U.S. manufacturers out of thisimportant market.
The United States has also lost important HRSimaging markets such as low-end copiers, as well asmany other pieces of the electronics industry,despite having a predominant market share in manyof these just a few years ago.
Startups in the U.S. electronics industry areincreasingly focusing on design alone and depend onforeign operations for the highly capital-intensivemanufacturing operations5—they cannot secure thecapital necessary to do the manufacturing them-selves. In contrast, a number of foreign firms withlittle expertise in advanced electronics are becomingimportant manufacturers of electronics through heavyand long-term investments and careful attention tothe manufacturing process. For example, NMBSemiconductor, a new subsidiary of a Japaneseball-bearing company, Minebea, in just 5 yearsentered and became the world leader in very fastDRAMs.6 Kubota, a Japanese agricultural equip-ment company, is now manufacturing rnini-supercomputers designed in the United States.7Similarly, Korean semiconductor f-s are nowbecoming “mportantproducers of commodity DRAMs.
The United States cannot survive by performingR&D alone. Manufacturing provides far more jobs,
‘@thers include non-impact printers and multiplexing automobile wiring. In particular, automotive applications are expected to become anincreasingly important driverof tbis technology in thenextfew years. Martin Gold, ‘Autos Drive Smart Power IC R&D,” Electronic Engineen”ng Times,Jan. 15, 1990, p. 39.
s~tep~ Cillxlit @*l@ Corp., “Mid-Term 1988,’ Scottsdale, AZ, lists 18 startups during 1984-87 that chose not to build their ownfabrication facilities. Of these, at least one-third were using fabrication facilities in Japu ‘hiwa~ and Korea.
6$~How~ Took @er ~ Fmt D~ -et,” Electronics, NOVedXX 1988, p. 188; Bob Johtone, “Chips and Sum’ FarEastern EconomicReview, July 20, 1989, p. 52.
~avid E. Sanger, “U.S. Parts, Japanese Computer, ” New York Times, Sept. 7, 1988, p. D].
Chapter $+Linkages Between HDTV HRS, and Other IndWries ● 63
greater value-added, and larger cash flows thanR&D, and these are needed if future investments areto be made in R&D or production. Nor can we expectevery American to earn a living as a design engineer.With the increasingly tight linkages between R&Dand manufacturing due to the exacting requirementsof modern manufacturing processes and qualitycontrol, and due to the need to design for manufac-turability, R&D is merging with the manufacturingprocess. In many cases, when the United States losesmanufacturing, the loss of R&D is not far behind.
The problems facing the U.S. electronics industryare much broader than simply HDTV. AlthoughHDTV maybe an important element of any broaderU.S. strategy in electronics, by itself HDTV willneither seal the fate nor save the U.S. electronicsindustry. There are undeniable linkages-somestrong, some weak-that should be recognized, butthe problems facing the U. S. industry extend intomany other financial and structural factors. Theseinclude: the higher cost of capital in the UnitedStates resulting, in part, in lower capital and R&Dinvestments than our competitors; inattention tomanufacturing process and quality, poor design formanufacturability, and separation of R&D hornmanufacturing; foreign dumping and foreign marketprotection; and smallness of scale and/or lack ofvertical/horizontal integration compared to foreigncompetitors. Some of these broader issues facingU.S. manufacturing are discussed in a recent OTAreport ‘‘Making Things Better. ’
SEMICONDUCTORSThe rapid technological advances and cost reduc-
tions of digital electronics will likely make HDTVaffordable in the not-too-distant future. During thepast 10 years, the capacity of leading-edge memorychips (DRAMs) has increased by 250 times whilethe cost per unit memory has decreased nearly 100times.9 Each generation of advanced TVs will useincreasingly complex digital semiconductors toprovide a better quality picture at a lower cost.
In turn, HDTV will directly push the state-of-the-art in various aspects of digital signal processor(DSP), display, data storage, and possibly semicon-
ductorpackaging technologies, among others. HDTVmay indirectly impact a much broader range ofcomponents as well as computer and telecommuni-cation systems as a result of these technologicaladvancements.
Digital Signal Processing
There are three steps in digital signal processing.First, the continuously varying analog signals of thereal world—sound, light, temperature, etc.—areconverted into a digital form usable by computerswith an analog to digital (A/D) converter. (See box4-2.) Second, the signal is processed with a DigitalSignal Processor-to decode the tightly compressedbroadcast signal back into a recognizable picture, orreduce ghosts and snow (noise) to produce anear-flawless picture. Third, the digital signal isconverted back into an analog form-sound andpictures, etc. that people can understand-with adigital to analog (D/A) converter.
A digital signal can be manipulated (as in signalcompression), analyzed, transmitted with greaterreliability, and stored in computer memory. Ingeneral, the more the broadcast signal is compressedto fit into a narrow bandwidth, the more digitalsignal processing power is required for its recon-struction.l”
Digital signal processing is used today in compactdisk players, facsimile mail, long-distance telephonelines, computer modems, and in other applications.Human hearing and vision are analog, so digitalsignal processing will play an increasingly impor-tant role in providing an ‘interface’ between peopleand information systems in the future as we come torely more on images and sound instead of alphanu-meric text.
Digital signal processing chips (A/D, DSP, D/A)and digital signal processors in particular for HDTVSare at the leading edge of many aspects of thetechnology (figure 5-l). For example, at the 1989International Solid-State Circuits Conference (themost important international conference for unveil-ing new chip technologies), some of the fastest DSPchips ever developed were specifically designed for
SU.S. Congms, Of3ice of Technology Assessment, Making Things Better: Competing in A4an@acturing, OTA-ITE-443 (Wash@on, DC: U.S.Government Printing Ofllce, Febrwuy 1990).
%tegrated Circuit Engineering Corp., op. cit., footnote 5.l~e development of ~tter video si~ ~~p~essio~ ~gori~s is ~ impo~nt non.~dw~ ~~t of cment work kl DTV tht k dSO Critid
to the success of interactive video systems.
64 ● The Big Picture: HDTV and High-Resolution Systems
Figure 5-1-Selected Applications of Digital Signal Processors v. the Speed of Operation
Satellite communicationsI I
Radar and sonarI I
Imaae cvocessina
Speech recognitionI I
Telephone, modems, and fax Conventional TV tfDTV Multiply-accumulate
I [ ~~ cycle time
Ordinary microprocessors1 I
DSP microprocessors
HDTV is among the most demanding of applications in terms of speedSOURCE: Jack Shandle, “Signal Processors Open Up New Territory in Communieations,” Hectronics, April 1989, p. 79. Used with permission.
HDTV or related color video signal processing.11The only comparably demanding applications today,at least in terms of speed and throughput, are militaryradar and sonar (highly specialized and low-volumemarkets), and image processing that is closely
related to HDTV.
An uncompressed HDTV signal contains about1.2 billion bits per second (1.2 Gbps) of information.In comparison, today’s advanced engineering work-station hits peak speeds internally of roughly one-half gigabit per second.12 This signal is compressed,transmitted, and is then converted back into aviewable picture by the receiver. The amount ofcomputation needed to decode this varies with thestandard chosen, but can be as much as 2 to 3 billionmathematical operations per second. DSPS are ableto handle these huge information flows and compu-tational speeds—roughly comparable to those at-tained by today’s supercomputers—at a cost con-sumers can pay only through specialized designstailored for specific tasks. Unlike HDTVS, super-computers are able to handle a broad range ofcomputations and to do so much more flexibly .13
The development of certain important computertechnologies may be aided in part through efforts indeveloping digital signal processing for HDTV. Forexample, the ‘Ctestbeds” built by the Japanese todevelop DSPS have required extensive work withmassively parallel processor systems—the ability tohook-up many microprocessors in parallel to speedup computations. At Nippon Telephone & Tele-graph, for example, the National Academy ofSciences Panel reviewing Japanese HDTV develop-ment efforts observed a system with 1024 processorsin parallel, far fewer than some U.S. systems but stilla notable achievement. Parallel processing has beena significant weakness in Japanese supercomputertechnology, and a primary area in which U.S. fmshave managed to maintain their edge. The experi-ence with parallel processing hardware that theJapanese have gained in their HDTV developmentefforts may have spinoffs to their supercomputersystems.
Similarly, HDTV research at the David SarnoffResearch Center has led to the development of avideo-supercomputer capable of an informationflow rate of 1.4 Gbps and computational speeds of
Illlickd Doherty, “At ISSCC, Pamllel Signal processing, “ Electronic Engineering Times, Feb. 27, 1989.lz~c~d McComc~ “Supercomputer Highway co~d Be a co~&y Road, Say Industry Executive, “ New Technology Week, Aug. 14, 1989.]sDavid Mes~m~~~tt, Dep~ment of El&~c~ En@eefig, univ~sity of C~ifomia-Berkeley, pmSOIXd COml!lUl unicatiom July 31, Oct. 12 and
13, 1989.
141bid.
Chapter 5-Linkages Between HDW, HRS, and Other Industries ● 65
some 1.4 trillion mathematical operations per sec-ond at a cost of less than one-tenth that of othersupercomputers.ls
The DSP market is growing rapidly. It is expectedto increase from about $650 million per year to $1.6billion by 1992 (figure 5-2). If the HDTV marketdevelops, HDTVS will use an enormous amount ofdigital signal processing. For example, the JapaneseMUSE-9 system uses some 500,000 gates—a meas-ure of processing power—to convert the highlycompressed signal back to a picture. 16 DSP require-ments may be less or more than this, however,depending on the eventual choice of transmissionand receiver standards.
If high rates of growth are realized for the HDTVmarket in the United States, Japan, and Europe,17then 15 years from now the use of digital signalprocessing chips in HDTVS alone could be 10 timestoday’s total world demand (measured by processingcapacity-’ gates”) for all microprocessor and re-lated applications and roughly 100 times today’sdemand for DSP.18 Such estimates are speculative;their qualifications are discussed below. Dependingon the relative growth of the computer, telecommu-nications, and other markets, this may or may not beimportant compared to the entire microprocessorand microcontroller market in 15 years. It would,however, almost certainly have a strong impact onthe cost of DSPS for video processing. Even underlow-growth scenarios a tenth as large, HDTV wouldlikely have a strong impact on the cost of DSPS forvideo processing.
The United States currently has a stronger posi-tion in DSP design than Japan and is about equal toJapan in the production and performance of DSPchips. Domestic firms have managed to maintain adominant market position in DSP because they havebetter software for developing these chips. TexasInstruments currently has 60 percent of the worldDSP market; NEC is second with 11 percent.19Production of DSPS for HDTV could significantly
1989
1990
1991
1992
Figure 5-2—Projected Growth of the Digital SignalProcessor Market, 1989-92
Small IDSP Markets Grow Fast— —
—
E=I:-mlO 65
IEzI090
H----”-i
1 25
— — - ..———— —. II 1 6 2 5 I
l— II —J–- -J— - . L– - . L . _ I I
72 () 4 06 08 1 0 1 2 1 4 1 6Worldw!de DSP-product revenues ($ bllltons)
_ Consumer ~ industrial ~ Government and mllltary
D $ lenLe a n d en{]lneerlng ~ Commumcations
SOURCE: Jonah McLeod, “DSP,” Ektronics, April 1989, p. 71.
change these market positions, depending on afirm’s presence in HDTV. Such concerns may havebeen a factor in Texas Instruments’ recent decisionto purchase the Japanese HDTV chip designs andtechnology from Japan’s NHK in order to participatein the Japanese HDTV market.20
DRAMs
HDTVS similarly place heavy requirements onmemory technology. Access times needed for HDTVmemory chips must be roughly 20 nanoseconds(ns)—20 billionths of a second. Today’s fastestDRAMs have typical access times of 60 to 80 ns.
Leading-edge PCs and workstations are providinga significant market pull for the special techniquesand faster types of memory devices such as SRAMS(Static Random Access Memory) necessag tooperate at high speeds. If the HDTV market devel-ops, it could provide an additional pull for leading-edge fast DRAMs. Matsushita’s 8 Mb Video RAM,for example, has a serial access time of 20 ns, 1.5times faster than current VRAM technology.
IsDavid s~off R~earch Center, “The Princeton Engine: A Video-Supercomputer, ’ SWIIOff ~bs, Sept. 28, 1989-
16David Lammers, “U.S. ASICS on TV, ” Electronic Engineering Times, Feb. 27, 1989.17Cone5pon&g Uket projections are those by the Electronic Industries Association (EM) for the United States and the Japanese ~stry of posts
and Telecommunications (MOPT). Similar sales rates were assumed for Europe with appropriate reductions for the observed lower rates of householdpenetration for color TVs.
Igworld logic Gate production in 1987 is estited by O’IA to be about 2 trillion gates for microprocessors, microcontrollers, and ASICS. Thecomparison is made by multiplying the high growth projections for HDTV by 500,000 gates per set.
Iwdfim I. s~auss, “Fe@-Genemtion DSPS Will Debut Next Year,” Electronic Engineenng Times, Aug. 7, 1989, p. 55.
‘Jacob Schlesinger, “Texas Instruments Agrees To Buy HDTV Technology From Japw ” Wall Sfreet.lourmd, Sept. 14, 1989.
& ● The Big Pictire: HDTV and High-Resolution Systems
HDTV could also become a si~lcant newmarket for DRAMs. An HDTV might use as muchas 32 million bits (Mb), equivalent to 4 million bytes(MB) of DRAM, to store the HDTV picture inmemory.21 Assurning rapid growth of the HDTVmarket as above, then the use of DRAMs in HDTVSalone in 15 years could be five times the total 1987world demand (by memory capacity-bits) for allDIUM applications.” Depending on the relativegrowth of the computer, telecommunications, andother markets, this may or may not represent asignificant tiaction of world DRAM use at that time.Japanese firms are, however, already establishingmajor new DRAM production facilities with theexpectation that their output will be used in Ad-vanced TVS.23
These scenarios for DRAMs and DSPS are subjectto a number of quali.flcations and uncertainties. If thestandard chosen for HDTV uses significantly lessormore memory and digital signal processing, then theprojections would be adjusted accordingly. If thedevelopment of the IDTV market, for example,substitutes for HDTV and prevents HDTV marketgrowth, then the DRAM projections would bereduced by a factor of 4 because the typical IDTV isexpected to use about a quarter of the memory usedin an HDTV.W If consumers instead move progres-sively upscale, buying IDTVS and EDTVS first andthen move to HDTV, relegating their IDTVs/EDTVS to use as a second set as they did B&W setswhen moving to color, then chip demands could be25 to 50 percent greater than projected. If strongcommercial markets for HDTV develop, as pre-dicted by the Japanese MPT, then chip demandcould be twice the projections above. Factoring inproduction and broadcasting equipment sales couldincrease these projections by 10 to 15 percent.mFinally, some believe that progressively more so-phisticated systems will be developed beyond HDTV,requiring even more memory and signal processing.
Fifteen years ago the United States had more than90 percent of the world market in DRAMs; today theUnited States makes less than 15 percent of theDRAMs purchased in world (merchant) markets.Texas Instruments, Micron Technologies, Motorolawith technology licensed horn Toshiba, and IBM(for internal consumption) are the only U.S. firmsstill producing DRAMs. The recent effort to formaconsortium, U.S. Memories, might have improvedsomewhat the U.S. position. For a variety of reasons,however, it failed to attract sufficient support hornU.S. firms to even be launched.
Gallium Arsenide and Other CompoundSemiconductors
Receiver-compatible HDTV systems propose useof the standard 6 MHz NTSC signal, which wouldthen be augmented with a second signal 3 to 6 MHzwide to provide the additional information for thehigher quality picture. The wider bandwidth of suchHDTV systems may require GaAs (Gallium Ar-senide) chips in the tuner due to their widerbandwidth capability and their ability to handleoverloads.~
GaAs and related materials are now used in avariety of applications, ranging fkom some leading-edge supercomputers to the lasers in CD-players andin fiber-optic systems. The use of GaAs remainslimited, however, due to the difficulty of producinghigh-quality stock material and fabricating semicon-ductor devices from it. If HDTV provides a largemarket for GaAs devices, the additional productionvolume might help some of these difficulties to beovercome. Improved GaAs materials production andfabrication techniques could have spinoffs to avariety of markets.
The United States seriously lags Japan in a varietyof GaAs and related materials, processing, and
ZISome re-hers &lieve that fairly high resolutions can be achieved without such large use of memory; Others insist tit th=e tie signifi~ntadvantages in having a full frame memory, or 20 to 32 Mbits, depending on the standard, etc. Ultimately, the mezno~ and digital signal processordemands will depend strongly on the particular standard chosen, but will also likely increase over time. (A Byte equals 8 bits and can represent onecharacter on a keyboard.)
%Estimated 1987 world DRAM production is 2.3x1014 bits. Integrated Circuit Engineering Corp., op. cit., footnote 5.Z+N~onKoWo S- J@y b, 1989, cit~~ B~ w’haltm and Mark Eato~ “Prospects for Development of au.s. ~~~dustry,” CO@~
on GovemrnentaJ Affairs, U.S. Senate, Hearing 101-226, Aug. 1, 1989, p. 516.
%Xuules L. Cohen+ “NEC ‘Ihkes An E!arly bad in Improved-Definition TV,” E2ecrronics, Dec. 17, 1987.fiw- G1e@ norida Atlantic University, Boca Raton, personal cmnmunicatiom Feb. 10, 1989.%~ Kel~ Natio@ s~~ndu~tor, pm~ ~mmuuication, ~. 21, 1989; Birney r)ayto~ Nvisio~ p~~ communi~o~ ~. & @t. 12
and 13, 1989; Ronald Rosenzweig, Testimony at hearings before the House Committee on Science, Space and Technology, Mar. 22, 1989.
Chapter Uinkages Between HD~ HRS, and Other Industries . 67
device technologies.27 Despite pioneering the devel-opment of many of these semiconductor devices, theUnited States today buys much of the unprocessedGaAs material horn Japan as well as the semicon-ductor devices fabricated from it. AT&T inventedthe solid-state laser, but in some cases has purchasedsemiconductor lasers from Japan to drive its fibercable.
Semiconductor Manufacturing
Highly disciplined and cost-effective manufactur-ing is required for a large HDTV market to developand for a firm to successfully compete within it.Technology will have to squeeze even more circuitryonto the same sliver of silicon to bring the cost ofHDTVS to reasonable levels. Reductions of the totalnumber of chips in this manner reduces costs-ofcomponents, of assembling and testing the HDTV,of repairing defects, and by increasing reliability.Reducing the number of parts in color TVs was animportant aspect of the competition between theU.S. and Japanese producers in the 1970s—and anaspect in which U.S. producers seriously lagged.The quest to reduce the number of chips in systemsis responsible for the explosion in ASIC (Applica-tion SpecKlc Integrated Circuits) production, whichnow accounts for about a quarter of all merchantintegrated circuit Production.x
Efforts to reduce the number of chips needed canalready be widely seen in ATV development. NEC,for example, reduced the number of chips in itsIDTV horn 1,800 to 30.29 An early prototype of theJapanese MUSE HDTV system had 40 printedcircuit boards, each centaining 200 chips for a totalof 8,000 chips. In contrast, the latest generation ofMUSE decoders unveiled in June 1989 has less than100 chips. Half of these were ASICS with 26
different designs. To minimke the burden on anyone manufacturer in developing these numerous andcomplex ASIC designs, NHK divided the effortamong 6 different manufacturers, and then distrib-uted the designs among all the participants (ch. 2).
Manufacturers are also pushing the design ofconventional memory chips. Matsushita recentlyunveiled an 8 Mb Video RAM designed spectilcallyfor application in HDTV, and intends to begincommercial sampling in 1990.30
Eventually, nearly all of the required memory andDSP for an HDTV might be combined on a singlechip. Increasing levels of chip integration willrequire a significant increase in current capabilities,and correspond to the expected leading edge ofsemiconductor technology for the next decade.31The extent to which this drives the state-of-the-artwill depend on the relative size of the HDTV,computer, telecommunications, and other markets.
A number of important studies have documentedthe current U.S. lag behind Japan in a broad range ofsemiconductor process technologies:
●
●
●
The Federal Interagency Thsk Force found theUnited States lagging Japan in 14 semiconduc-tor process and product areas; the United Stateswas ahead in just six categories and its lead wasfound to be slipping in five of these (figure1-1).32The National Academy of Sciences found theJapanese leading in 8 of 11 semiconductorprocess technologies that will be critical in the‘ihture.33 -
A recent study by the Department of Commercefound Japanese semiconductor plants had a5-year lead over the United States in the use ofcomputer integrated manufacturing techniques.x
27Rqoti Ofthe F~~ ~temgency SW working Group, ‘ ‘The Semiconductor ~dus~, “ National Science FoundatioxL Washington DC, Nov. 16,1987.
~Mm~~nt pr~ucem we those ~hic.h ~11 on tie open market and ficlude ~ Japanese and most U.S. wmiwnductor producers. Captive prOdUCerSare those which use the semiconductors they produce themselves and do not sell them outside the fm. The world’s top three ASIC producers are Fujitsu,‘lbshib& and NI?C.
3Cohe~ op. cit., footnote ~.3oM,iyoko Mku@ “8 Mbit VRAM for HDTV,” Electronic Engineering Times, Apr. 3, 1989.slDa~o~ op. cit., fm~ote 26; and* onpmj~tions by Craig R. Btumt~ “TechnoIow Dixwtions for~tegfated circ~ts~” Se minar at O’IA, June
15, 1989.32RepoII of a R&ml Interagency staff wO&@ @XIp, Op. cit., fOOtnOte 27.qsNatio~ Rmearch Comc& ~$Advanc~ fio~ss~ of Elec@onic Mate& in the United States and Japa~” National Academy Pm.w, 1986.
%T.C. Holto~ J. Dussaul~ D.A. Hodges, C.L. Liu, J.D. plummer, D.E. Thomas, and B.F. Wu, “Computer Integrated Manufacturing (CIM) andComputer Assisted Design (CAD) for the Semiconductor Industry in Japa%” JTECH Panel Repo@ Department of Commerce, Science ApplicationsInternational Corp., December 1988.
68 ● The Big picture: HDW and High-Resolution Systems
These techniques have allowed the Japanese toreduce turnaround time by 42 percent, increaseunit output by 50 percent, increase equipmentuptime by 32 percent, and reduce direct laborrequirements by 25 percent.
The Japanese have also rapidly improved theirplant and equipment to take advantage of the moretechnically demanding, but cost-saving, large wafertechnology. From a position of parity in 1984, theynow use, on average, wafers that are nearly 35percent larger in area than their American competi-tors.35 IBM, however, is pioneering very large,8-inch, wafer technology.
DISPLAY TECHNOLOGYHDTV drives display technology perhaps more
than any other single area. To truly appreciateHDTV, much larger high-resolution displays areneeded than are generally available today. Indeed,some analysts believe that the HDTV market willnot take off until large display s40-inch diagonaland preferably larger-are available at reasonablecost. A HDTV display must have fairly highresolution—lOOO lines or more; superb color; rapidresponse times; large size; good brightness, contrast,and efficiency; and low cost.
Numerous display technologies are being devel-oped, including: improvements in conventionalpicture tubes; advanced projection displays usingeither CRTs, LCDS, or deformable membranes3G;and large-area flat panel liquid crystal displays,among others.
Conventional picture tubes will undoubtedly con-tinue to be the display of choice over the next fewyears. They perform well, they are efficient, and theyare low in cost due to the many years of experienceman~facturing them. In the longer term, however,there will be a shift away from direct view picture
tubes. In the larger sizes desired for HDTV, direct-view CRTs are bul@ and heavy in addition to beingfi-agi.le. Although work is being done to reduce theirdepth37, CRTs currently are nearly as deep as theyare wide-few houses have either doors wideenough to accommodate large CRT displays (40-inch or more) or living rooms large enough toconveniently house them. Furthermore, the weightof a 40-inch CRT display is several hundred pounds.
By the mid-1990s, many analysts expect thathigh-performance projection systems will be availa-ble that provide the larger viewing areas needed forHDTV. Toshiba, NHK, Hitachi, Sanyo, Mitsubishi,and Philips have all developed projection systemsfor HDTV with screen sizes as large as 50 feetdiagonal. 38 LCD and deformable membrane projec-tion systems are also under development with someindications that the deformable membrane may haveadvantages in efllciency, brightness, contrast, andresponse time.
By the late 1990s, yet another display technologymay become available-the active matrix flat panelliquid crystal display, or AM/LCD. Nine Japanesecompanies demonstrated 10- to 14-inch color LCDdisplays with resolutions of 640 by 400 pixels at the1989 Tokyo Business Show.39 IBM recently un-veiled an experimental high-resolution 14-inch di-agonal color liquid crystal display that it co-developed with Toshiba.40 Figure 5-3 illustrates thecurrent progress in developing AM/LCD displaysand a few of the firms that have led the way.41 TheJapanese recently began a 7-year, $100 millioncollaborative research program to develop verylarge, 40-inch diagonal, color flat panel AM/LCDdisplays for HDTV and other applications (ch. 2).
Other display technologies are being investigated,but barring fundamental breakthroughs,42 they areless likely to be applied to HDTV. (In contrast, for
350T- ~~~p.
~cRTs ~e cathode-ray tubes, the basic technology used today for TV picture tubes. LCDS are liquid crystal displays, the basic twbology used todayin pocket watches, calculators, and many of the laptop PC displays. Deformable membranes work by reflecting light off a membrane that is deformedpoint by point to either focus or diffuse the light so as to generate a picture.
sTDavid tires, “IWMSUShita Flattens CRT,” Electronic Engineen”ng Times, my 29, 1989, p. 29.3s~WenceEo ~ws, “~TvandHDTVDisplays: Evolv~gfiJap~” JT’Ec!HP~el Review of HDTV~Jap~presen~tion atNational Academy
of Sciences, July 26, 1989.sgElec&onics, Septaba 1989, p. 100; Special Advertising SmtiOn on JaP~.@AshOk Bin@ 46F1at panel AdV~ceS had SID,” EleC~Onl’c Engineering Times, Apr. 24, 1989, p. 35.
41wfl~ c. Sctieider, ~l~L. Resor, “~gh.volume fioduction of ~ge F~l.Color Liquid-Crys@l Displays,” Znfonnution Display, February1989, p. 8.
AzFor e~ple, if a good blue phosphor can be fo~d for ekctrol uminescent displays, they might become a strong competitor with LCDS.
Chapter S-Linkages Between HDW, HRS, and Other Industries ● 69
Fiaure 5-3--’ ’Ressr’s Rule” for Active Matrix Licauid“
C r y s t a l D i s p l a y s (AM/LCDs) -
MITI goal
‘1
(1 meter by 1996) ,’”/ ’/
/ ’/// ’
/ ’/
/ ’/
Research//
&/ ’
D e v e l o p m e n t ,/” ,4‘% /’ /
/ //,/ // /// //
—19” TV—— -- “/-—— — —–//’/ /’/ ,’
/ /,.9/’ T o s h i b a ,/’ I
ing cost are critical factors for each of these displaytechnologies. Progress in these technologies, how-ever, is being made: the Japanese NHK Laboratoryrecently announced an experirnenta.1 20-inch diago-nal plasma displaf18, and some believe that newtechnologies might give plasma displays an edgeover A.M/LCDs.49
Production of large flat-panel displays will re-quire signii3cant advances in a number of importanttechnologies. These include: high-throughput low-cost lithography tools for large area, high-performance patterning of circuitry onto the panel;large-area, high precision glass sheet production;and large-area, high precision thin-film technology.
/T Sharp .’ I
+ M a t s u s h i t a , . ” < Commercial I Such capabilities may have applications in many/Seiko Epson ~
I /Production I/ Seiko Epson
J ‘-
‘ S e i k o E p s o n M a t s u s h i t a
-+ i - - l - l
1982 1984 1986 1988 1990
Year of Introduction
This shows the trend in display size by ye:
t t A
992 1994 1996
r for displays underR&D and in mmmercial produ~on. Display size, as measured bythe diagonal inch, is increasing at the rate of 3 inches (75mm) peryear. Commercial introduction follows R&D production by approx-imately 3 years.SOURCE: MRS Technology, Inc., Chelmsford, MA. Used with permission.
flat panel LCD displays the technical limitations arenow believed to largely be in engineering.) Amongthese displays are light-emitting diodes43, PlasmaDisplay Panelsw, thin-film electroluminescent dis-plays45, and fiber-optic expanders4G, among others.47
Size and weight, cost, brightness, power consump-tion, viewing angle, response time, and manufactur-
other arefi as well. High prec~~on control of thinftis, for example, is generic to a variety ofindustries, from semiconductors to the production ofoptical disks. Large-area lithography is expected tobe used to develop very high-density printed circuitboards, or “chips on glass” by the Japanese KeyTechnology Center AM/LCD research consortium.This could be a very important development and isdiscussed below.
Linkages such as these are difficult to anticipate.Simple accounting may overlook them, but they cansometimes lead to enormous new markets. Considerthe example of the Power htegrated Circuit. ThePower IC was initially developed and its costs weredriven down, in part, by the demands of such devicesas plasma and electrolti escent displays, amongothers,so for IC drivers capable of handling medium-level voltages (100 volts instead of the typically 5 orso volts used in computer circuits) and relativelyhigh currents.
43LEDs ~ve gener~y IOW efficiency in emitting lighg and there is an enormous variation in output and efilciency between different colo~, m*gthem hard to match. LEDs also have a fairly variable light output from device to device There is relatively little new research going on now in LEDs.
44plasm Display panels io~e a gas ~~ ~ medi~.level vol~ge (i.e., 100 volw) ~using the gas to glow. Disadv~tag~ include the cost C)f theelectronics to deliver this voltage, low efficiency, and a limited color range.
4sElectrol uminescent Displays work by applying a medium voltage across a material causing it to glow. Limitations me its low eftlciency and limitedcolor range. Planar (U.S.), in particular, hm been working on developing a better “blue’ color as well as better manufacturing processes and electronicsthat can vary the intensity of the colors. Tom Manuel,“A Full-Color EL Display Is Demonstrated by Planar,” Electronics, May 26, 1988, p. 73.
46Fi&r-optic expanders ~ve co~iderable profi5e, but we Cmenfly embrofied fi a patent ~pute. See G~rge Gilder, “Severed H-ds and WastedResources,” Forbes, June 26, 1989.
470~em ~clude Vacum fluoreSmnt display5, ~ which Fu~&., ~c, and Ise hold the #l, z, ‘3 market positions; cold cathode emitter d.kphiyS; ~delectrophoretic displays.
48Ashok B~&% ‘fFlat panel Advances bad SID, ” E/ec~onic Engineering Times, Apr. 24, 1989, p. 35.
A9David Lie~~~ ‘ ‘Plasma’s HDTV FOCUS, “ Electronic Engineering Times, June 12, 1989.
~othem include nonimpact printers and multiplexing automobile* g, etc. Athough apphCatiOfL3 in automobiles ordy began to be realized muchmore recently, they were a long-term goal for some manufacturers.
70 ● The Big picture: lfLITV and High-Resolution Systems
Plasma and electrolumin escent displays area tinyfraction of the display market and Power ICS forthese panels are a still smaller market. Despite suchhumble beginnings, Power ICS are now beginnin g tofmd applications across a host of industries withimportant benefits. These range horn potentiallysigniilcant reductions in the weight and cost ofaircraft wiring and controlssl to large improvementsin the efficiency of refrigerators, room air-conditioners, and a host of other appliances.52
With the expected transition to flat panels andother advanced display technologies, the UnitedStates has a fleeting opportunity to regain a strongmarket position in display technologies. In recentyears the United States strength in display technolo-gies and markets has slipped away to the Japanese.The United States still has a few scattered experts inbasic CRT technology at Zenith, Tektronix, GE,Raytheon, Corning and a few other fins, but it lacksthe broad-based talent of the Japanese, particularlyin manufacturing.
U.S. fms are still competitive in some aspects offlat-panel display technologies: in design, in theproduction of the basic materials, in some of themanufacturing equipment, and in some state-of-the-art displays. Coming makes the best glass substratein the world for AM/LCD displays and currentlyholds 90 percent of the Japanese market.53 MRSTechnologies, a Massachusetts venture startup, cur-rently makes the world’s best lithography tools forproducing large flat panel displays.54 U.S. fms alsoheld half the 1988 world market in electroluminesc-ent displays compared to Japan’s 29 percent; and aquarter of the 1988 world market for plasmadisplays—down from 57 percent in 1984—compared to Japan’s 68 percent.55
These strengths are unlikely to last. In AM/LCDdisplays, there are five significant R&D groups in
the United States-Sarnoff Labs, Ovonic, Magna-screen, Xerox, and Philips Labs (Briarcliff).5G GErecently dropped its research program in LCDS.There are more than a dozen fii in Japan doingR&D in AM/LCD displays, most of them with moreprojects and people involved than any of the U.S.teams.57 Tables 5-1 and 5-2 illustrate the disparitybetween U.S. and Japanese efforts in developing andproducing flat-panel displays.
No AM/LCD production lines are now operatingin the United States; essentially all of the world’sproduction comes from Japan-which holds roughly96 percent of the world market for (small pixel)passive matrix LCDS and virtually IOOpercent of themarket for active matrix LCDS.58 Even if the UnitedStates were to make breakthroughs, the manufactur-ing infrastructure to produce the displays would notbe in place. Without production, there will likely belittle revenue to continue a long high-level researchprogram-particularly considering the large capitalinvestment and engineering effort required forproducing very large area screens.
Already, the remainin g U.S. strengths are beingchallenged. For example, a MITI-sponsored consor-tium, the New Glass Forum, was begun in 1985 to doR&D in glasses, some of which may have applica-tions to AM/LCDs. Nikon (Japan) recently an-nounced a lithography system with a higher through-put than that of MRS Technologies.59
The market for displays for all purposes is large.Worldwide sales of flat panel displays of all typeswas $2.4 billion in 1988, out of a total display marketof $8.2 billion. The flat panel market is expected toapproach $6.3 billion by 1995, out of a total displaymarket of $14 billion.a Other estimates place the
slvl~~ Rume~ “power Devi&s Are In The Chips,” ZEEE Spectrum, July 1985.
Szswuel F. Baldw@ “Energy.EfflCient Electric Motor Drive Systems, “ in Electricity: Eflcient End-Use and New Generation Technologies, andTheir Planning Zmp/ications, Thomas B. JohanssoP Bi-rgit Bodlund, and Robert H. Williams (eds.) (Lund, Sweden: Lund University Press, 1989).
SSGW Stk ‘‘~n~ac~g Hurdles Cwenge Large-LCD Developers, ’ IEEE Spectrum, September 19*9, P. ~.
mNote, however, that it has lower throughput than the recent Nikon offering. See Stix, op. cit., footnote 53.
ssHeidi Ho- U.S. Department of Commerce, personal communication, NOV. 29, 19*9.fi~lti Resor, ~S T~~ologies, persomd communicatio~ Oct. 12, 19*9.
57Lawrence Tannas, l?mnas Electronics, personal communication, Aug 9, 1989.sgBob w’his~ BIS wc~tosh I.xmdo~ personal communicatio~ Mar. 29, 1989.59s~, op. cit., fOOmOte 53.
6f)~~u.s. Brains for a De~rate Fight ~ F~t p~els,” EZec~onics, DWember 1988.
Chapter Uinkages Between HDIY HRS, and Other Industries ● 71
Table 5-lAtatus of U.S. Producers of Flat Panel Displays in the 1980s
Company EL LCD PDP Other
Alphasil . . . . . . . . . . . . . . . . . . . . .AT&T . . . . . . . . . . . . . . . . . . . . . . .Cherry . . . . . . . . . . . . . . . . . . . . . . ProductionColoray . . . . . . . . . . . . . . . . . . . . .Control Data . . . . . . . . . . . . . . . . .Crystal Vision... . . . . . . . . . . . . .Eleotro-Plasma . . . . . . . . . . . . . . .EPID/Exxon . . . . . . . . . . . . . . . . . .Kylex/Exxon . . . . . . . . . . . . . . . . . .GE . . . . . . . . . . . . . . . . . . . . . . . . .GTE . . . . . . . . . . . . . . . . . . . . . . . . Closed 1987IBM . . . . . . . . . . . . . . . . . . . . . . . .LC Systems . . . . . . . . . . . . . . . . . .Magnasoreen . . . . . . . . . . . . . . . .NCR . . . . . . . . . . . . . . . . . . . . . . . .Ovonic . . . . . . . . . . . . . . . . . . . . . .Panelvision . . . . . . . . . . . . . . . . . .Photonios . . . . . . . . . . . . . . . . . . . .Planar . . . . . . . . . . . . . . . . . . . . . . . ProductionPlasma Graphics . . . . . . . . . . . . . .Plasmaoo . . . . . . . . . . . . . . . . . . . .Safnoff Labs . . . . . . . . . . . . . . . . .Sigmatron Nova . . . . . . . . . . . . . . Closed 1988TI . . . . . . . . . . . . . . . . . . . . . . . . . .Xerox . . . . . . . . . . . . . . . . . . . . . . .
Closed 1988
Closed1984
Sold 1983Sold 1989
Closed1988Researoh
ResearohSold1986
Researoh
Research
Closed 1987
Seeking fundingClosed 1980
ProductionClosed1986
Sold1987
CIosed 1984
Production
Closed 1985Research/Prod.
Closed 1983
SOURCE: DaveMentley, Stanfofd Resources,lncvSanJose,CA,personalcommunication,Apr. 25,1990;JimHurd,Planar Systems, Inc. Portland, OR, personal communication, May 1, 1990; Larry Weber, Plasmaco,Highland, NYj personal eommunieation, May 1, 1990; and Defense Scienee Board, “High DefinitionSystems Task Force, Final Report,” Washington, DC, forthcoming.
Tabie 5-2—Recent Investments in AM/LCD Production Facilities in Japan
Investment FaotoryCompany (U.S. $million)a Operational Iooation Technology
Alps . . . . . . . . . . . . . . . . . . . . . . . .Fuji-Xerox . . . . . . . . . . . . . . . . . . .Hitaohi . . . . . . . . . . . . . . . . . . . . . .Matsushita . . . . . . . . . . . . . . . . . . .NEC . . . . . . . . . . . . . . . . . . . . . . . .Seiko Epson . . . . . . . . . . . . . . . . .Seiko Instruments . . . . . . . . . . . . .Sharp . . . . . . . . . . . . . . . . . . . . . .Toshiba/lBM-J . . . . . . . . . . . . . . . .
$3380
134+67/yr2306716720447134
1992NDND
4Q1991NDNDND
3Q19932Q1991
IwakiEbina
MobaraOsaka
KagoshimaNagano Pref.
Akita Pref.Tenrie/Mie
Himeii
TITa-Si TRa-Si TtT
Poly-Si TITT~TIT
Diode Matrixa-Si TITa-Si TH
Worwertecf at U.S.$1.150 Yen.KEY: TIT-thin fiim transistor; a-Si-am orphous-sii.kxm; Poly-S+poiy-siiicon
SOURCE: Dave Mentley, Stanfotd Resources, Inc., San Jose, CA, personaf communication, Apr. 25, 1990; andDefense Science Board, “High Definition Systems Task Force, Finai Report,” Washington, DC, forthcoming.
flat panel market as high as $11.7 billion by 1996.61 The market for displays for HDTV could also beThe display market is also primarily driven by large. For example, if AM/LCDs became the displayconsumer applications: 70 percent of the 1988 of choice, using the high-growth scenario above,market was for consumer electronics; just 18 percent screen production for HDTV would be nearly 6,000for computer applications.b2 times total world production of active matrix LCD
61Da~d fiemm Cccolofi FUtUK for Flat panels,” Electronic Engineering Times, Jm 8, 1990, P. 71.%dmond Branger, “Flat Panels in Focus, ’’Electronic Engineering Times, Mar, 20, 1989.
72 ● The Big picture: HDTV an.d High-Resolution Systems
screens today. 63 HDTV might then be an irnport~tdriver of flat panel display technology as well ascontributing to economies of scale in production.Even with a small market for HDTV, the specialrequirements of HDTV will drive large-area AM/LCD or other flat panel technology development.
STORAGEThe state-of-the-art in magnetic and optical stor-
age technologies for studio use is already beingpushed by the large volume and high rate ofinformation flow requirements for HDTV and, to alesser extent, related HRS.@ Digital VCRs forstudios will require much higher magnetic recordingdensities and information transfer rates through theuse of improved magnetic materials, recordingheads, and other techniques.G5 Sony, for example,has developed a prototype studio VCR that has arecording speed of 1.2 Gbps—five times faster thanthe previous record.fi To similarly extend recordingtimes on compact disks, the semiconductor lasersused will have to operate at higher frequencies thanthose used today, requiring advances in semiconduc-tor lasers and reductions in production costs.G7Matsushita has recently succeeded in storing 2.6 GBof video information on a single 12-inch opticaldisk.G8 These recording technologies will have manyapplications in the computer industry.G9
Such spinoffs horn consumer electronics havealready been widely seen. Magnetic and optical(compact disks) storage technologies were bothoriginally developed for the consumer electronicsmarket, but are now used widely in the computer
industry. In particular, compact disks are expected tohave a profound impact on information handlinggenerally.
Similarly, Digital Audio Tape (DAT) drives,originally developed for the consumer market, areexpected to have a si@lcant impact on thecomputer data storage market. DAT sales in the U.S.consumer market have been limited due to U. S.-Japan trade friction and issues of copyright protec-tion; therefore, prices are expected to remain higherthan if large volume sales had already been achieved.If approved, legislation currently pending in theCongress that requires copy-controlling devices inDAT machines may open up U.S. markets. DAT willbe able to store about 1.3 GB of data on a cassette thesize of a credit card and about 3/8-inch thick and willhave data rates of roughly 1.4 Mbps.70
There are also spinoffs between technologies. Thehard drives used in computers are made by coatinga very thin, high-quality layer of magnetic materialon a metal disk. The technology to do this, and eventhe processing equipment, originally came hornsemiconductor wafer fabrication. A key technologyfor Iarge-area, flat-panel displays will similarly beputting extremely thin, high-precision coatings oververy wide areas. Once developed, this could have animpact on the production of semiconductors, and onmagnetic and optical storage. The converse is alsotrue. Thin-film technologies developed for thesemiconductor industry could initially have animpact on flat panel production, although theimpacts will likely decrease as the panel areasincrease.
Gsworld toti pr~uction of small pixel (not watches or calculators) passive LCD displays in 1988 was roughly 6.13 million tits, of which Japmesefirms produced 96 Pemeng U.S. firms 4 percent. World production of active matrix displays was 0.21 million units in 1988, essentially by all Japanesefirms. Assuming a distribution of 60 percen~ 3- to 6-inch diagonal, and 40 percent, 6- to 12-inch diagonal and averaging, the area is approxirnately4.6million square inches of active matrix panels; and 135 million square inches of passive matrix LCDS. If high growth rate projections are realized,assuming each HDTV has a40-inch diagonal (or 768 square inches for a 16:9 aspect ratio), then the total active matrix screen area produced will be 5,844times today’s production of active matrix screens; 200 times today’s production of passive matrix. For the large areas required for HDTV, active matrixscreens will be necessary. World LCD production figures for 1988 were supplied by Bob Whisk@ BIS Mackintosh London.
‘Note that for consumer use, the extensive signal compression brings storage densities to near the level of current technology, which is doubling itsstorage capacity every 2.5 years. The video signal is not compressed before storage in the studio, however, to prevent the introduction of errors or artifacts.
GsOthe~ might include improved error correction algorithms or data coding techniques.66’’HDTV Faces Many Hurdles in Japan, ” Electronic Engineering Times, Apr. 10, 1989. And Messerschmi~ op. ci~ footnote 13.67~~HD~ Faces ~ny H~dles in Japa~” op. cit., footnote 66. Increasing information stomge on today’s ~S k diffhdt beCaUSe the spot she on
the CD is approaching the wavelength of the lasers used-the system is approaching the diffraction limit. To increase &ta storage, higher frequenciesare needed.
@“Digital HDTV Goes On Record,” ZEEE Spectrum, September 1988, p. 26. Please note that gigabytes (GB) is a volume of information-like abucket of water; whereas gigabits per second (Gbps) is a rate of information flow—like the rate of flow through a hose. These cannot be compared.
6s~~~c~ves ~ -tie,” PC Magazine, Jan. 31, 1989.70Terry COStiOW, “TechDiscord Rips Tape Industry,” Electronic Engineering Times, May 29, 1989, p. 53; Eng T&I and Bert Vermeule~ “Digital
Audio Tape for Data Storage,”IEEE Spectrum, October 1989, p. 34.
Chapter Minkuges Between HDW, HRS, and Other IndMries . 73
Precision motors like those used in HD-VCRSwill also be used in computer tape and disk drives,robotics, and elsewhere. Ttiay, Japan is the world’slargest producer of precision motors due to thissynergy of uses among electronic products.71
The high-precision helical scan drives for VCRsare now made primmily in Japan, although a few aremade in Korea. Exabyte of Boulder, Coloradopurchases off-the-shelf 8mm camcorder-type tapedrive mechanisms from Sony and uses them in a 2.3GB tape system (the highest storage capacity to date)for computer data storage;72 they are totally depend-ent on the Japanese source. With the continuingmove to higher density storage systems, firms thatproduce computer tape storage systems, but do nothave access to helical scanning (VCR-type) tapedrives are unlikely to survive.73
The United States continues to hold a strong R&Dand market position in some storage technologies,but has seriously lost ground in others. The UnitedStates has largely lost the floppy drive business,holding just 2 percent of world sales in 1987; but inthe hard drive market, U.S. fms have fought backsuccessfidly and still hold a 60 percent or bettershare. 3M continues to be a major world-classproducer of magnetic tape; but no U.S. fm pro-duces the high-performance helical scan recorderdrives. Only one domestic fii, Recording Physicsin California, has the capability to produce the veryhigh-performance materials needed for read/writeVCR heads.74
The United States lags in many areas of opticalstorage research, and has littIe presence in the
manufacture of optical storage devices. Over a dozenJapanese firms are developing or selling advancedrewritable optical disks and/or drives.7s The opticaldata storage device market is expected to grow from$400 million in the United States in 1988 to $7.3billion by 1993.76
COMMUNICATIONSWith the declining costs of fiber, the extension of
fiber to the home is expected to become moreaffordable. Some estimate fiber may be cost-effective for large, new housing developments by1992. There are optimistic projections that 17million homes and small businesses could be hookedup to fiber by 1999.77 The very high informationcarrying capacity of fiber may make it the carrier ofchoice for HDTV in the fhture; and if the HDTVmarket develops, it could further stimulate the use offiber-initially in the cable backbone and later to thehome.78 In the near- to mid-term, however, coaxialcable will continue to be the most important mediumfor carrying video signals to the home.79
As HDTV begins to be networked via fiber-optic, it could be an important force behind the nextgeneration of telecommunications equipment. HDTVwill require wide bandwidths and, correspondingly,a wide bandwidth switching and control system at acost consumers will pay (figure 5-4). Similarly,low-cost techniques will have to be developed forinstalling optical fibers to households. Software tooperate and manage a fiber network must also bedeveloped. Today, sofsvare is a significant fractionof the expense of telecommunications80 and recent
71Yuji Akiy% “tiO* D
emands of Precision Motors Are Supported by AV Equipment Industial Robots,” JEE, February 1986, pp. 3@35. IiIparticular, he notes that 40 percent of Japan’s precision motors go to audio video equipment. Further, fewer than 10 producers in Japan are able to makemotors of the precision needed for home VCRs, with a core deviation of less than 2 microns, a speed deviation of less than 0.02 percent and a price lessthan 1,500 yen.
~Pleme note that this is the standard VCR tape drive ti.s used by COn8umers today. Its recording capacity is in GBytes and cannot be compared to theexperimental Sony HDTV studio tape drive above with a recording speed of 1.2 Gbits per second.
TsCkk Johnso~ consultan~ personal communication, Aug. 11, 1989; Oct. 26, 1989.
7%id.75~e5e ~clude Sony, S-, (&nou Nikoq ol~pus, ~~ushi~, ~shib~ Mitsubishi, fitachi, Fujitsu, NEC, Ricoh, P i o n e e r . EZectroru”cs,
September 1989, Special Advertising Section on Japan, p. 105. Roughly one-sixth of optical disk drive production for computer data storage applications,on a $-basis, is done in the United States but essentially none for consumer applications is made here, and this is a far Imger market today. Ron Powel~NIST personal communication, Nov. 14, 1989.
7sManny Fernandez, “Forecast ‘89-Mixed Emotions,” Electronic Engineering Times, June 12, 1989.
“‘Fiber Optics: Getting Cheap Enough To Start Rewiring America,” Business Week, July 31, 1989, p. 86.
78Larry Wailer, “Fiber’s New Battleground: Closing the Local Loop,” Electronics, February 1989, p. 94.7~mes~~ broadcas~g, of cow~e, is the second most important in terms of household amess and wfi confiue to serve a VW hrlporbnt rOle.
Whether delivered by cable or terrestrial broadcasting, the three major networks are the most important source of programmhyz.~Ju]es Be~sio, BellCore, personal cOJnmlmiCatiOn, Mar. 15, 1989.
74 ● The Big picture: HDTV and High-Resolution Systems
Figure 5-4-Data Transmission Rates and HoldingTimes for Different Types of Data Communications
Entertainment- – - v i d e o video— a u d i o Compact-disk I~ – — _ -— d a t a audio / ! 1
nl I II1
Teleconferencing II
wdeo / II I I
L
ndton
TelemetryI
1 \ I III I
r , J, , 1 , I I
10 10 10’ 104 10 10’ 10 10” 10’ 10 ‘(1 kdoblt) (1 megabit) (1 gigabit)
Channel data rate btts per second
High-resolution video is among the most demanding of applica-tions in terms of channel data rates, and also requires long holdingtimes.SOURCE: Stephen B. Weinstein, “Teleeommunieations in the Coming
Decades,” IEEE Speetrum, November 1987, p. 62. Used withpermission.
failures in the telephone system caused by softwaresuggest that additional development may be needed.
Consumer electronics has already had an enormousimpact on opto-electronics. Much of the leadingresearch in solid-state lasers has been, for example,by the companies that produce compact disk players.(Related high-performance solid-state lasers powerfiber-optic networks.) These revenues have in turnprovided the capacity to further advance the state-of-the-art.81
Despite pioneering fiber-optics and the electron-ics that drive signals through the fibers, the UnitedStates now lags Japan in many aspects of R&D andthe production of fiber and associated electroniccomponents. The United States still leads Japan,however, in linking these components together intocomplete communications systerns.82
PACKAGING/INTERCONNECTPackaging/Interconnect (P/1) is the set of technol-
ogies that connect all of these component8-semiconductors, displays, storage, and communica-tions-into fictional systems. To connect a siliconchip to the outside world, the chip is mounted in ada.stic or ceramic package that has tens to hundreds~f metal leads. ~-ese p;ckaged chips axe mountedon printed circuit boards which, in turn, me intercon-nected via standard multipin connectors on a moth-erboard or a backplane (figure 5-5).
The cost of these comections increases rapidly ateach level. Within the chip itself there are millionsof tiny wires of aluminurn comecting the transistorsformed in the silicon. Despite their complexity,these comections typically cost just $0.0000001each because they are all formed in a single stepusing a photomask. The cost of the comectionsbetween the chip and the package it is mounted onare roughly $0.01. The cost of comections betweenthe package and the printed circuit boaxd are roughly$0.10 each. And the cost of comecting the printedcircuit board to the backplane are roughly $1.00each.83 Overall, the cost of packaging/intercomect-ing and assembling the electronic components,together with testing the system, accounts forroughly 30 to 50 percent of the total for a complexelectronic system. Most system reliability problemsare due to intercomect failures, and P/I technologiesare a principal barrier today to achieving highersystem performance.
Today’s printed circuit board technology, forexample, etches individual circuit patterns in copperfoil laminated to sheets of fiberglass-reinforcedepoxy. Multiple layers of unique circuit patterns canbe kuninated together with more epoxy. Packagedsemiconductors and other components are thenmounted on the board and intercomected viacopper-plated holes to specific circuit patterns ondifferent layers of the board. These holes account fora large fraction of the total board area and limitwiring densities and component spacing, and thusslow attainable system speeds while increasingsystem size, weight, and costs. The mechanicaldrilling process used to form these holes limitsfurther reductions in their size.
81JWe~ J+ ‘1’ietje~ David S~offRese~h cater, Testimony kfore the Senate COtitt= on ~v~en~ Afffi> ‘ug. 1, 1989”82, ,Asesstig Japan’s Role in Telwomticatiom,“ IEEE Spectrum, June 1986.
‘sJohn S. h&yO, “Materials for Information and Cornmu.nicatiou” Scientific American, October 1986, p. 61.
Chapter 5-+5inkages Between HDW, HRS, and Other Industries ● 75
Figure 5-5-Levels of Packaging/Interconnect
Hybrid Mlcrocircult/Multichlp package
IC chips
e
!-v
BoardUnit assembly
BackplaneSingle chip package
Transistors are interconnected via metal leads on the integrated circuit and is then mounted in a plastic or ceramic package. The packagedintegrated circuit is mounted on a printed circuit board which, in turn, is mounted on a motherboard or a backplane.
SOURCE: Adapted from: National Security Industrial Association, Electronics Packaging/lntereonneet Task Force, “Eleetronies Packaging/lntereonneet: TheNext Crisis for Cost-Effeetive Military Eleetronies,” Washington, DC, draft, 1989.
As an example, intercomecting the 200 or sochips of a supercomputer processor, each chiphaving 250 input/output leads or more, wouldrequire a board with 40 layers of circuitry.~Advanced ICS may require 500 to 1,000 inches ofintercomect wiring per square inch of board-twoto three times the current practical limit. Designingand building such boards reliably is very difficult.
Small, high-density multichip modules are onemeans of improving P/I performance that is nowgaining favor.85 IBM’s 3090 mainframe, for examp-le, combines many chips in a ceramic module with44 layers of wiring. These modules are then mountedand interconnected via a printed circuit board withrelatively few layers.
In the longer term, the Japanese MITI and KeyTechnology Center flat-panel display consortium(ch. 2) intends to use the lithography and thi.n-ftitechnologies developed for large-area flat paneldisplays to advance these printed circuit boarddensities through improved and lower-cost ‘Chip-on-Glass” technologies.
C’hip-On-Glass technology mounts “bare” inte-grated circuits directly on lithographically printedglass substrates. This has several important advan-tages. Fine-line lithographic printing can provideperhaps ten times the wiring density attainable withthe conventional copper-epoxy printed circuitboards described above. Increasing the wiring den-sity also reduces the number of layers necessary.This reduces the space that must be allotted to theinterconnections between layers. The savings aremultiplicative. A conventional printed circuit boardwith 40 layers of copper-epoxy interconnect mightbe replaced with a lithographically printed glasssubstrate with just two layers of intercomect.8c Thisprovides substantial cost savings in both design andproduction.87
Mounting the bare IC directly on the substratebypasses several conventional packaging and inter-connect steps with further corresponding cost sav-ings and improvements in reliability. Together, thehigher wiring density and use of bare chips can allowsubstantial increases in how close components are
~Samuel Wekr, “For VLf+I, Multichip Modules May Beeome the Wckages Of ~oice, ” Electronics, April 1989, p. 106. This is roughly equivalentto a Fujitsu supercomputer.
8%id.
~~ically, a total of five layers might be needed: one for bonding pads for the chips, two for intereonnee~ one for power, and one for ground.LWBW ~m MCC, ~so~ Comxmmicat,iou My 12, Oct. 12 and 13, 1989; Jan. 23, 1~.
76 ● The Big Picture: HDW and High-Resolution Systems
packed. This allows higher speeds and reducessystem size and weight.
Chip-On-Glass technologies are used in special,high-performance cases today, but could be appliedmuch more widely if large area lithography tools andrelated technologies were available. These tech-niques would allow many glass substrates to beproduced at once on a large sheet, rather thantediously one-at-a-time.
The complexity and high speed of the chips usedfor HDTV will require the use of high-performanceprinted circuit boards. Complex multilayer printedcircuit boards will be necessary and new materialsmay have to be developed to handle the high speedsat an affordable cost.88 Although these are allavailable today in high-end commercial and militarymarkets, manufacturing in volume for the HDTVmarket might force rapid improvements in produc-tion technology and dramatically lower their price.
The United States seriously lags Japan in many ofthese P/I and related assembly technologies. Arecent National Academy of Sciences study foundthe majority of U.S. companies 4 to 5 years behindJapanese competitors in manufacturing process con-trol and in factory automation for fabricating,assembling, and testing electronics products.89 TheUnited States also lags Japan and Europe in the useof surface-mount technology for connecting the chipto the printed circuit board (figure 5-6).W Thistechnology saves space, increases reliability andperformance, and reduces assembly costs. TapeAutomated Bonding (TAB) technologies for pack-aging semiconductors, invented in the United Statesby GE but used more widely in Japan, offmsignificant increases in reliability at greatly reduced
labor and cost.gl In addition, TAB significantlyimproves semiconductor Performance.g*
Producing P/I equipment and materials for HDTVor, more generally, for the flat-panel display marketmay provide economies of scale to a firm as well.Shindo Denshi, the largest Japanese producer ofTAB tape, currently gets half of its sales fromsupplying producers of LCD displays. Some Japa-nese companies, such as Toshiba and Matsushit%have also developed proprietary “outer leti bon-ders” for comecting wires to the display. Thistechnology is not for sale and might make it moredifficult for U.S. firms to enter the market.93
Many P/I and related technologies-assefily,test, surface-mount, tape-automated bonding-havebeen pushed the hardest by the consumer electronicsmarket. The Sony Watchman television, for exam-ple, uses higher performance TM than the NECSX-2 supercomputer.w The consumer electronicsmarket demands high reliability, small size, and lowcost, but at the same time provides very large volumemarkets that allow even expensive, yet innovative,technologies to pay for themselves through long-term productivity improvements as experience isgained.
Because of these characteristics, consumer elec-tronics often pushes the state-of-the-art in manufac-turing technologies harder than lower-volume buthigher-profit markets-especially for assemblingcomponents or systems. If the HDTV market devel-ops, it may similarly provide manufacturers a testingground for developing new assembly technologieswith the volume needed to pay for themselves, aswell as gaining valuable experience in assembly ofsophisticated electronic systems that can be trans-ferred to many other products.
88J~&W, s~~ff~bs~~~~communi~tioq h. 15 ~dzo, 1989; Daytoq op. ci~, footnote 26. Attheveryh@ expected prOct%+ShlgspWds,better PC board material maybe needed than standard epoxy as it is too absorptive at these high frequencies. Teflon boards have desirable dielectricproperties and are now used in high-speed computers, but will be too expensive for the co nsumer market. It may be necessary to develop new materialsthat: are low-loss dielectrics; have good tempmture characteristics; and bond well to copper. Alternatively, some labs are trying to avoid these highspeeds on the board itself, which would require higher quality materials, by putting bus speed multipliers on each chip and thusrunning at the higherspeeds only within the chip itself.
g-~aca Studies Bored, “The Future of Electronics Assembly: Report of the Panel on Strategic Electronics Manufacturing Teclmologies”(W@@tOU DC: National Academy Press, 1988),
%esley R. Iversoq “Surface-Mount Technology Catches On With U.S. Equipment Makers-Finally,” Electronics, December 1988.gIJq L= ~J15 ~m ~elu=i*5 Big @le F~y Pay@ Off’?” Electroru”cs, Feb. 17, 1986; “~ON ~ghlights * Do*t ‘O1e
‘Ih~Automated Bonding Is Taking, “ Electronics, Feb. 18.1988. David Lsmmers and Ashok Bin@ “Tape Automated Bonding Sparks RenewedInterest” Electronic Engineering Times, July 10, 1989, p. 53.
%?~e~r Burg-, “~ for ~gh U() ad figh S@,” Se~”conductor Znternationul, June 1988.
g3J~~ G. pm= ad E. J~ V=- “HDTV D~elopm~ts in Jap~” TechSearch Internatio~ h., AUStiXL w 1989.
~“l%e Future of Electronics Assembly,” op. cit. footnote 89.
Chapter S=Linkages Between HDW HRS, and Other Industries ● 77
Figure 5-6-Fraction of Integrated Circuits Used in Surface-Mount Packagesfor Japan, Europe, and the United States
‘“~-–– –60
50
40
30
20
10
DB
B
B
‘~~1987 1988 1989 1990 1991 1992
SOURCE: Wesley R. Iverson, “Surface-Mount Technology Catches On With U.S. Equipment Makere-Finally,” E/ecfronic.s, December 1966, p. 116. Usedwith permission.
Chapter 6
Advanced TV Markets and Market Uncertainties
INTRODUCTIONThe future growth and ultimate size of the various
advanced TV and related markets is unknown. Thismay be unimportant for ATV-related R&D. Themere expectation of a large market seems to beinducement enough for many manufacturers to pushthe state-of-the-art in ATV technologies (ch. 5). Incontrast, the growth rate and size of these variousmarkets directly determines the economies of scalein production of the component technologies (e.g.,displays, semiconductors, and data storage) as wellas in the manufacturing processes for sophisticatedelectronic systems.
Development of HDTV markets depends on manycomplicated and unpredictable factors.1 How muchconsumers will be willing to pay for HDTV-qualitypictures is a matter of speculation. Consumers mayprefer to spend their disposable income on a widervariety of progr-g choices, other entertain-ment systems, lower resolution interactive videosystems, or more basic staples or amenities. Forthose consumers that are interested in purchasingHDTVS, there remain a series of “chicken-and-egg’ problems to overcome:
●
●
Consumers are unlikely to buy receivers until awide variety of HDTV broadcasting is availa-ble; but broadcasters hesitate to offer HDTVprograms until there are enough viewers tomake it worthwhile.Consumers are unlikely to buy receivers untilthe cost comes down to acceptable levels; butmanufacturers can’t get the cost of receiversdown until millions of consumers are purchas-ing HDTVS and enable large-scale mass pro-duction and streamlined manufacturing proc-esses.
In contrast to the U.S. market-driven approach,the Japanese are actively trying to overcome theseproblems in the interest of developing their domesticmarket. This approach carries both risks and poten-tial benefits: they may fail at a large scale but theirchances of success are improved.
The uncertainties in the details of how the HDTVand other ATV markets will develop should notobscure important underlying trends in the technol-ogy. No matter what form video entertainmentsystems take in the future, they will increasingly usedigital electronics for higher performance at reducedcosts. For example, digital microprocessors arealready common in television tuners and IDTVShave recently come on the market.
The rate at which fiber-optics is extended to thehome and how it is connected may be similarlydebated, but not its eventual widespread use. Imp-roved economics and superior performance of fibersystems make this transition inevitable. The merg-ing of digital video entertainent systems and fibermight then make a host of new information servicespossible. The question is how to provide a flexiblefiameworkfor linking fiber to the home amibringingdigital video systems into the home with the leastoverall disruption to the national telecommunica-tions tiastructure-all at an affordable price.
A few of the possible market scenarios will firstbe examined qualitatively; then market projectionsby several leading groups will be discussed. Thetrend in technology towards digital/fiber systemswill be reviewed. Finally, some of the marketconflicts generated by the coming introduction ofATV services will be examined.
SCENARIOS FOR MARKETDEVELOPMENT
No one knows how the ATV market will develop.HDTV might prove irresistible to consumers and themarket may grow rapidly, with corresponding econ-omies of scale quickly lowering costs and makingHDTV available to most families-perhaps even asecond set. The HDTV market might grow moreslowly: growing only after much larger and im-proved display technologies are available; afterconsumers are sensitized to the value of its higher
1~~ iS fO~u.sed on ~p~~~~~yhme ~~~se it represen~ a si~lc~t c~~tive c~ge from to~y’s TV ht~y Or may nOt h made ill mU@yone step, depending on the choice of transmission standards.
-79–30-368 - 90 - 4 : QL 3
80 ● The Big Pictlu-e: HDTV and High-Resolution Systems
picture quality; and after quality programming isavailable.2 Less advanced versions of ATV such asIDTV or EDTV might prove more than adequate formost consumers, limiting HDTV to a small high-endresidential market and to movie theaters, bars, orrestaurants. Or perhaps lower resolution interactivevideo systems will reach a level of performancesufficient to stimulate signiilcant consumer interestand vie with advanced TVs for consumer dollars.
An advanced form of HDTV might become thehome information center, providing entertainment,computer, and telecommunications services in asingle generalized piece of equipment. Altern-atively, and perhaps more likely, these differentservices will instead continue to be primarily pro-vided by separate, somewhat specialized pieces ofequipment. Even in this case, the continuing declinein the cost of computer and advanced telecomsystems will allow them into an ever larger fractionof households; while the home ATV continues toprimarily provide entertainment.
In the longer term, having an advanced HDTVwith the capability of interactivity-the equivalentof a computer, but for video images-might be veryattractive to consumers. These systems might allowconsumers to create personalized newspapers;browse distant databases; request more in-depthinformation on a news program; transmit an interest-ing movie clip to a friend; or even use an electronicyellow pages to see a video clip of the inside of arestaurant they want to try. Alternatively, suchservices might be provided by a telecom/computersystem while the HDTV, again, simply providesentertainment. Or consumers may decide that suchinteractivity is not worth the cost or effort in eithercase.
Consumer Demand for HDTV
Little reliable information is available on con-sumer demand for high-quality video entertainment;the consumer response to higher quality pictures isambiguous. Many consumers watch TV without thebenefit of either rooftop antennas or cable, depend-ing instead on rabbit ears or internal antennas. Manyviewers still watch black and white TVs, at least asa second set—some 3.5 million B&W TVs were soldin the United States in 1987.3 On the other hand,cable TV has now penetrated over half of Americanhouseholds—both for the higher quality of picture itprovides, and especially for the increased variety ofprograms it offers. Further, large screens are populardespite their high cost? The market for 25-inch andlarger screens increased from 2.8 million in 1982 to5.5 million in 1987.5
HDTV market research is similarly ambiguous.The few studies performed indicate that whencompared to studio quality NTSC on today’s rela-tively small displays, the viewer preference forHDTV is tempered by program content and viewingconditions.G For example, when these (18- and28-inch) displays were viewed horn a distancewhere the additional detail provided by HDTV waslost, then viewers naturally had little preference. Butviewed up close, 75 percent of the viewers prefemedHDTV. These tests may simply reaffirm that toappreciate HDTV large displays are needed. Sony,for example, believes that a 72-inch diagonal isrequired to portray the full capability of their1125/60 system.7
It is important to note that the pictures shown inthese tests were studio quality. For NTSC, this is asignificant improvement over what viewers nor-mally see at home: the quality of NTSC TV is oftenpoor due to transmission degradations.8 For any new
WCC Advisory Committee on Advanced Television Services, PIarming Subcommittee Working Party 7: Audience Researcb Repo% Feb. 14,1989.Tbis report notes the potentkd di.ftlculties posed by an audience tbat becomes more sensitive to flaws after a fixed standard is locked into place.
sElec~onics ~dus~es Associatio~ Consumer E2ecrronics Annuu/ Review, 1988 cd., Wmhingtoq DC.
d~eMc~ght, W. Russell Neurn~ Mark Reynolds, Shawn O’Domell, and Steve Scbneider, ‘‘The Shape of Things to Come: A Study of SubjectiveResponses to Aspect Ratio and Screen Size,” ATRP-T-87, Massachusetts Institute of Technology Media Lab, May 17, 1989.
5~omson Commer E1ec~ofics, E]~@onics rndus~es Association, HDTV ~ol-mation packet 1989. Elec&onics Industries AssociationConsumer Electronics Annual Review, 1988 cd., Washington, DC.
6W+ Ru~sell Ne~an, Summe Ctibliss Neil, he Mc~@~ and s~~ O’Donnell, “ A c t i o n Merno,” H o u s e &lbCOmXdfiee O nTelecommunications and Finance, Feb. 1, 1989; W. Russell Neurnan,‘‘The Mass Audience Looks at HDTV: An Early Experiment,” paper presentedat National Association of Broadcasters’ Annual Conventio~ Las Vegas, NV, Apr. 11, 1988.
7LW ~oqe, Sony Cow., perSOnaI communicatio~ Oct. 12, 19*9.
gwi~w F. Sctieiber, Massachusetts IrMitute of Technology, ‘‘Comments Before the FCC on Advanced Television Systems and Their Impact onthe Existing Television Service, ” Nov. 30, 1988.
Chapter Udvanced TV Markets and Market Uncertainties ● 81
standard, its susceptibility to transmission degrada-tions are a key consideration (ch. 4).
Consumer demand for HDTV receivers willdepend on the cost of the set and the availability ofprogram material to view. In turn, receiver costs andprogram availability will be determined by the sizeof the market. These factors delayed the takeoff ofthe color TV market by some 6 or 7 years, fromroughly 1955 to the early 1960s. Only the persever-ance of RCA, led by David Sarnoff, and theinvestment of about $3 billion (1988 dollars) keptcolor TV alive until the necessary critical mass ofcolor programming became available and enabledthe market to grow. Similarly, VCRs took off whenrental stores for tapes became common (figure 6-1 ).9
The response of broadcasters to market demandfor color programming in the early 1960s was lessambiguous. Broadcaster costs are roughly constant.A small change in market share (or ratings) then hasa signitlcant impact on profit margins. Therefore,broadcasters must compete vigorously for everypossible market advantage. Between 1963 and 1966the share of prime-time hours offered in color wenthorn 20 to 100 percent, even though less than 10percent of all households had a color TV (figure 6-2).Small changes in market share had enormousleverage over the broadcasters.
Consumer Demand for Interactive Video
Similarly, little reliable information is availableon consumer demand for interactive video. Comput-ers, rather than HDTVS, could be the platform forinteractive services in the future. Interactivity islargely a fimction of flexible computing power andgood software-in sharp contrast to the very highspeed but relatively inflexible signal processingdone by an HDTV. Declining costs have alreadybrought computers into about 20 percent of Ameri-can homes.l” A recent survey, however, found thathome computers are used, on average, just twice perweek; and many question whether they will pene-trate the remaining 80 percent of households any
Fiaure 6-l—Sales of VCRs, VCR Prices, and theGrowth of VCR Tape Rental Stores
1 500 ~ -- T– ----––-. 4 Mean >.....,& 1 250
/
VCR “...,: : 1 alla-l
. . .price . .
4
. . .~Qm. 7’)0 Video ““”””-...;g
500
i
specialty “-...“%. ..,..
: stores .. ...,,
2(>0
o }-—--— —---144
J1 2
21L
- 3 0 (000s)-25- 2 0
- 1 5
- o
- 5
75 76 77 78 79 80 81 82 83 84 85 86 87
Note that the growth of the VCR market corresponds more oloselywith the opening of VCR tape rental stores than with the declineof VCR prices.
SOURCE: Boston Consulting Group, “Development of a U.S.-based ATVIndustry,” May 9, 1989 for the American Electronics Associa-tion. Used with permission.
time soon. On the other hand, the potentially largehousehold penetration of advanced TVs might allowthe delivery of a limited range of user friendlyinteractive information services to more of thepublic.
Although interactive videotex systems interestmany, experimental systems tested by Knight-Ridder in Miami (1983-86) and Times Mirror insouthern California (1984-86) failed. New effortsare underway by Bell South in Atlanta, NYNEX inVermont, and IBM/Sear’s Prodigy throughout theNation.1 1
In France, the Government-backed interactiveMinitel system has been widely noted as a success.There are already some 4 million subscribers,although figures for 1988 showed an 8 percentdecrease in residential use of Minitel services overthe previous year. ~z Minitel succeeded by makinguse of the existing telephone network; by giving thesubscribers interactive terminals; by keeping the
gsome~venot~ ~t~eBe~~ VCR format and RCA videodisc were hurt by a corresponding lack of pro~g. Othersnote ~ttheProblemswitb both those systems were inadequate technology: the Betamax format only provided 1.5 hours of recording time; the VHS format could givesign.i.flcantly more. Similarly the videodisc technology generated fuzzy pictures compared to alternatives. John Weaver, Liberty Television personalCOIIlllMdCdiO~ Oct. 12, 1989.
lw.s. Dep~ent of Commerce, Bureau of the census, “Statistical Abstract of the United States,” table 1286, 1988 ed.llFr~c Sal.mier, “The~blic Network Goes On-Line, ” Te2ephony, Apr. 3, 1989, p. 26; John Markoff, “Betting on a Different Videotex Ida” New
York Times, July 15, 1989, p. D5.12~e.Mfie RouSSel, “~tel Gets New T~~~,” Communications Week, Apr. 17, 1989, p. 52.
82 ● The Big pic~re: HDIVand High-Resolution Systems
Figure 6-2—Penetration of Color TVs Into American Households, Average TV Prices, and the Share of Prim*TimeProgramming Devoted to Color TV
Color TV penetrahonprogramming and real prices
‘001——————
90
I
Price of“ \ average
\ color TV80 \
\I \
I\
Percent of 10 \ \prime time \ ----and CTV \households G. - .
\\
‘ - — .50
r —Pnme- t imeh o u r s In
/ c o l o r ( % ) I
2,500
/
- - -1-— \ TV households\ \ with color TV
\ (%) ,-.
40-Price of
30-averageB&W TV
20-
lo-
7. \\ \ \o I I I I - - - - -~~1954 1957 1960 1963 1966 1969 1972
Year
-2,000
Prices Inreal dol lars(1988 )
-1 ,500
-1 ,000
-500
Note that the prime-time hours in color increased very rapidly beginning in 1963 and that the household penetration of color TVs followed,while the price of color TVs deolined only slowly.SOURCE: Boston Consulting Group, “Development of a U.S.-based ATV Industry,” May 9, 19S9 for the Ameriean Electronics Association. Used with
permission.
terminals simple13 and their use intuitive; and byproviding service on a pay-as-you-go basis.14 Thiskept total investment down; allowed potential usersto experiment without having to invest heavilybefore knowing what they were going to get; andavoided the ‘chicken-and-egg’ problems with usersdemanding abroad range of services before signingon, and service vendors requiring many users beforeproviding services. At this point, Minitel appears tobe no more than a niche market, which has yet to payfor itself (at least in accounting terms), and maybeshifting towards more of a business market.15
Government Policies and Market Development
Government policies, particularly those of theFCC, could have a powerful influence on the way the
HDTV market develops. The marketplace mightestablish early standards for VCRs (which are notoverseen by the government), for example, that areincompatible with government-backed standards forterrestrial and satellite broadcasting. If such incom-patibilities arise between different media or if thereis confusion in implementing the standards, consum-ers may hesitate in purchasing an expensive ATVthat has only limited applicability. This couldsign~lcantly slow ATV market development.
If the FCC allows ATV standards to improveincrementally, with IDTV, EDTV, and other gradualimprovements in performance, then consumers maydecide to stop short—at least in the near-term-ffull HDTV quality, depending on price and viewing
13~~ t~ are now believed to have been too simple and are being upgraded.ldJ~n ~efiw and Gorges Nshou “A Videotex Suczess Story,” Telephony, July 27, 1989, P. 46.
ISJsrnes Ma&harq “France’sMinitel Seeks a Niche,” New York Times, Nov. 8, 1988, p. D1.; Roussel, op. cit., foofnote 12; Steven J. Marcus, “TheFrench Videotex Connection” Zssues in Science and Technology, fall 1987.
Chapter t%4dvanced TViWarkets andillarket Uncertainties ● 83
preferences. If the standards do not include interme-diate enhancements, but move directly to fullHDTV, then that market may or may not developdepending on consumers’ perceptions of the prod-ucts usefulness. If NTSC were phased out over anumber of yearn in order to make better use of thebroadcasting spectrum, then consumers would haveno choice but to buy A~s.
If the FCC requires stations to show the samepicture in both NTSC and in HDTV simultaneously,then consumers may decide that the additionalquality for the same programming is not worth it.Alternatively, HDTV quality programs and broad-casts might be reserved for special high-end mar-kets, such as pay-cable. Producing and broadcastingHDTV quality programs on special, high-end chan-nels alone presents the same “chicken-and-egg”problems: who invests first-viewers, broadcasters,or program producers? Even detailed consumerstudies and market tests can not resolve some ofthese questions. Ultimately, risks must be taken,particularly by program producers and broadcastersto provide pictures that consumers will want to see.Clear and consistent government policies may beimportant if potential investors are to take theserisks.
MARKET PROJECTIONSDespite the lack of audience-based HDTV market
research, a number of analysts have projected HDTVmarkets in the United States. Their forecasts havebeen made by analogy, i.e., by modeling HDTVpenetration rates after those of previous successfulconsumer products such as black and white TV inthe late 1940s, color TV in the early 1960s, andVCRs in the late 1970s.16 There are a number ofremarkable market successes such as these. Therehave also been many market failures, such as thevideodisc, quadraphonic sound, stereo AM radio,and others, due to factors ranging from poortechnical performance and confusion over standards,to misperception of consumer interests.17
Most market projections for HDTV begin byexamining the high-end markets where expensive,early production sets will likely be sold. Marketanalysts then estimate the rate at which cost reduc-tions from economies of scale can bring the price ofsets within reach of middle and lower incomemarkets. These models implicitly assume that HDTVwill eventually be successful. Furthermore, most ofthese models assume that the availability of IDTVSand EDTVS will not dampen the market for HDTV,but will instead be part of a natural progression toHDTV.
The projected market penetration rates for suc-cessfulproducts normally follow an S-shaped curve.18Sales are initially slow. When sales reach a market“take-off” point-typically observed to be a fewpercent of household penetration-they then growrapidly. After several years of rapid growth, themarket begins to saturate and sales level off toreplacement levels. This behavior can be seen in theB&W TV, Color TV, and VCR markets for theUnited States and Japan (figure 6-3). Eventually,new products may come along that displace theearlier model and cause its sales to declin~such asthe case of Color TV replacing B&W. The twoprincipal variables in such scenarios are: 1) at whattime the market ‘takes off’ and 2) how rapidly themarket grows after take-off.
HDTV market projections generally follow thisform. The Japanese Ministry of Posts and Telecom-munications (MOPT) assumes that the Japanesemarket will take off in 1993. In the United States, theElectronic Industries Association (HA), NationalTelecommunications andInformationAgency (NTIA),and American Electronics Association (AEA) as-sume that the market will take off in roughly 1994,1997, and 2000 respectively. All four of thesestudies have reasonably similar rates of growth oncethe take-off point is reached (figure 6-2).
16E~ples of such st@ies include: Robert R. Nathan Associates, kc., “Television Manufacturing in the United States: EconomicContributions--Past, PresenC and Future,” Washington DC, Electronic Industries Associatio~ Nov. 22, 1988; David Russell, “High DefinitionTelevision @D’f’V): Economic Analysis of Impact,” Washington DC, American Electronics Associatio~ November 1988; Larry F. Darby, “EconomicPotential of Advanced Television Products,” Washington DC, National Telecommunications and Information Administration+ Apr. 7, 1988. Reviewsof these and other market studies are included in: Shawn O’Donnell, “Forecasting the HDTV Marke6° Neumam et al., Feb. 1, 1989, op. cit., footnote6, appendix; and in Kenneth R. Donow, “EIDTV: P1arming for Actio%” Satellite Systems Engineering Inc., Bethesda, MD, for National Associationof Broadcasters, April 1988.
17Steven p+ scm, M~ga*”~take~: F~r~~~~li~~ ~~ the MYth of Rapid Technological c~nge (New York NY: Free press, 1988).
WIISUCCCSSfUI products foffow other paths, including upside down “u” S.
84 ● The Big Picture: HDTVand High-Resolution Systems
Figure 6-3-Sales and Projections of Video Entertainment Technologies in the United States and Japan
U.S. domeatic sales (miilion.a)20
1 5
10
5
o45 55 65 75 85 95 105 115
Calendar year
— Black 6 white TV ‘+ Color TV * V C R
= HDTV, Darby - HDTV, Nathan + HDTV, Ruatell
8
6
4
2
0
Japanese domestic aale.9 (milliona)
4 5 5 0 55 60 65 70 75 80 85 90 95 100 105 110Calendar year
— Black & white T V + C o l o r T V - V C R a + HDTV
SOURCES: United States: Electronic Industries Association, “Consumer Electronics Annual Review,” various years; Robert R. Nathan As-”ates, Inc.,“Television Manufacturing in the United States: Economic Contributions-Past, Present, and Future,” Washington, DC, Electronic industriesAssociation, Nov. 22, 1988; David Russell, “High Definition Television (HDTV): Economic Analysis of Impact,” Washington, DC, AmericanElectronics Association, November 1988; Larry F. Darby, “Economic Potential of Advanced Television Produets,” Washington, DC, NationalTeleeommunieations and Information Administration, Apr. 7, 1988.
SOURCES: Japan: Eleetronie Industries Association of Japan, various years; James Howard Wooster, “lndustnal Policy and International Competitiveness:A Cass Study of U.S.-Japan @repetition in the Television Reeeiver Manufacturing Industry,” Ph.D. Thesis, University of Massachusetts, 1986;Bob Johnstone, “Programming Better Quality TV,” Far Eastern Econornk Review, Aug. 11, 1988, p. 53.
Others have suggested that the HDTV marketcould be a big flop,lg emphasizing the high initialprice of HDTVs and questioning whether consumerswill see enough additional value over their currentsets to justify paying the difference. IDTV andEDTV may themselves fill consumers’ needs. Inter-active video might create a new competing market,limiting sales of HDTVS (if Open ArchitectureReceiver designs are chosen as a standard theninteractive video and HDTV might be the samemarket). The experts-the managers of many of theworld’s consumer electronics fins-have alreadybet more than $1 billion that there will be a bigmarket for HDTV. Whether they are right or wrongwill only be clear in hindsight.
The value of the HDTV market is potentially verylarge. The Electronic Industries Association hasestimated that HDTV receiver sales in the UnitedStates could total 13 million units worth about $12billion20 retail (1988 dollars) in 2003, $4 billion
more than would be the case if the market continuedwith standard TVs alone. The manufacture of movieproduction and broadcasting equipment would in-crease these values even more. Lower estimatesinclude the NTIA Report with sales of $5 to $10billion in 2003 for 12 million receivers and $3 to $6billion for 10 million VCRs; and the AmericanElectronics Association report with 2003 sales of 5million HDTV sets in the United States worth about$5 billion and world sales of about $14 billion. AEAprojections give VCR sales of 4 million units addingabout $3 billion in the United States; with $8 billionin sales worldwide in 2003. In comparison, theoverall U.S. consumer electronics market was worth$30 billion in factory sales in 1987.21
These estimates are for home entertainment alone.Commercial and industrial uses may also be impor-tant. The Japanese MITI, for example, estimates thatsales of consumer goods will be just 60 percent ofthe total sales of HDTV by 2000.22
lgsee, for example, Bob Davis, “Will High-Definition TV Be A Tbrn-Off?” Wall Street JournuZ, Jan. 20, 1989, p. B1.20M1 v~um here me given h 1988 dollars. EL+ values were discounted to 1988 dollars at 3.5 percent -ud ~ation rate ~m tie ~ es~te ‘f
$20 billion in 2003. NTIA and AEA published values were assumed to already be in 1988 dollars as they cited no inflation rates.21 Ekxtrcrnics Industries Association, op. cit., footnote 3.
earning of a Wide Crispness, ”=4 <Television Makers h ‘r
Business Week, Dec. 21, 1987; and James G. Psrker, TechSearch International, Inc.,personal communicatio~ Oct. 6, 1989.
Chapter (A4dvanced TV Markets and A4arket Uncertainties ● 85
Many widely circulated projections also containhighly questionable assumptions and numerousinternal inconsistencies. For example, one reportprojects that the total TV market in the United Stateswill double in the next 15 years, from 21 to 40million sets-without identifying who will purchaseall these sets as population growth slows and thepopulation ages. It insists that most of the valueadded will take place in the United States, butglosses over the fact that the high-value manufactur-ing of electronics for tie consumer TV market istoday almost entirely offshore-leaving primarilyscrewdriver assembly operations in the UnitedStates. This report also ignores that with the longerterm transition to flat panels or projection systems,much of the manufacturing and assembly remainingin the United States could also move offshore. Thereport assumes that IDTV and EDTV will have noimpact on the HDTV market.
The precise course that Am technology will takeand the pace that it will enter cannot be predicted.The development of the ATV market will depend ona variety of factors, including: the rate of technolog-ical advance and the achievement of large econo-mies of scale and learning; the consumer response toIDTV, EDTV and HDTV; the development ofmovies and programs at the appropriate level ofvisual quality for the ATV technology of choice; andthe policies and standards chosen by the gover-nment, among others. Improvements in televisiontechnology are inevitable, but when and in whatform they will take is largely unknown in the longerterm.
MARKET PROJECTIONS AND THERELATIVE IMPORTANCE OF HDTV
Forecasters have typically compared the relativeimportance of the HDTV, computer, and othermarkets by: 1) projecting constant growth rates intothe future; 2) comparing the semiconductor contentor other component demands for assumed marketsizes; or 3) comparing ultimate market penetrationrates and corresponding needs per user. None ofthese are very satisfactory or illuminating.
Extrapolating Current Growth Trends
Extrapolating current or expected growth trends atthe same compound amual rate far into the futurehas been one of the most common techniques formaking projections, but this form of analysis can
lead to highly exaggerated claims if done carelessly.For example, one report (“A”), has projectedconstant growth rates of about 5 percent in U.S. TVsales-with about 40 million TVs sold by 2003. Asecond report (’‘B”) has projected a constant annualgrowth rate of 10 percent for the computer industryand 7 to 8 percent for the semiconductor industrythrough 2008. Similarly, a third report (“C”) hasprojected constant annual growth rates of HDTVSand PCs, resulting in annual world sales reaching180 million HDTVS and 135 million PCs by 2008.
These projections ignore the simple point thatsales cannot grow exponentially forever. Marketsinevitably saturate. For example, assume that theHDTV market will primarily be in the industrializedcountries over the next 20 or so years and thatreplacements rates will be similar to the 7 year normexperienced for color TV today. With these assump-tions, Report C’s projection of 180 million HDTVSsold annually in 2008 is roughly equivalent to anultimate penetration rate of 1.25 HDTVS for everyman, woman, and child in the industrial countries.Their projection for personal computer sales, 135million annually in 2008, gives similarly highpenetration rates. These are remarkable, and perhapsunrealistic, penetration ratios.
One can foresee a time that most people in theindustrialized world will have a personal computerand an advanced TV. Most people in the UnitedStates today have direct access to both a TV and atelephone. There will also be a growing demand forcomputer technology from the developing countries.Making constant growth projections, as these reportsdo, is reasonable over short periods of time. Overlonger periods markets are likely to saturate andgrowth projections must be reduced for when thisoccurs. Even if these very high levels of penetrationare reached in the next 20 years, some marketsaturation will likely occur; but this is not consid-ered in the constant growth rate projections made bythese groups.
Comparison of two markets in the distant futurebased on constant annual growth rates is extremelysensitive to the assumptions. For example, report“C” assumes that personal computer sales willgrow from current annual sales of 20 milhon at a rateof 10 percent; and that HDTV sales will grow fromsales of about 5 million in 1994 at a rate of about 30percent annually. At these rates, HDTV sales aregreater than PC sales after the year 2000. But if
86 ● The Big picture: HDTV and High-Resolution Systems
HDTV sales instead grow at a still remarkable paceof 20 percent annually, their total sales are slightlyless than half those of PCs in the year 2008. Suchsensitivity to (as yet) totally unknown underlyingparameters makes such forecasting unreliable.
Comparing the Semiconductor or OtherComponent Requirements for Assumed
Market Sizes
Forecasters have also compared the importance ofdifferent markets by the relative requirements forsemiconductors and other components. This com-pounds the uncertainties of market growth rates witha lack of regard for the relative importance of varioustypes of components. For example, some havecompared the semiconductor content of HDTVS ona simple dollar basis with that of the entire semicon-ductor industry-rather than within the relevantmarket segments-and then dismissed HDTV asunimportant as its total semiconductor demandmight be relatively small.
With such logic almost any market segment, whencompared to the entire industry, can be dismissed astoo small to be important. Supercomputers, minisu-pers, and parallel machines combined were a marketof just $1.36 billion in 1988.23 Assumin g a semicon-ductor content of 10 percent, they would consumejust $136 million in semiconductors, compared to atotal world semiconductor market of some $54billion in 1988. Few would argue that the supercom-puter industry is unimportant, however; it drives thestate-of-the-art in a variety of chip, packagingintercomect, and other technologies as well as thatof computers.
Similarly, (MOS) DRAMs were just a $2.5 billionmarket in 1987 (before the price was driven up by theMITI-coordinated production cutbacks),24 or lessthan 5 percent of the total world semiconductormarket. But DRAMs drive many important semi-conductor manufacturing technologies and the lossof DRAM manufacturing has been a serious concern
for both U.S. semiconductor producers and U.S.computer makers. Few would argue that we can dowithout this market segment despite its small size.
Semiconductors are expected to account for some10 percent of the retail cost of HDTV receivers.~World sales of HDTVS in the various high-growthscenarios might reach $25 billion or more by 2003.The corresponding HDTV semiconductor contentwould be $2.5 billion. In comparison, the worldmarket for semiconductors today is ah-eady morethan $50 billion with continued dramatic growthexpected. Obviously, HDTV would not drive theentire semiconductor industry, but it could add todemand.
It would be a mistake, however, to dismiss theimportance of HDTV within specific market seg-ments. In high-growth scenarios, world HDTV useof DRAMs in 2003 could be as much as five timestotal world DRAM production in 1987 (ch. 5). Butthis says nothing about the relative importance ofthis market for DRAMs compared to other uses in2003. Consider the following scenarios of demand.Both high and low (a tenth as large as the high)scenarios are presented to show the sensitivity of theforecast to underlying assumptions and to cautionagainst the potential pitfalls in extrapolating con-stant growth trends into the future described above.
Assuming that the world computer, office auto-mation, and telecom/fax markets (the principal usersof DRAMs today) grow at an annual rate (by numberof units) of 10 percent over the next 15 years and thatthe intensity of memory use within these productsalso increases, the resulting total annual growth ratein memory demand (by capacity) over this periodwould be 20 to 40 percent.2G At a 20 percent growthrate for all DRAMs, the high-growth scenariosuggests that HDTV would represent 25 percent ofthe total world demand in 2003. But at a 40 percentgrowth rate for all DRAMs, HDTV would accountfor just 3 percent of the world DRAM demand. Low
~Te~s. perry and Glenn Zorpette, “Supercomputer Experts Predict Expansive Grow” IEEE Spectrum, February 1989, p. 26.24C~1es H. Ferguso@ “D~,s, Component Supplies, and the World Electronics Industry: An hematiOMl Stitegic ~Ysis*” ‘1 ‘emo
89-554, August 1989, Massachusetts Institute of Technology, Cambridge, MA; Kenneth Flamm, ‘‘Policy and Politics in therntemational SemiconductorIndustry,” SEMI 1SS Seminar, Newport BeacL CA, Jan. 16, 1989; and Kenneth Flamm, Brookings Institution personal conmmnicatiow Oct. 25,1989.
~El~@ofics ~dus~es Association “Commer Elec@onics, HDTV, and the Competitiveness of the U.S. Wonomy,” House SUbCOlllrrdttee OnTelecommunications and Finance, Feb. 1, 1989. Semiconductors will account for a much higher fraction of factory costs, typically 20 percent or more.In the early stages of HDTV production the semiconductor content will probably be higher yet.
z6B=ause the -ket v~ue per~t capability has decked so rapidly for computers and their underlying semiconductor t~hnologies> it is ‘i~~there to specify that this is by unit capacity rather than by dollar value.
Chapter (L4dvanced TV Markets and Market Uncertainties . 87
growth scenarios for HDTV would reduce thesemarket shares to 2.5 and 0.3 percent respectively.
In high-growth scenarios, the demand for digitalsignal and other processing chips for HDTV in 2003(measured by processing capacity) could similarlybe as much as 10 times world production of all logicchips in 1987 (ch. 5). HDTVS might contribute asmall to signMcant additional demand to the totallogic market, depending on the assumptions onechooses about growth in the computer and othersectors. In 1989, however, DSPS are forecast to bejust 4 percent of the total dollar-value of the logicmarket. 27 DSP production for all other purposeswould therefore have to increase at a phenomenal 45percent annually for the next 15 years, or 250 timestotal, just to equal the high-growth scenario demandof HDTV.X Even in low-growth scenarios one-tenththis size, HDTV would still provide a siqi.ficantdemand for digital signal processing chips.
More generally, High-Resolution Systems (whichinclude many computer, office automation, andtelecom/fax markets) are likely to account for animportant share of DRAM and DSP use in almostany scenario, due to the special processing require-ments of video imaging.
If HDTV does provide a large market share forparticular semiconductors such as DSPS, it couldprovide a firm significant economies in production.This has been widely observed for other consumerelectronics components and markets. Producing aparticular 16-bit digital-to-analog (D/A) converterchip for the large CD-player market, for example,drove their price from $75.00 each to just $3.75,while the price of less complicated 14-bit D/A chipsthat did not benefit born such a large market demandstayed at $60.00.29 Such production economies canalso extend to individual firms that have the benefitof supplying several markets. This will also tend toinsulate the firm from the cyclical swings in demandfrom a single narrow product segment.
As notedin ch. 5, HDTV is driving the state-of-the-art in many aspects of display technology, and theproduction volume of displays for HDTV is poten-
tially enormous-as much as 6,000 times larger areaof AM/LCD display than is currently produced inthe world. Whatever form video entertainment takesin the future, it is likely to continue to be a principaldriver of technology and production economies ofscale for displays.
Finally, the production of HDTVS may be impor-tant in driving the state-of-the-art in some aspects ofthemanufacturingof sophisticated electronics. High-volume production of consumer electronics has longdriven the state-of-the-art in system manufacturingand has had important spinoffs to many othersectors. For example, all the Apple Macintosh’sproduced for the North American market are madeon a TV assembly line in California that wasimported from Japan.30 The development of TapeAutomated Bonding, surface mount technologies,and much automated insertion equipment was simi-larly driven by the consumer electronics industry(ch. 5). Large-volume manufacturing of HDTVSmight also teach important lessons in manufactur-ing, especially because of the sophistication of itselectronics.
Comparing Ul$i”mate MarketPenetration Ratesand Corresponding Needs per User
Although technological progress will continue,the total production volume of computers, AWS,and telecommunications equipment, etc. will atsome point in the future likely slow to roughlyreplacement levels: there will be a limit to thenumber of computers, ATVS, videophones, etc.people will be able to put to effective use. At thattime, the number of such devices used per personand their relative semiconductor and other compo-nent requirements will determine the importance ofeach market to the semiconductor industry.
Each household, for example, might have anATV, computer, and videophone at home; a car withsome electronics; and a computer/videophone at theoffice. In addition, there would be a number of‘‘smart’ devices such as microwave ovens, washingmachines, coffee makers, and others that would use
mS= Ch. 5; and k~gmted Circuit Engineering Corp., “Mid-Term 1988, ” Scottsdale, AZ.~su~h ~~ is po~~ible, ~ s=n ~ tie Pmt in tie D~ ~d other ~ets. H -h ~o@tion or other such system enter the market ill hge
numbers over the next 15 years, DSP market growth might be signifkantly larger than this. This assumes that the cost per gate for DSP is the same asfor the larger logic market.
mAI Ke~ck Natio@ Semiconductor, personal communication, MM. 15, 1989.
~erguso~ op. cit., foomote 24.
88 ● The Big Picture: HDTV and High-Resolution Systems
some semiconductors.31 In this scenario, the AWwould probably have the highest value display. Itmight then be the principal driver of displayelectronics and screen technologies and of theirproduction costs. The computer would continue tobe the principal driver of microprocessor technologyand costs. The computer, videophone, and ATVwould all play a role in driving fiber and widebandswitching technologies, with the Am perhapsplaying an especially important role in leveragingthe extension of fiber to the home. The computer andAW would share in driving magnetic and opticalstorage media technology and costs, with the ATVtaking a leading role in the near future due to itshigher rate of information flow. From this perspec-tive, Advanced TV could play an important role inseveral sectors.
Basing policy decisions on precise predictions offhture markets in the electronics sector is clearlyimpractical. The market for HDTV, as envisionedtoday, may or may not develop. HDTV might openup entire new markets unforeseen today; or evensubstitute for personal computers and telecommuni-cations equipment both at home and in the office.Alternatively, HDTV might be a flop and interactivecomputer systems prove to be a winner. Such marketuncertainty makes investment decisions and policyformulation difficult; in the final analysis, faith maybe an important factor in market success.
TECHNOLOGY ANDMARKET TRENDS
The important underlying trends in the technol-ogy should not be obscured by the uncertainties inhow the HDTV and other ATV markets willdevelop. No matter what form video entertainmentsystems take in the future, they will increasingly usedigital electronics.
Digital systems allow the manipulation and proc-essing of video information in ways that analogsystems camot. Significant improvements in thequality of the picture presented (ch. 4) and thepotential for interactivity will be the result. Digitalmicroprocessors are already common in televisiontuners, and IDTVS are now on the market. In thelonger term, these advantages make the increasinguse of digital electronics in TVs inevitable.
HDTV currently exerts greater demands thancomputer or telecommunications equipment forfull-color, high-resolution, W-motion, real-timevideo. As a result, HDTV is pushing the state-of-the-art in a variety of display, display processor, storage,and certain related technologies. The mutual conver-
gence of Advanced TVs, computers, and telecom-munications towards digital video systems and thestrong technological push provided by consumervideo may result in significant linkages betweenthese sectors.
Technological linkages between components andmarket segments are undeniably important and noteasily quantiiled. Such linkages, however, can beoverstated. For example, report “B” cited abovestated that if the U.S. share of the HDTV maxket is10 percent or less (weak), then the U.S. share of theworld PC market would decline horn today’s 70percent to just 35 percent in 2010; and the U.S. shareof the world semiconductor market would declinefrom 41 to 20 percent in the same time period. Thereport is done so haphazardly, however, that it showsthese drops in U.S. marketshare beginning in 1990.Thus, in their scenario the U.S. share of the world PCmarket even next year oscillates between $34 billionand $17 billion depending on whether or not we haveless than 10 percent or greater than 50 percent of anHDTV market that is all but nonexistent-accordingto their estimate, worth just $4 million in the UnitedStates and $92 million worldwide.
As discussed in the previous chapter, there arenumerous and important technological linkagesbetween certain aspects of the technologies underly-ing HDTV, computers, and telecommunicationsequipment. Some of these linkages are strong; someare weak. In some areas, HDTV will drive thetechnology; in others computers or telecom equip-ment will drive the technology. For example,automatic test equipment will be primarily driven bythe demands of the computer and telecommunica-tion industries due to the greater complexity, variety,and flexibility of the circuits. Some HDTV propo-nents have sacri.iiced credibility and trivialized theimportance of these linkages by arbitrarily assuming
31~s i=ores tie i.lldus~ and military segments which are much s~kr.
Chapter tidvanced TV Markets and Market Uncertainties ● 89
an instant, across-the-board reduction in marketshare.32 Linkages are usually much more subtle, aremore often identtiled only in hindsight, and theirimpact on related industries and markets is oftenslow to develop.
The consumer demand for video entertainmentand for consumer electronics equipment will alwaysbe large regardless of the spectilc course taken byAm markets. The Japanese semiconductor industrytoday is strongly supported by sales of consumerelectronics. Roughly one-fourth of total Japanesesemiconductor output is currently used in consumerelectronics, and TVs and VCRs are a major portionof this.33
Perhaps more significantly, large volume massproduction, typical of consumer electronics, drivesthe state-of-the-art in manufacturing technology forsystems. These manufacturing technologies, as wellas the management practices that go with them, willbe important for producers of computer and telecom-munications equipment.
In the past, U.S. firms could ignore the consumerelectronics market at relatively less cost becauseanalog technologies were used. With the shift ofconsumer equipment to digital electronics, the
linkages to the computer and telecommunicationindustries are becoming quite important and U.S.fiis can no longer ignore this market with impu-nity.
The rate at which fiber-optics is extended to thehome and the details of how it is comected and madeuse of is also uncertain. Fiber has already begunreplacing the backbone of cable TV and telephonesystems. In the midterm, mixed fiber, coaxial cableand copper pair systems (ch. 3) of the telephone andcable TV companies could provide a limited rangeof interactive video services.
In the longer term, most agree that the improvedperformance and ultimately lower costs of fiber (andassociated electronic equipment) guarantee that itwill eventually be used in all new construction andperhaps to replace copper pairs in existing buildingswhen maintenance costs become excessive. Ulti-mately, fiber will likely penetrate most householdsand could then provide a wide range of high-qualityinformation services, however long that may take.The policy question is not whether or not to allow orpromote this, but rather how to provide a fimneworkfor it to happen most flexibly with the least overalldisruption to society.
szhp~es of this IOSS of credibility can be seen in: Congressional Budget office, “The Scope of the High-Defiition Television Market and ItsImplications for Competitiveness,” July 1989; David Wessel, “Mossbacher’s Initiative on HDTV Is Getting Scuttled, Sources Say,” Wa2Z StreetJournal, Aug. 2, 1989, p. B2; and Peter Passell, “The Uneasy Case for Subsidy of High-Technology Efforts,” New York Times, Aug. 11, 1989, p. Al.
SSF1-, op. cit., fwmote 24; Japan’s v~ pr~uction a~olmts for 12 percent of toM Jap~ese chip production: “~’s H@-S@kW, Hi@-TechBattle,” Fortune, Oct. 24, 1988. .
Appendix A
The Decline of the U.S. TV Industry: Manufacturing
U.S. firms led the world in TV technology andproduction until the early 1970s. Then, foreign-ownedfirms gradually took control of the U.S. market. Inparticular, superior design and manufacturing techniqueswere implemented by Japanese manufacturers, in afinancial and business environment that the JapaneseGovernment took pains to make favorable to technologydevelopment and capital investment. Their governmentalso assisted in developing overseas markets through avariety of export promotion incentives and protected theemerging Japanese companies from competition. Thedecline of the U.S. television industry was also hastenedby dumping on the part of Japanese (and, later, otherforeign) firms, and by the sluggish and inadequateresponse of the U.S. Government to that unfair tradepractice. Finally, the diversification of Japanese produc-ers from small, portable televisions into larger, highervalue segments probably was speeded by Orderly Mar-keting Agreements, which limited the number of unitsthat could be sold here.
Solid-state electronics were almost entirely invented inthe United States, yet the Japanese pursued solid-state TVdesigns more vigorously. To reduce energy consumptionby TVs, manufacturers requested that MITI sponsor amulti-company project to study the replacement ofvacuum tubes with transistors. MITI assigned particularresponsibilities to specific companies, and then circulatedthe results among all participants. The new designs,introduced in 1969 and 1970, used less than half thepower of the older models.
MITI followed this with another multi-company re-search program on the use of integrated circuits (IC) incolor TV. The change to ICS allowed dramatic reductionsin the number of electrical components used in a TV.From an average of 1,200 components per color TV in1971, Japanese firms reduced the count to just 480 by1975. In the same time period, U.S. firms only managedto reduce the component count from an average of 1,150to 880 (figure A-l).
This enabled all the components to be squeezed onto asingIe printed circuit board, which reduced the number ofcontact points and made the TV both easier to assembleand more reliable. In contrast, American firms continued
to use multiple-board designs for their ease of service.Thus, while Japanese fiis designed faults out, Americanfirms accepted faults as inevitable, designed for them, andin the process introduced them.
This difference in design philosophy could be seen atevery stage of the manufacturing process. Japanesemanufacturers worked to ensure ‘‘zero” defects by:performing elaborate pre-production testing of new de-signs; coordinating closely with suppliers to eliminatedefects in incoming parts; and adopting automatedassembly. By 1978,65 to 80 percent of components wereinserted automatically in Japan compared to 40 percent inthe United States. This insistence on perfection evenextended to the boxes that the finished TVs were shippedin. When Sanyo took over the Arkansas facilities ofWarwick (U.S.), over a year was spent working withsuppliers until the cardboard was flawless and thelettering perfect.
In contrast, American manufacturers allowed a certainpercentage of incoming components to be defective, andrelied on testing during and after the production processto catch the problems. As a result, typically less than 1 in100 TVs had to be reworked during production due to
Figure A-l—Average Number of ElectricalComponents Per TV for U.S.-and Japanese-Made
Color TVsElectrical component per standard set
1600
1400- &
1000-
800
600 1
400
L , , , , , ,
K
200
0
64 6 6 6 8 7 0 7 2 7 6 7 8 8 0
Calendar yea~4
— Us. h Japan
SOURCE: Based on data in lra C. Magaziner and Robert B. Reich, MindingAmerica’s Bushess (New Yorkr NY: Random House, 1983).
l~mes for MS ~tion iIKIu&: Harvad Business School Cases, “The U.S. Television Set Marke~ Wwti to 1970, “ “The U.S. Television Set Market, 1970-1979,”“SanyoMauufaeturing Corporation-Forest City, Arkansas,““TheTelevisimSet Industry in 1979: Jaw Europe and Newly Industrializing Countries.” Ira C Magazinerand Robefi B. Reich “Minding America’s Business” (New York NY: Random House, 1983). J.M. J~ ‘‘Japanese and Western QuaLity-A Contrast,” @ahiy Progress,December 1978. James Howard Wooster, “Industrial Policy and International Competitiveness: A Case Study of U.S. Japanese Competition in the Television ReceiverManufacturing Industry,” Ph.D. l%esis, University of Massachusetts, 1986. Developing World Industry and Technology, Inc., “Soumes of Japan’s InternationalCompetitiveness in the Consumer Electronics Industry: An Examir@‘on of Selected Issues,” contractor report for the Office of Technology Assessment, 1980. James E.Millste@ “DecliueinanExpauding Industry: Japanese Competition in Color Televisio~” inAmericanIndustry in Internatiomzl Competition, John Zysman and bura~son(eds.) (Ithaca, NY: Cornell University Press, 1983).
–91–
92 ● The Big picture: HDTV and High-Resolution Systems
defects in Japan; many U.S. factories had to rework 50percent or more. Strict control of the quality of incomingparts is also important if automation is to succeed. Forexample, Philco’s automated circuit board assembly plantin Brazil ran for 3 months before they discovered thatevery IC from one of their suppliers had been defective.The cost of replacing the circuit board of every color TVproduced during this period was an important factor intheir owner’s (Ford Motor Co.) decision to sell Philco toGTE-Sylvania and Zenith in 1973-74.
Reduced re-working of TVs during production, auto-mation, and other factors reduced the assembly time forcolor TVs to 0.8 hours in Japan versus 2.6 hours in theUnited States in 1979 (figure A-2). Their TVs were alsomore reliable. Service calls for U.S.-made color TVs werefive times more frequent than for Japanese-made TVs in1974, dropping to two times the frequency by 1979 asU.S.-made sets improved.
These changes in failure rates had other impacts aswell. Because of early reliability problems, many U.S.firms had developed dealer networks to sell and serviceTVs in the 1950s, as had Japanese electronics firms inJapan. The increasing reliability of TVs and the ability tomarket them through mass merchandisers in the UnitedStates, however, freed Japanese firms from supportingsuch dealers in the United States, and thus converted thesedealer networks from a potential barrier to market entryinto a liability for U.S. firms.
Meanwhile, the Japanese producers were protectedfrom foreign competition at home. The governmentmaintained strict controls on foreign direct investment,restricted import and currency exchange licenses, and setimport quotas. Even without them, Japan’s distributionsystems and traditional business practices posed a formi-dable practical barrier to American exporters.
Some U.S. firms were competitive in manufacturing,however, and survived till quite recently (GE, RCA) or to
Figure A-2—Man-hours To Assemble Color TVs in—U.S. and Japanese Factories
Aa8embly man-hours per atandard set
‘;[5;,<\
4
2
0L k-—–
64 66 6 8 70 72 74 76 78 8 0Calendar year
— U s . + J a p a n
SOURCE: Based on data in in Ira C. Magaziner and Robert B. Reich,Minding America’s Business (New York, NY: Random House,19s3).
the present (Zenith) even in the face of dumping by someforeign firms. Had this dumping and other trade violationsnot occurred, other U.S. firms might have had the time tolearn and apply superior quality control and productiontechniques, and have had the profit margins needed toinvest in R&D and automation to be fully competitive. Inthis context, it is important to note that U.S. manufactur-ing practices were superior to those of the Japanese interms of component counts and assembly man-hours(figures A-1 and A-2) until the early 1970s.
Foreign-owned firms now dominate the U.S. TVindustry and much of the skill-intensive design andproduction of electronic components and subassembliesis done abroad, in many cases leaving primarily screw-driver assembly operations to be done in the UnitedStates.
Appendix B
The Decline of the U.S. TV Industry: Trade
Foreign t.mde pmctices were an important factor in thedecline of U.S. television manufacturers. Domestic pro-tection and coordinated pricing policies permitted Japa-nese firms to sell televisions in the American market atprices substantially below what comparable sets weresold for in Japan. The U.S. response to this dumping wassluggish and ineffective-in part because a statutorychange that enabled a more effective response was notmade until 197%xposing American firms to competi-tion that drove down profit margins and increased thedifficulty of making the needed investments to improvetheir manufacturing. The Japanese Government protecteddomestic producers flom imports as welI as foreign directinvestment, and the nation’s distribution system acted asa barrier to any firm that could surmount the govern-ment’s restrictions.
Japanese TV manufacturers have high fixed costs, inpart due to their customs of providing permanent employ-ment and of maintaining exclusive distribution outlets.High-volume production was needed to cover these fixedcosts but also provided significant economies of scale.Government fiscal policies and other factors encouragedheavy investment in plant capacity by Japanese manufac-turers-in excess of domestic needs-to achieve thesevolumes. The large output was sold at a high profit in theprotected domestic market and at low- or no-profitabroad: Japanese producers maintained domestic retailprices at least 50 to 60 percent higher than for comparablesets sold in the United States. 1 Imports would have brokenthis arrangement but were blocked.
The Japanese Government protected the profitability ofdomestic sales through tariffs, quotas, import and foreignexchange licensing, and restrictions on foreign directinvestment. Tariffs on color TV imports were 30 percentin Japan until 1968 compared to 10 to 7.5 percent in theUnited States over the same period. Commodity taxeswere reduced to less than 10 percent on sets entirely ofdomestic origin, but were maintained at 30 percent onthose with larger imported picture tubes. Import certifica-tion took much longer and was more costly and stringentthan the U.S. equivalent.
The Japanese industry blocked imports by denyingdistribution through their extensive network of franchiseddealers, who carried only one manufacturer’s products,
excluding all others. Sales through large retailers alsoproved difficult. In 1973, for example, Zenith attemptedto export to Japan to take advantage of the exceptionallyhigh prices for TVs there, but MITT reportedly pressuredZenith’s trading partners and retail chains to limitdistribution efforts.
As a result of these and other factors, sales of color TVimports in Japan totaled only 16,000 in 1974, 11,000 in1975, and 452 in 1976-out of total color TV sales ofalmost 5 million.
At the same time, Japanese sold TVs at rock bottomprices in the United States. To ensure that the impact ofthis was on U.S. rather than Japanese producers, Japanesefirms (including Hitachi, Matsushita, Mitsubishi, Sanyo,Sharp, and Toshiba) a.IIocated U.S. retailers amongthemselves according to the so-called “five company”rule to eliminate intra-Japanese competition.
U.S. firms challenged the unfairness of Japanese tradepractices, both in courts and through administrativeprocesses. Action was taken on at least five separatefronts.
The first and longest proceeding began in 1968 whenU.S. Customs received complaints about dumping viola-tions. In 1971 the Department of Treasury found Japaneseproducers guilty of dumping,2 but virtually no duties werecollected or other actions taken until Congress overhauledthe antidumping duty 1aw in 1979.3 At that point, theSecretary of Commerce negotiated a settlement of ap-proximately $77 million for antidumping duties and otherpenalties. Zenith, having estimated its own damages asmuch larger than this, unsuccessfully appealed thesettlement. The case was finally closed in 1987 when thegovernment unsuccessfully tried to force Zenith to forfeitthe $250,000 bond it had been required to post inchallenging this settlement.4
This and subsequent dumping findings against Korea,Taiwan, and Japan (most recently for 1986-87)5 resultedin the imposition of duties on TVs imported from thesecountries. Foreign efforts to rescind the duties have failed,but duties have reportedly sometimes been avoided byshipping TVs or components to the United States throughthird countries. For example, by transshipping through
IKozo Yamamura and Jan Vandenberg, “Japan’s Rapid-Growth Policy on Trial: The Television Case, ” Luw and Trade Issues of the Japanese Economy, Gary R.Saxonhouse and hzo Yamamura (eds.) (We, WA: University of Washington Pxess, 1986).
2T.D. 71-76,5 cw. B. and Dec. 151,36 Fed. Reg. 4597 (1971).
3Ke~Kennedy,finithR&io Corp. v. UnitedStates: TheNadir@the U.S. TradeReliefProcess, Nofi CiUOlirMJbof bh~d Law md CO~ xial Re@tion(1988), VOI. 13, p. 226.
4&ni~h R@~ COT. V. Unit&states, 823 F.2nd 518 @d. Cir. 1987)
s“Televisim R~iVers, M~~~ and Color, Fmm Jaw” 54 FR 13917, 1989.
-93–
94 ● The Big picture: HDWand High-Resolution Systems
Mexico, Matsushita was reportedly able to cut its tariffbill horn 15 to 5 percent on color picture tubes.6
Second, the National Union Electric Corp. (U. S.) filedsuit in 1970 and Zenith filed suit in 1974 against eightJapanese firms and their subsidiaries for violations ofantitrust and antidumping laws. Most of the evidence inthe case was ruled inadmissible by the District Court in1981, including, for example, thousands of pages ofdocuments seized by the Japan Fair Trade Commission inraids on corporate offices.7
The District Court’s decisions were largely reversed bythe Third Circuit Court of Appeals, which concluded thatthere was direct evidence of concerted action among theJapanese, including price fixing, use of the “five-company’ rule, and false-invoice-and-kickbacks to avoidU.S. Customs regulations and antidumping penalties. TheCourt recommended that the case be sent to a jury.8
The Third Circuit Court of Appeals findings werenarrowly reversed, 5+ by the Supreme Court in 1986.The Supreme Court ruled that the direct evidence ofconcerted action among the Japanese found by the ThirdCircuit Court was irrelevant, because the Japanese couldnot have a motive to engage in predatory pricing in theUnited States—it would require the Japanese to sustainyears of “substantial losses in order to recover uncertaingains. “9 Pricing behavior in the Japanese market was alsodeemed irrelevant to the antitrust charges because ‘Ameri-can antitrust laws do not regulate the competitiveconditions of other nations’ economies. “1°
The legal arguments notwithstanding, the Japanesefirms’ actions at issue in the case caused significantdamage to U.S. firms. Both U.S. and Japanese firms hadto cover their fixed costs. The Japanese Governmentpermitted, if not encouraged, the creation of a protected
domestic market in which the Japanese firms wereallowed to recover all of their fixed costs, Iiee of anysignificant foreign competition. Japanese firms couldthen charge much lower prices in foreign markets,including the United States, while U.S. firms were forcedto charge prices that covered their average total costs.
Third, Zenith petitioned the Treasury Department in1970, requesting the imposition of a countervailing dutyto offset the effective rebate of the commodity taxgenerally due on TVs which Japanese producers werereceiving from their government for exports. In 1976, theTreasury Department found that the rebate was neither a“bounty” nor a “grant,” either of which would havetriggered the imposition of a duty. The decision wasupheld unanimously by the U.S. Supreme Court in1978.11
Fourth, in 1972 the U.S. Department of Justiceundertook investigations of collusion and of false-invoice-and-kickback schemes run by Japanese firms to circum-vent U.S. dumping tariffs and penalties.12 In 1977, Justiceconcluded that there was no evidence of collusion, but itbrought charges against a number of fiis for thesekickback schemes. At least one U.S. retailer pled guilty,13and at least one case was still unresolved in September1989. 14
Fifth, in 1976 a GTE request to the U.S. InternationalTrade Commission (ITC) to investigate unfair acts by fiveJapanese firms was dismissed, but a request by COM-PA(X15 led to a 1977 ITC holding that there was injurydue to increased imports. While the ITC thereforerecommended higher tariffs on color TVs, the Adminis-tration responded instead by negotiating a voluntaryOrderly Marketing Agreement.
6Co~~ ~ pmwme AIUSriC~ Color Televisio~ U. S. Co~ss, House subcotittee on Tekcmnmwn “cations and Finance, Ckuruittee on lhergy and Cormmrce,“Recommended Oovemment Policies to promote the Evolution of a U.S. High Definition Television Industry,” Cornmittoe Print 101E, March 1989.
7~nith R~io COP. v. Matsushita Elec. Id. CO.,” 494 F. Supp. 1190 (1980), 505 F. Supp. 1125 (1980), and 513 F. Supp. 1100 (1981)s’c~ Re Japse Hectroldc kdlCtS,” 723 F.2nd 238 (1983). See also “Japanese Electronic Products Antitrust Litigation,” 807 F.2nd 44 (1986) for the discussion of
the case following the Supreme Court decision.9~ leg~ Scholw ~~d ~ ~ ~~ @ysis of @ov tit ~w ~ l~ge pm -d ~ fie -St “c~serv~ve S&OOl of ~~ght ti the debate Over Pdation,”
that of Bork-McOee-Easterbmok. See: Randolph Sherman, “TheMatsushitaC ase: Tightened Concepts of Conspiracy and Predation?” Curdozo LzwReview, vol. 8, p. 1121,1987.
~OMatsushita E[ecwic Industrial CO. v. Zenith Radio COW., 475 U.S. 574 (1986).
ll~nith V. United States, 437 U.S. 443 (1978).
12’iWm&& ~ckbxks k Living Color,” Time, June 13, 1977, p. 63.
13David Staelin et al., “The Decline of U.S. Consumer Electronics Manufacturing,” Massachusetts Institute of Technology Cmnnis sion on Industrial Productivity,December 1988.
l~s cW~ ~~ck ~d fofimmP@ apP~s ~d ~-ds. 647 F.&d 902 (9~ Cir. 1981); m~&d to 5]8 F. Supp 179 (C!. J). Cal 1981); reviewed 719 F. 2nd1386 (9th Cir. 1983); cext. dm 104 S. Ct. 1441 (1984); 579 F. Supp. 1055 (C.D. Cal 1984); reviewed 785 F.2d 777 (9th Cir. 1986); cat den. 479 U.S. 988 (1986); 677 F.Supp. 1042 (CD. Cal 1988); reviewed 866 F. 2nd 1128 (9th Cir 1989).
Appendix C
The Decline of the U.S. DRAM Industry: Manufacturing
U.S. firms led the world in DRAM technology until theearly 1980s. Japanese firms then gradually took control ofthe world market, in part because many U.S. producerscould not match Japanese efforts at critical points in thetechnology’s lifecycle. Heavy investment of money andmanpower and close attention to highquality manufac-turing were important factors in the Japanese success. Inaddition, Japanese efforts have been abetted by violationsof trade law. As a result, Japanese firms now control 70percent of the total world DIL4M market and 85 percentof the advanced 1 Mbit DRAM market.
Two engineers designed Intel’s (U. S.) pioneering IK(1,000 bits or binary memory cells) DRAM in 1970-71and just three engineers designed Intel’s 16K DRAM. Incontrast, one of today’s major Japanese DRAM producersreportedly assigned 50 select engineers to design their IKDRAM and 100 to design their 16K DRAM. This allowedgreater specialization, more careful attention to issues ofmanufacturability, and more rapid development of thedesigns.
Japanese firms invested heavily in pIant and equipmentin the mid-1970s. In contrast, U.S. producers cut invest-ments due to the 1974-75 recession and were then unableto meet the demand when U.S. semiconductor marketsboomed in 1979. Japanese producers stepped in—offering 16K DRAMs as licensed second sources for theindustry-standard Mostek (U. S.) design—and, by the endof 1979, had captured40 percent of the world 16K DRAMmarket.
Manufacturing quality began to appear as an issue inthe late 1970s. Japanese 16K DRAMs, for example, hadmuch lower failure rates than those of U.S. firms (tableC-l)-even though nearly all began with the Mostekdesign. It took several years for U.S. firms to reducefailure rates to comparable or lower levels.
Table C-l—U.S.-Japan 16K DRAM Failure Rates(parts per miiiion)
1978 1979 1980 1981 1982
Japanese vendors . . . . . . . . . 0.24 0.20 0.16 0.17 0.05U.S. vendors . . . . . . . . . . . . . . 1.00 1.32 0.78 0.18 0.02
SOURCE: Hewlett Packard, WNiam F. Finan, and Annette M. Lamond,“Sustahing U.S. Competitiveness in MicroElectronics: TheChallenge to U.S. Policy,” in U.S. Competitiverress in the 144ddEdortorny, Bruce R. Scott and George C. Imdge (eds.) (Boston,MA: Harvard Business School Press, 19S5).
U.S. firms lost even more of the DRAM market in thenext generation due, in part, to relatively less competitivemanufacturing. Chiprnakers normally design chips assmall as possible to reduce the likelihood that any onechip is contaminated by a stray microscopic dust particleand to increase the number of chips produced per waferprocessed. Determined to leapfrog the Japanese in qualityand cost, most U.S. producers designed much smaller andmore sophisticated 64K chips than the Japanese, but wereconsequently slower than the Japanese in completing thedesigns and in solving related manufacturing problems.
In contrast, the Japanese successfully produced 64KDRAMs by slightly modifying and scaling up &eir 16KDRAM designs. The resulting chip was nearly 50 percentlarger than the leading American designs, but theyachieved good yields by using higher purity chemicals, bygreater capital investment in cleanrooms and automation,and by superior quality-control techniques.
The simple design allowed the Japanese firms to get tothe market first. High yields also lowered the overall costper chip and gave them a greater production output perunit of capital investment and per labor hour than U.S.firms. By the end of 1981, the Japanese held 70 percentof the world 64K DRAM market. U.S. firms cut theJapanese share to 55 percent by rnid-1983 after enteringthe market in volume, but most U.S. firms subsequentlyabandoned the market due to Japanese dumping in themid-1980s and/or due to problems they encountered inmanufacturing subsequent generations of DRAMs com-petitively.
Issues of manufacturing process arose again as firmsmade the transition from the 256K to the IM (1,000,000bits) DRAM. Table C-2 compares the production yieldsand costs for a lower-yield U.S. manufacturer-a majorU.S. firm that subsequently dropped out of the DIWMbusiness-with that of Toshiba, the world leader in IMDRAM production.
The U.S. firm’s design and manufacturing process hadseveral serious shortcomings that allowed no rnarg. forerror. For example, in etching the silicon wafer to createthe circuit elements, the U.S. firm’s process formed sharpvertical walls. In previous generations of DRAMs thiswould not have been a problem. But with the dimensionsof the IM DRAM circuitry shrinking to just 1/100 thediameter of a human hair, sharp vertical walls prevented
ltiesforthis section i.nclu&: PeterD. N- *mi@cLFrson~m~ .cations, May 11, June 22, Aug. 4, andoct. 10, 1989; Brian Sante, “lK-Bit DRAM, 1970, ”IEEE Spectrum, vol. 25, No. 11, 1988; Wfiam F. F% Jeffrey Frey, “Study of the Management of MieroEkctmnics-Related Research and Development in J~”cuntractorqortpre pared forthe OfficeofTechnology AssessnnmL November 1988; CompetitiveEdge: TheSemiconductorInAstry in the US. andJapan, Dauiel I. Oldmoto,Takuo Sugano, and Franklin B. Weinstein (eds.) (Stanford, CA: Stanford University Press, 1984); William F. Finan and Annette M. LsMon& “Sustain@ U.S.Competitiveness in Microelectronics: The Challenge to U.S. Policy,“ in U.S. Competitiveness in the World Economy, Bruce R. Scott and George C. Lodge (eds.) (BOXMA: Harvard Business Press, 1985).
-95–
96 ● The Big picture: HDTVand High-Resolution Systems
Table C-2—U.S.-Japan 1 M DRAM ManufacturingCost Comparison
Lower yieldU.S. manufadurer
Operation 3Q 1986Toshiba3Q 1986
Start wafer costa . . . . . . . . . . .Prooessed wafer cost . . . . . . .Chip size (square mm) . . . . . .Total chips possible (assuming
a 125 mm wafer) . . . . . . . . .Wafer probe yield . . . . . . . . . .Number of good chips . . . . . .Packaging cost . . . . . . . . . . . .Assembly yield . . . . . . . . . . . .final test cost . . . . . . . . . . . . .Final test yield . . . . . . . . . . . . .
Total manufaoturina oost . .
(Bulk) $25.00$300.00
54
15125Y0
$::592Y0$0.2085’XO
$11.83
20568Y0139$0.2592Y0$0.2085?40$3.31
NOTE: These are representative values to indicate relative manufacturingcosts for these two firms at a particular time. These firms are atdifferent points on the learning curve for IM DRAMs in 1986, butprocess design flaws probably would have prevented much higheryields for the U.S. firm.
~he starting wafers, Epi and Bulk, refer to different types of wafers.
SOURCE: Peter D. Nunan, Sematech, personal communications, May 11,June 23, Aug. 4, and Oct. 10, 1989.
subsequently deposited material from being effectivelyetched out of the comers, causing the circuitry to short-out(figure C-l).
The Toshiba engineers recognized this pitfall anddeveloped a new technique which formed sloped ratherthan sharp vertical walls (figure C-l). The resultingprocess was highly robust. When transferred to Siemensin Germany and to Motorola in the United States, yieldswere high even with the very first wafers processed andeven with relatively less experienced line workers.
Technical management and quality philosophy provedto be key problems for the U.S. firm. Its design engineersdeveloped their DRAM process and prototypes in thelaboratory, and then ‘threw the design over the fence’ tothe manufacturing engineers. The design engineers recog-nized the difficulties of producing IM DRAMs with theirdesign: they attempted to compensate by specifying ahigh-quality starting wafer, by keeping the chip sizerelatively large, and by including a very large number ofredundant memory cells on the chip as backup (table C-2).They were relying on inspection and correction afterproduction to provide usable DRAMs rather than design-ing quality in.
The manufacturing engineers were unable to get (waferprobe) yields up to competitive levels under factoryconditions. They protested that the process had no marginfor error and was not readily manufacturable, but didn’thave the resources or knowledge to do proper analysis andimplement improvements. The designers insisted thatthey had developed a robust and manufacturable processand stayed away from correcting the problem. In contrast,the engineers and scientists who developed the Toshiba
Figure C-l-Cross-section of IM DRAM Storage Cell
Metal or oxide insulator
/on vertical Si(fe ~a,,
//////////////////////~Residual polysilicon
-J/.— — — — — — — — — — -. — — — -——
Transistor
Lower-yield U S. manufacturer design
Metal or oxide insulator
/ “ ’ ”
on sloped sidewall.
~ ———-—Toshiba design
A. Cross-section of lower yield U.S. manufacturer’s designshowing the sharp mrners from which it was dificult to etchresidual polysilicon.B. Cross-section of Toshiba design shovhg the sloped sidewall.SOURCE: Peter D. Nunan, Sematech, personal communication, Oct. 10.
1989.
process were also responsible for improving the yield onthe factory floor. Even if the Japanese had not beendumping DRAMs in the United States in the mid-1980s,the IM DRAM design and manufacturing process of thislow-yield U.S. manufacturer might never have beencompetitive.
The loss of the DRAM market may be particularlydarnaging to the U.S. chip industry. DRAMs are knownas a technology driver because they push the limits incertain kinds of process technologies. Loss of DRAMproduction will likely cause U.S. fiis to lag the Japanesein developing certain kinds of manufacturing processesimportant in the production of many types of chips.
U.S. firms face formidable obstacles should theychoose to reenter the DRAM market. From $5 to $10million in the mid- to late-1970s, the cost of a singleminimum-efficient scale, state-of-the-art DRAM produc-tion facility has risen to roughly $200 million today, andis expected to approach $400 million for the next-generation 16M DRAMs. The human skills to design andproduce leading-edge DRAMs also take many years todevelop.
Siemens has committed $1.6 billion to develop oracquire IM and 4M DRAM technology and productionfacilities. Even armed with IBM’s DRAM designs,however, U.S. Memories failed to raise $1 billion to enter
Appendix C—The Decline of the U.S. DRAM Industry: Manufacturing ● 97
DRAM production. Similar investments needed forproducing state-of-the-art semiconductors generally areall but impossible for small- and medium-sized firms. Asa result, many American companies are forced to rely onJapanese and other foreign firms to produce their chipdesigns.
Some observers see the relatively low level of fundingfor Sematech—roughly $200 million per year-and thecorresponding decision to not pursue large-volume pro-duction of DRAMs as critical constraints. They argue thata high-volume operation is essential for testing yield and
reliability, and that issues of technical management andquality techniques-such as for the lower yield U.S.manufacturer described above-can otherwise be sweptunder the rug. As one Sematech engineer, frustrated bywhat he feels is an inadequate response to the Japanesechallenge, put it,
It’s as if the Soviets, having already taken the lead in thespace race, had announced in 1961 that they were going tosend a man to the moon, and the U.S. response was to focuson selected aspects of rocket science.
Appendix D
The Decline of the U.S. DRAM Industry: Tradel
The Japanese Government protected the Japanesesemiconductor industry when it was weak and, as itstrengthened, supported its move into international mar-kets. Conversely, American trade policy largely failed toprevent serious darnage to U.S. industry caused by tradeviolations.
Scientists and engineers in the United States inventedessentially all of modem solid-state electronics. By oneaccounting, of some 103 major product and processinnovations in the semiconductor industry between 1950and 1978, 90 were by American fiis. Despite thetechnological lead of U.S. firms, they were never able toconvert it into market share in Japan as they did in Europe.
U.S. firms found it all but impossible to establishsubsidiaries in Japan or joint ventures with a significantshare of the equity-unless the Japanese partner wasgiven access to new technology or other appreciablebenefits. Thus, whereas U.S. firms had established 46subsidiaries in Europe by 1974, including 18 manufact-uring operations, only Texas Instruments (’H) had amanufacturing operation in Japan.
TI succeeded where other U.S. firms failed because ofits strong U.S. patent on the integrated circuit (IC). TheJapanese Government refused to give TI permission toestablish a wholly owned subsidiary in the early 1960s; inturn, TI refused to license its IC patent in the United Statesto Japanese firms. This generally stopped exports ofJapanese ICS to the United States, but it did not stopJapanese firms tiom producing ICS for their domesticmarket as TI’s application for a Japanese patent wasrefused. (’II applied for a Japanese patent on its inventionof the IC in February 1960 but did not receive the patentuntil November 1989. The patent is estimated to be worth$500 million annually in royalties to T’I.)
Japanesefirmsmpidly gainedexpertise inlCfihication-rnaking TI’s entry into their market evermore difficult thelonger it waited to settle. In 1968, TI settled for a 50:50joint venture with Sony; licensed NEC, Hitachi, Mitsub-ishi, Toshiba, and Sony; and agreed to limit its share of theJapanese IC market to less than 10 percent. Firms withweaker patent positions did not succeed in establishingmanufacturing subsidiaries in Japan until much later.
Tariff and non-tariff barriers limited imports into Japanas well. Tariffs were roughly double those of the UnitedStates until the early 1980s. Imports of ICs with more than
200 elements were bamed until 1976, limiting the importmarket to the most simple types.
Unable to penetrate the Japanese market, most U.S.firms licensed their technology to Japanese firms as ameans of realizing some earnings. The Japanese Gover-nment required that foreign firms license all Japanese firmsso requesting at a single royalty rate. This prevented thecompetitive bidding up of the license fees; and the broadlicensing prevented any one firm from capturing monop-oly revenues. Competition among firms was thus effec-tively shifted fkom innovation downstream into manuhctur-ing.
The battle for control of the DRAM market began in thelate 1970s with the strong Japanese push in 16K DRAMs.The 1979 boom in semiconductor demand created acapacity shortage among both Japanese and @encanproducers. Japanese firms responded by increasing theirexport of chips to capture and hold market share abroadwhile importing the same chips to ffl their domestic needsuntil they could expand their production capacity. U.S.firms rushed to fill Japanese orders in the hope that thiswas a market opening, resulting in the highest levels ofsemiconductor imports into Japan at anytime between theearly- 1970s and 1989 (figure D-l). As additional capacitycame on line, however, the vertically integrated Japaneseproducers cut back imports of U.S.-made DRAMs whilehanging onto their market share gains abroad.
Japanese firms pushed DRAM prices down sharply inthe early 1980s (figure D-2). Repeated allegations ofdumping were made, but no formal action was takenbecause it was difllcult to distinguish the effect on priceof dumping, if any, from that of the high value of thedollar. Over half of the U.S. DRAM producers droppedout of the market during this period.
The remaining American firms followed the Japaneselead in heavily investing in capital equipment in 1983-84.When recession hit the semiconductor industry in 1985-86, the large overcapacity and other factors drove downDRAM prices. Japanese firms lost an estimated $3 to $5billion during 1985-86, while American fms lost anestimated $2 billion.
Severe price competition also occurred with othercommodity chips such as EPROMS. In 1985, Hitachi tolddistributors to quote 10 percent lower prices—irrespective of costs-than competing American firms
l-es fi~s s=tim iIKJU& hose cited~ ~. C ~d: Johu E. Til~ “Iu@tiod l)ifhi~ of Tecbology,” Brootigs btihti~ 1971; Es J. ~ “%“Cress-Investment: A Second Front of Economic Rivalry,” CalifomiaManagement Review, vol. ~ No. 2, Wider 1987; The Departmmt of Cothe Us. Semimndlllx
nnnerce, “A Report onor Indmtry, ” 1979; “The united states Government Trade Policy Response To Japanese Competition in Sanbnductors: 19821987,” O’Ill
Background Rem September 1987; Clayton K Yeutter, “lhe Japanese Left Us With Little Choice,” New York Times, Apr. 5, 1987; Andrew Pollac~ “Chip Pact FallsShort of Ooals,’’iVew YorkTimcs, Aug. 2, 1988; WilliamF. F- Chris B. Amundscw “Modeling U. S.-Japau Competition in Semiconduc tors,” Journal qfPolicy Mcxieling8(3): 305-326 (1986); John Burgess, “Japan Oives U.S. Firm Circuit Patent,” Washington Post, Nov. 22, 1989;
–98–
Appendix>The Decline of the U.S. DR.AMIndustry: Trade ● 99
Figure D-l —U.S. Share of the Japanese Semiconductor Market, 1973-66
PercentLosing the Chips: The Semiconductor Industry
100% I \90%
t I80% t I70%60Y050%40%30%20%10%
Substantial Substantial Substantial
yen appreciation yen appreciation yen appreciation
# – \9%
/
o% I I I I I I I I I I I I I 173 74
ElFirst‘‘liberalizationof imports
75 76 77 78 79 80 81 82 83
\ P!!!ns I / /nReduction ofJapanesetariff to4.2%
\Third Im
84 85 86
/t
U .S, -Japansemiconductoraccord
Elimination ofJapanese duties
SOURCE: Clyde V. Prestowitz, Jr., Trading P/aces: How We A//owedJapan To Take the Lead(New York, NY: Basic Books, 1988). Used with permission.
until the sale was won, while guaranteeing a 25 percentdistribution profit. From January 1985, when the Japaneseentered the 256K EPROM market, to August 1985 pricesfell from $17 per chip to less than $4, while the estimatedJapanese production cost alone was $6.34.
In response, U.S. semiconductor firms filed anti-dumping cases against Japan in 1985 for below-cost salesof 64K DRAMs, and the Department of Commerce itselffiled an antidumping case against Japan for below-costsales of 256 DRAMs and EPROMs. The U.S. Intern-ational Trade Commission (ITC) found that dumping hadoccurred. For example, constructed prices indicated thatDRAMs were being sold at half their estimated produc-tion cost. A trade agreement was subsequently reached inSeptember 1986. When this failed to stop below costsales, the Reagan Administration imposed sanctions.
These remain in place in early 1990 on the issue of thelack of access to the Japanese market. Despite thisrelatively quick action, the only American firms in themerchant DRAM market today are TI, Micron, andMotorola.
Following the trade agreement, prices on the spotmarket rose sharply to as much as four to five timeslong-term contract prices. Further, prices charged to U.S.purchasers have typically been 30 percent higher thanthose for Japanese users. In sharp contrast, EPROMprices-where U.S. producers still have 40 percent of theworld market and 70 percent of the U.S. market-havebeen much more disciplined.
Some analysts believe that Japanese producers, whonow control the world DRAM market, have subsequentlyacted like a carte12: driving prices up to capture excess
%hades Fergu~ “DRAMs, Component Supplies, and the World Electronics Iadustry: Aa international Stmtegic Analysis,” VLSI Memo 89-554, Ausust 1989,Massachusetts Institute of Technology, Cambridge, M& Kenneth Fl~ “Policy and Politics in the International Semiconductor Industry,” paper presented at SEMI ISS
minar, Newport Beac4 CA, Jan. 16, 1989.se
100 . The Big Picture: HDTV and High-Resolution Systems
Figure D-2—Average CostMillicent8 per bits
120 1 1
Per Bit for DRAMs, 1976-89Millicente per bit
40
3 0 -
2 0 -
10-
0 1 I I 1 1 1 1 $ 1 ! 1 1 r
76 77 76 79 80 81 82 83 84 85 86 87 88 89 90 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90Calendar year Calendar year
- E x p e c t e d p r i c e ~ Actual price + E x p e c t e d p r i c e + Actual price
Showing the large price erosion that took place in 1981 and 1985.SOURCE: Based on data in: “Mid-Term 1988,” Integrated Circuit Engineering Corp., S@tedale, AZ, 1988.
profits in 1988 and 1989, and driving prices down, as in ary 1990, in part due to this sharp decline in DRAMfall 1989 to perhaps wam any would-be entrants into the prices; the same day, many of Japan’s largest chipDRAM market with the specter of enormous financial producers announced DRAM production cutbacks to turnlosses. The effort to launch U.S. Memories-a DRAM the price declines around.3production consortium-was finally abandoned in Janu-
3David E. Sanger, “Contrasts on chips, “ New York Times, Jan. 18, 1990.
Appendix E
The Development of the Japanese Computer Industryl
Japanese researchers at the University of Tokyo, theElectrotechnical Laboratory (run by MITI after 1952),NEC, Fujitsu, N’IT and elsewhere resumed pre-war workon computing machines in the late 1940s and early 1950s.The computers they developed were generally small,low-cost, and technically well-behind the better financedefforts in the west. In 1955, a MITI-sponsored committeerecommended that the computer sector be given morefinancial support, be protected by limiting imports, and beassisted in the acquisition of foreign technology. The firsttest came the following year.
In 1956, IBM requested MITI’s permission to create awholly owned manufacturing subsidiary (it already had asales subsidiary) in Japan with the right to return royaltypayments and profits to its parent company. Permissionwas denied. A settlement was not reached until 1960,when IBM was allowed to establish its desired subsidiaryand to repatriate 10 percent royalties back to its parent inreturn for licensing its patents to all interested Japanesecompanies for a 5-year period at a single reduced rate--5percent on computer systems, and 1 percent on parts,among other concessions. MITI negotiated these licens-ing rights on behalf of the individual companies toprevent competitive bidding-up of the royalties and toprevent the establishment of a domestic monopoly.
Even with the 1960 settlement, IBM-Japan operationswere closely controlled by the government on the groundsthat it might hurt domestic industry. IBM-Japan was notallowed to begin production until 1%3; its 1964 requestto produce the 360 series was delayed for a fullyear-until after Fujitsu and NEC had introduced theirown “family series”; its importation of critical partswhich could not be produced locally was slowed; and theentry of capital that it needed to build facilities wasrestricted.
Beginning during this same period, the JapaneseGovernment began an extraordinary series of initiatives toenable Japanese firms to become world class competitorsin computer technologies and markets.
First, the Japanese Government provided domesticfirms direct financial support. Subsidies and tax breakstotaled about $130 million and loans totaled more than$400 million during the 1960s. Together, this was nearlytwice what domestic firms themselves invested in R&D,
plant, and equipment for commercial computer develop-ment.
That funding was often not used at the firms’ discre-tion. Much of it went to spectilc investments thatgovernment, business, and university researchers agreedwould contribute most to technical progress and produc-tion efficiency. Support was also targetted towardsspecific firms to develop certain classes of computers(Fujitsu, NEC, and Hitachi) and particular pieces ofperipheral equipment (Oki, Mitsubishi, and Toshiba).This divided the market and improved the scale econo-mies for the firms in each segment. Firms chosen by thegovernment to lead the effort in specfilc segments variedover time on the basis of competitive proposals and pastperformance. For example, Hitachi was chosen to lead the1966 Super High-Performance Computer Project—intended to develop a domestic counter to IBM’s 360Series-on the basis of its design proposal.
The government-backed Japan Electronic ComputerCo. (JECC) was another important source of directsupport. Most of its directors were former MITI or JapanDevelopment Bank offlcia.ls; and it was financed withlow-interest loans either directly through the JapanDevelopment Bardq or through a MITI-organized privatefinancing cooperative with the loans guaranteed by theJDB. As of 1978, the JECC was the 20th largest firm inJapan in terms of capital; yet had no sales division, did notadvertise, had just 120 employees, and averaged annualprofits of less than 0.1 percent of rental assets.
The JECC (est. 1961) purchased computers at rela-tively high fixed values to prevent price competition andprovide producers reasonable profits; and then rentedthem to users at values designed to undercut IBM. Thisgave computer makers their cash up front, and shiftedmuch of the financial burden from the computer firms tothe JECC. While 15 companies had licensed IBM’s basicpatents, only the top seven firms were allowed to enterJECC in order to prevent excessive competition such as“redundant investment and cut-throat pricing. ” The topthree-Fujitsu, Hitachi, and NEC-were given pref-erential treatment. Following the establishment of theJECC, Japanese companies share of the domestic marketjumped from 18 percent in 1961, to 33 percent in 1962,
l~ci~ SO-S include: Marie Anchordoguy, ‘The State and the Market: Industrial Policy Towards Japan’s Computer Industry, ” draft, 1986 and Computers, Inc.(Cambridge, IvIA: Harvard University F%Ess, 1989); Jonah D. bxy and Richard J. Samuels, “Institutions and Innovation: Research Collaboration As Technology Stmtegyin Japan,” MITJSTP 89-02, Massachusetts Institute of Technology, 1989; Kenneth Fl~ Targeting the Computer: Government Support andhternational Competition,Creating the Computer: Government, Industry, and High Technology (Wbsbingto% DC: Brookings Institution, 1988); Robert Sobel, IBM vs. Japan: The Struggle for theFuture (BriarcliffManor, NY: Stein & Day Publishers, 1986); Charles H. Fergusou “Technological Development, Strategic Behavior, and Government Policy inroformatl “onTechnology Industries,” PhD. Thesis, Massachusetts Institute of Technology; ha C. Mag aziner and Thomas M. Hout, “Japanese Industrial Policy,” Report #585 (J-xmdmuEngland: Policy Studies Institute, 1980); “A Worldwide Strategy for the Computer Market,” Business Week, Dec. 14, 1981; John Mzukoff, “Fujitsu Will Pay $833.3 Millionto IBM To Settle Software Fight, “ New York Times, Nov. 30, 1988.
–lol–
102 ● The Big Picture: HDTV and High-Resolution Systems
to 52 percent in 1965 despite the technical inferiority oftheir computers.
The JECC only bought the specific machines that usersordered for rent and, further, required computer makers tobuy back at book value any computers that users wishedto trade in after the minimum 15 months. This forced theproducers to compete for customers through continuouslydeveloping better computers. At the same time, domesticcontent requirements were only slowly increased. Whenthese buybacks became excessive for computer compa-nies, however, the government accelerated depreciation,further lowered the interest rate charged JECC, andallowed these companies to put money for trade-ins intotax-free reserves. By 1972, for example, some 60 percentof the cost of a computer could be depreciated in the firstyear.
Firms began developing their own rental systems in theearly 1970s to circumvent (undercut) JECC’S price carteland thus gain market share. Hitachi, for example, putsome $180 million into its rental system in 1973 alone,using profits from its consumer electronics division.
Second, the government organized cooperative R&Dbeginning in 1962. Cooperative R&D reduced the finan-cial burden on individual firms, promoted the diffusion ofcritical technologies, and increased competition by pre-venting any one firm from gaining control of criticaltechnologies. Many projects, including the first-theFONTAC project—fell far short of their goals. With eachproject, however, more was learned about managing suchcooperative ventures, the R&D was done at a lower costthan if firms had each done it individually, and, withexperience, better computers were developed. Projectswhich maintained a strong competitive environmentbetween firms were generally more successful than thosewhich placed all bets on a single horse.
The Japan Software Co., for example, was establishedin 1966 as a joint venture between Hitachi, NEC, Fujitsuand the Industrial Bank of Japan. It was intended todevelop the software needed for the MITI-organi.zedeffort involving all of the Japanese computer firms tomatch the IBM 360 computer. It failed. The companypresumed it had an assured market and made little effortto build up outside customers. Software technology iscomplex and abstract and realistic goals were difficult toformulate. The company was left with little direction. Inaddition, software technology changed rapidly and be-came increasingly important in overall system cost,increasing the desire of firms to keep software develop-ment within their own company rather than contractingfor it outside. The presence of the Japan Software Co. alsodiscouraged other firms from entering the softwaremarket. When the project ended in 1972, its ordersdropped precipitously leading to bankruptcy and dissolu-tion in December 1972.
Third, the government allowed firms to establishagreements with foreign partners for technological coop-eration, while at the same time denying foreign firms(with the exception of IBM) direct entry into their market.Other firms had less market power and fewer patents totrade upon and were thus generally unable to get termseven as favorable as IBM’s-IBM was the only computerfirm to get a wholly owned subsidiary during the 1960s.Sperry Rand, for example, was able to enter the Japanesemarket only by accepting a minority interest in a jointventure with Oki Electric. Between 1961 and 1964,Hitachi, Mitsubishi, NEC, Oki, and Toshiba formedagreements with RCA, TRW, Honeywell, Sperry Rand,and GE respectively. This dependence on U.S. firmscaused considerable turmoil in the 1970s when firms suchas RCA and GE abandoned their computer businesses.
Fourth, the government increased protection for thedomestic computer industry. Tariffs on imported comput-ers were raised from 15 percent to 25 percent in June 1960and tariffs on computer peripherals were raised to 25percent when the government decided to enter that marketin the late 1960s. The tariffs on computers were loweredin 1964 when Japan entered GA~ and the OECD.Quotas also limited imports, and were not ended until theearly 1970s. As already noted, IBM’s production in Japanwas similarly limited. Foreign firms, IBM in particular,were also excluded from certain data-processing marketswhich developed in the late 1960s by changing variouslaws that had restricted NTI”s entry. This allowed NT’I’to begin a cooperative research project in 1968 to developa large, high performance computer for on-line dataprocessing and to subsequently provide these services anddominate this market. NTI’ has also been a major sourceof R&D funding as well as a major market for computerfimls.
Fifth, government control over computer imports gaveit strong leverage over firms applying for import licensesto instead buy a Japanese made-computer. These effortswere effective. Purchases of foreign computers (includingthose made in Japan) were reduced horn 93 percent in1958 to 43 percent in 1%9 despite the technologicalinferiority of Japanese-made machines.
This “Buy Japan” policy did cause inefficiency andhardship, particularly in the 1960s when production wasjust getting underway and the technological gap was thelargest. Firms objected to this pressure from MITI to buydomestic computers, usually unsuccessfully. The gover-nment allowed, however, the import of some foreigncomputers to prevent excessive damage to critical sectorsand to push firms to do better by showing them the levelof technology needed to compete in world markets.
Sixth, government procurement played an importantrole in Japan just as it had in the United States. In the
Appendix&The Development of the Japanese Computer Industry ● 103
1%0s, the Japanese Government purchased or rented 25percent of all domestic computers.
These efforts helped. The U.S. hardware advantage wasreduced fkom some 10 years in the mid-1960s to perhaps4 years by the early 1970s. The Japanese share of theirdomestic market increased to some 60 percent by 1970.The introduction of the IBM 370 in the early 1970s,reduced the Japanese share of their domestic market to 48percent in 1974. RCA, GE, and others left the market atthis time due to the heavy investment that would havebeen required to remain even somewhat competitive withIBM.
Japanese producers might have left the market as wellhad it not been for government protection and support.Indeed, IBM had enormous advantages in the scale of itsoperations. In the late 1960s, the top three Japanesecomputer firms each manufactured about 2 percent of thenumber of computers made by IBM for any given type. Atthe same time, their currency was revalued, they wereunder increasing pressure from the United States to opentheir markets, and oil prices were crippling their heavyindustries.
In response to IBM’s 370 Series, the governmentorganized the firms into three groups: Fujitsu and Hitachifocused on large computers; NEC and Toshiba on smalland midIevel machines; and Mitsubishi and Oki onspecialized scientific and industrial machines. From1970-75, more than $600 million in subsidies, includingtax breaks, and over $1 billion in low-interest loanshelped these firms make the investments needed tocompete with IBM. Indeed, these subsidies and loanstotaled nearly 1.7 times what the firms themselvesinvested in R&D plant and equipment.
Similarly, the computer firms would have had tomassively increase their debt in order to finance theircomputer sales directly rather than through the JECC.Fujitsu, for example, would have had to more than doubleits long-term loans during the 1960s, and then nearlytriple them again in the 1970s-pushing its debt-equityratio to 21—in order to provide this financing itself.
A major opportunity also arose when a former topdesigner for IBM spun off a startup firm in 1970 toproduce IBM compatible rnainhames. Unable to securesufficient funding, he turned to Fujitsu for help in 1972
and received $54 million between 1972-76 in exchangefor technical information. In 1974, Fujitsu announced itwould produce computers in Japan for this company,Amdahl, to market in the United States. Amdahl is now49 percent owned by Fujitsu and sells over $1 billion ofIBM-370 compatible mainframes annually.
The intensive internal effort and external technologyacquisitions helped Japanese manufacturers produce com-puters competitive with the 370 series within 3 to 4 yearsof IBM’s offering. When these machines became availa-ble beginning in the mid-1970s, Japanese users quicklybegan trading in their IBM systems for those of domesticproducers. The number of IBM systems rented outactually declined for some models while the comparableJapanese offerings showed increasing usage.
The role of the Japanese Government continued to beimportant even after the market was officially opened in1975. Direct subsidies totaled some $1 billion between1976-81+ual to a quarter of private-sector investmentin R&D, plant, and equipment. If low-interest loans areincluded, government support nearly equaled privatesector investment during this period.
The Japanese computer firms have grown enormouslyin strength. They offer IBM compatible equipment that isoften as good, sometimes even better, than IBM itself, andthey are willing to drastically cut prices to capture marketshare. Hitachi, for example, has offered central banks,government agencies, and others discounts of 50 to 60percent below IBM prices in order to win customers.These tactics have worked. Between 1975-85, Japanesecomputer exports increased 35 times, while imports onlydoubled.
World reliance on IBM-compatible hardware andsoftware continues to be a serious weakness for Japanesefirms+ne which they have sometimes gone to greatlengths to circumvent. In 1982, for example, an FBI stingoperation caught Hitachi and Mitsubishi stealing IBMtechnology. Recently, Fujitsu was required to pay IBMnearly $1 billion for its ongoing unauthorized use of IBMsoftware. These and other incidents have led to Japan’scurrent Fifth Generation, Supercomputer, and otherprojects which include the goal of ending their depend-ence on IBM-compatible software.
Appendix F
Research and Development Consortial
The scale of investment required to catch up withJapanese and European HDTV projects naturally leads tothe consideration of R&D consortia. U.S. industry (andgovemrnent) have traditionally viewed consortia as un-necessary (given a technological lead), inefficient, andpossibly illegal (for antitrust reasons). But times change,and there are clear signs that the climate of opinion in theUnited States is changing. Many consortia have sprungup: industry-university consortia for basic research, likethe Semiconductor Research Corp.; private sector consor-tia aimed at long-term basic research, like the Microelec-tronics & Computer Technology Corp. (MCC); andconsortia aimed at developing manufacturing technology,notably Sematech and the National Center for Manufac-turing Sciences (NCMS).
Aside from simple catchup, there are good reasons forthis flowering of consortia in the United States. Asmanufacturing processes become more complex andtechnology more sophisticated, the cost of R&D is risingrapidly. Consortia offer a way of sharing these costs andrisks, which are becoming increasingly difficult to bear.For example, Yasuro Matsukara, general manager of~C’S WI diVisiOn,2 said that it would have cost hisfirm five times as much to develop their electron-beamlithography system independently. In contrast, six U.S.semiconductor equipment manufacturers attempted todevelop e-beam systems independently in the early1980s. Only Perkin-EImer succeeded, with the otherseventually writing off losses of more than $100 millionand quitting the market.
There are other reasons why consortia may be usefuland attractive. They can help to correct the well-knownU.S. tendency toward short-term thinking and strategy.3
They may generate externalities not captured by anindividual firm but available either to the economy or theindustry-such as a source of competition for foreignmonopoly suppliers of high-technology inputs.4 They cantrain people in methods and practices not prevalent intheir home corporations. They may help to diffuse newtechnologies, especially where consortia are designed tohelp companies catch up in areas of technologicalweakness. Finally, they can help participants create
formal and informal ties and alliances which may becritical for international competitiveness and technologydevelopment.
LESS tangible benefits of consortia may also beimportant. Some consortia may offer a forum for industryto discuss common problems and a framework forindustry to quickly establish technical standards andcommon equipment interfaces.
Initiating cooperation among otherwise strongly com-petitive firms can be difficult. Strong firms may hesitateto join, fearing loss of their proprietary technologies withlittle to gain fkom weaker firms. In Japan, governmentsupport then plays an important role both symbolicallyand substantively in enabling such collaboration—reassuring strong firms that they will get back at least asmuch as they give.
Companies similarly fear the loss of their best person-nel to a consortium and may consequently send theirsecond-best. Admiral Bobby Inman initially rejected 95percent of the researchers sent by member companies tothe Microelectronics & Computer Technology Corp.(MCC—a private U.S. consortium formed in 1982).Firms are likewise reluctant to share their in-house ideasor technologies. IBM’s and AT&T’s donations of impor-tant leading-edge technologies to Sematech suggest howvital they view its mission.
All these elements may be achievable during theenthusiasm of starting up a new consortium, but maintain-ing them-especially once the high-level champions inthe member firms have moved onto other problems—isa different and more difficult matter still. It is often hardfor firms to agree on an R&D agenda or to maintain along-term perspective. For example, managers are oftenforced to concentrate on the near-term bottom line withintheir firm, and therefore may in turn demand quicksupporting results from a consortium-though its purposeis longer term R&D, When R&D is successful, it can stillbe difficult to transfer technologies from the consortiumto the member firms, particularly when the firm has notmaintained good in-house technical expertise. Finally,even if the member firms are confident that these barriers
ls-e~for~s ~timfiIuA: Mark~oq “~andtie Entrepreneurial State,” unpublished monograph, MCC, Aust@ ‘IX; Joti D. ~VY ~d Ric~J. S~lS.‘Tnstitudcms and Innovation: Reseamh Collaboration As Technology Strategy in Japan,” and Richard J. S amuels, “Reseamh Collaboration in Jaw” MlT-Japan Scienceand Technology Pmgrarrx George R. Heaton, Jr., “The Truth About Japan’s Cooperative R&D,” NAS Issues in Science and Technology, fall 1988; “G-Operathg toCompete,” The Economist, Apr. 5, 1980, pp. 74-75; Charles H. Ferguson, “Technological Development, Strategic Behavior, and Government Policy in InformationTechnology Industries,” Ph.D. Thesis, Massachusetts Institute of Technology, 1988; Lee Smith, “Can Consortiums Defeat Japan?” Fortune, hne5, 1989; Sheridan Tatsuno,The Technopolis Straregy (New York NY: 1%.mtice-Hall, 1986); and interviews with personnel at the Semiconductor Reseach Corp., Reseamh Triangle Pink, NC; theMicmelectmnics & Computer Technology Corporation (MCC), Austiq TX; and SemakdL Austin, TX-
2Japm~5 ~~condmtor Cortsorhq 19 7 6 - 7 9 0
3For a di5m55im of ~ ~=a ~~d ~s ~~cy, so= of tie most impo~t of wtich are found in the U.S. fiancid ~ vimnment, See, U.S. Congress, Office ofTechnology Assessmmt, Making Things Better: Competing in Manufacturing, OTA-ITE-44 3 (W%shingtom DC: U.S. Govemmen t Ptiting Office, February 1990).
4~s Wm & aimof the U.S. Memories projec~ which failed in part because it aimed at generating externalities that might have goneuncaptumdby the firms themselves.
–lo4-
Appendi.x F4esearch and Development Consortia ● 105
can be overcome and that the consortium will besuccessful, in some cases they may have antitrustconcerns to deal with.
Consortia pose potential risks as well-of beingineffective and wasting money; of reducing competition;or of hampering creativity. Cooperative long-term R&Dalso does not address the areas where U.S. firms have hadthe greatest difficultieein manufacturing process andincremental product improvement.
Japanese consortia have provided many of these samebenefits to member firms, and have suffered many of thesame problems of initiating and sustaining cooperation.For example, judging itself to lead in the technology,Hitachi refused to collaborate in a $60 million 8-yearMITI sponsored R&D project to develop high-powerC02 lasers for flexible manufacturing. MITIneverthelesshelped fund Toshiba and Mitsubishi. Today, all threefirms have comparable C02 laser technology and thelevel of interfirrn competition has no-doubt accordinglyincreased. Some Japanese consortia have been abjectfailures as well. The Japan Software Co. (app. E) is anexample.
Japan is famous for its consortia, and justifiably so.One-third of Japanese industrial R&D is collaborative.But these consortia take a form different from thoseusually used to describe consortia in the United States.Fully 90 percent of collaborations are between two firmsonly, usually in the same keiretsu (industrial grouping).Only 6 percent of Japanese industry R&D is donecollaboratively between firms in the same business (i.e.,direct competitors). Much of the research done withinthese consortia is also done on a partitioned rather than acollaborative basis, with the results being shared but theresearch being done by individual firms in their own labs.Fewerthan 1 in200f Engineering Research Association~
horizontal R&D consortia-have had joint laboratories.Such partitioned efforts have nevertheless been importantin reducing duplication and the needed investment by anyindividual firm: Chapter 2 notes several examples of thisfor HDTV.
The critical element in the Japanese equation has beenthe role of MITI, not necessarily for providing the fundingfor collaborative R&D, but as a facilitator, supporter, andlong-range voice. During the Japanese catch-up phase,MITI often negotiated for patent rights on behalf of allJapanese firms, andin many cases demanded patent rightsas a condition for a foreign firm to have even a limitedpresence in the Japanese market. This kept patentlicensing fees uniformly low for Japanese firms andprovided all interested firms access to the teehnology—preventing any one from gaining monopoly control. Injoint R&D efforts, partitioning research between firmsallows the government to assign more difficult portions tostronger fins-effectively holding them back whileimplicitly aiding weaker firms. With all the participatingfirms having access to the same basic technologies bysuch mechanisms, competition is heightened and bynecessity also moves downstream into manufacturingprocess-the area where U.S. firms have most laggedtheir competitions.
Simple emulation of the Japanese model is not possibleor desirable in the U.S. environment. But as the Europeaninitiatives for science and technology show, there aremany ways in which the positive attributes of the Japanesemodel can be captured in a different setting. It is likely,therefore, that the right consortia operating under theappropriate conditions can help U.S. industry in somecritical sectors-perhaps including HDTV. The trick is tomake sure that the circumstances and conditions areIight.s
5A mm &~ed ~view of some of these issues can be found in OTA’s report, Maktig Things Better, op. ~~, footoo~ 3.
Appendix G
Acronyms and Glossary
Acronyms
—heIiCan Electrom“CS ksociatiq table z-s.
fk&cD —Active Matrix Liquid Crystal Display—ArnericanNational Standards Institute, table
2-3.ASK —Application Specific Integrated CircuitATSC —Advanced Television Standards committee,
table 2-3.AmC —Advanced Television Test Center, table 2-3.Am —Advanced TelevisionBTA —Broadcast Technology AssociationCM’S —Center for Advanced Television Studies,
table 2-3.cm -Charge-Coupled DeviceCCIR —Comite Consultatif International des
RadiocomrnunicationsCD -Compact DiskDARPA —Defense Advanced Research Projects
Agency (DoD)DAT —Digital Audio TapeDBS —Direct Broadcast SatelliteDoD —U.S. Department of DefenseDRAM —Dynamic Random Access MemoryDSP —Digital Signal ProcessorDVI —Digital Video InteractiveEBU —European Broadcasting UnionEDTV —Extended- or Enhanced-Definition
TelevisionEIA —Electronic Industries Association, table 2-3.EIAJ —Electronic Industries Association of Japan.FCC —Federal Communications CommissionFSS —Fixed Satellite ServicesGHz -GigahertzHD-MAC —High Definition Multiplexed Analog
HRsHDTVICIDTVIEEE
ISDNITFs
LCDLPTV
MSO
Component—High Resolution Systems—High Definition Television—Integrated Circuit.—Improved Definition Television—Institute of Electrical and Electronics
Engineers, table 2-3.—Integrated Services Digital Network—Instructional Television Fixed Service—International Telecommunications Union—Kilohertz—Liquid Crystal Display—Imw-Power Television—Megahertz—The Japanese Ministry of International Trade
and Industry—Multiple System Operator
MMDs
MFAA
MUSENAB
N(XA
NTIA
NTscPALSECAM
SMATVSMPTE
TvRo
VCR
VHs
—Multichannel Multipoint DistributionService
—Multi-media Terminals—Motion Picture Association of America,
table 2-3.—The Japanese Ministry of Posts and
Telecommunications—Multiple sub-Nyquist Sample Encoding—National Association of Broadcasters, table
2-3.—NationalCableTelevision Association, table
2-3.—Nippon Hoso Kyokai (Japan)—National Telecommunications and
Information Administration—National Television Systems Committee—Phase Alternation by Line—Sequential Encoded Color Amplitude
Modulation—Satellite Master Antenna Television—Society of Motion Picture and Television
Engineers, table 2-3.—Television Receive Only—Ultra High Frequency—Video Cassette Recorder—Very High Frequency—Video Home System—Video Random Access Memory
Glossary525/59.94; 625/50: The number of scan lines followed by
the field rate for the existing NTSC (U. S., Japan, etc.),PAL (Europe except France, China, etc.), SECAM(France, Soviet Union, etc.) color TV systems.
1050/59.94; 1125/60; 1250/50: The number of scan linesfollowed by the field rate for various HDTV systemproposals, corresponding to the United States, Japan,and Europe, respectively.
Active Matrix Liquid Crystal Display (AM/LCD): Anadvanced type of liquid crystal display.
Advanced Television (ATV): Refers generically to allthe improvements in TV over today’s system, includ-ing IDTV, EDTV, and HDTV.
Application Specific Integrated Circuit (ASIC): A typeof integrated circuit produced in relatively limitednumbers for a specific application.
Artifact: An audio or video error or defect introducedduring the processing or transmission of a TV signal.
Aspect Ratio: the ratio of a screen’s width to its height.Today’s TVs have a 4:3 aspect ratio. HDTV systemstypically call for a 5:3 or 16:9 ratio.
–lo6-
Appendix Mcronyms and Glossary ● 107
Bandwidth: The range of frequencies available for orused to carry an electronic signal.
Beta: The first, but less popular, format for home VCRs.Bit Rate: The rate at which digital data is carried or
transmitted, measured in units of bits (binary digits)per second. This is the digital equivalent of an analogbandwidth.
Broad Band: A signal that requires a large bandwidth tobe transmitted or equipment that must be capable ofreceiving and transmitting accurately a signal with alarge bandwidth.
Broadcast Technology Association (BTA): An or-ganization of private Japanese broadcasters and equip-ment manufacturers.
C-Band: The range of frequencies from 4 to 6 GHz. Seefigure 3-1.
Cable TV Labs: A lab setup by the N(2TA to test cablesystems, including those for ATV, table 2-3.
Charge-Coupled Devices (CCD): A type of solid-stateelectronic device used as a sensor in some types ofcareers.
ClearVision: The Japanese EDTV system.Comite Consultatif International des Radiocommu-
nications (CCIR): An organization under the ITUwhich studies technical questions and issues recom-mendations for international radio matters.
Compact Disk (CD): An opticaI storage medium used formusic and for computer data, among others.
Compatibility: The ability of one type of TV set toreceive and display the signals designed for anotherTV system. See box 4-3.
Digital Audio Tape (DAT): A new technology thatrecords music on magnetic tape in a digital format.DAT has many applications to computer data storage.
Digital Signal Processor (DSP): A type of digital chipthat manipulates a video (in the case of HDTV) signal.For the purposes here, this manipulation is usually toeither compress the signal so that it can be transmittedor decompress it and turn it back into a viewablepicture.
Digital Video Interactive (DVI): A digital technology inwhich the viewer can interact with the image beingshown. For example, a viewer might take a video‘‘walk’ through a building being designed by anarchitect and see the details of the interior or view thebuilding from any desired angle, at that person’sdiscretion.
Direct Broadcast Satellite (DBS): It transmits TVsignals directly to satellite receiver dishes at viewer’shomes. DBS is a high-power system that requires onlysmall dishes.
Downlink: The transmission (or receiver system) from asatellite.
Dynamic Random Access Memory (DRAM): A com-puter memory chip. The capacity of DRAMs ismeasured in bits—lKb (1,000 bits), IMb (1 million
bits), etc. The storage capacity of a computer or othersystems is measured in bytes-1 KB, MB, etc.—andone byte equals 8 bits.
Eureka 95: The joint project to develop Am systems forEurope.
European Broadcasting Union (EBU): A union ofEuropean broadcast organizations whose purpose is,among others, to develop standards for the exchange ofprogram material among its members.
Extended- or Enhanced-Definition TV (EDTV): Aform of TV that provides a better picture than today’sTV or IDTV using today’s broadcasts, but somewhatless resolution than HDTV. EDTV requires modestchanges in today’s NTSC broadcast signal, but iscompatible with it while remaining within today’schannel bandwidths. EDTV usually has a greater than4:3 aspect ratio.
Federal Communications Commission (FCC): TheU.S. Government agency dealing with communica-tions issues and allocation of the radio frequencyspectrum.
FCC Advisory Committee on Advanced TelevisionService: The industry committee set up by the FCC tomake recommendations on advanced television systembroadcasting standards.
Fiber: Optical fibers used to carry information, usually inthe form of pulses of light.
Field: The alternate lines that compose half of a completetelevision picture or frame. In the United States, fieldsare shown at a rate of 59.94 fields per second; inEurope, fields are shown at a rate of 50 fields persecond.
Fixed Satellite Services (FSS): Satellites that are as-signed geostationary orbits and provide informationtransmission services.
Frame: A complete television picture, including botheven and odd alternating scans. The frame rate in theUnited States is 29.97 ties per second; in Europe, itis 25 frames per second. If frames are shown at tooslow a rate, there can be an annoying flicker to thepicture.
Gigahertz (GHz): One billion cycles per second.Headend: A cable TV system’s control center where
incoming signak from satellites and other sources areput on cables going to subscribers.
Hertz (Hz): Cycles per second. One kHz is 1,000 cyclesper second; one MHz is 1 million cycles per second;one GHz is 1 billion cycles per second.
High Definition Multiplexed Analog Component (HD-MAC): The European HDTV system for DBS deliv-ery.
High Definition TV (HDTV): Usually defined as havingroughly twice the resolution of today’s TV systems, awider aspect ratio of 5:3 or more, and compact diskquality sound.
High Resolution Systems or High Definition Systems
108 . The Big Picture: HDTV and High-Resolution Systems
(HRS): Information systems that provide a highresolution visual image. See box 1-1 for examples.
HiVision: The Japanese HDTV system based on theirMUSE standard.
Improved Definition TV (IDTV): A television that usesdigital technologies to improve the picture seen evenwith today’s conventional broadcasts.
Instructional Television Fixed Service (ITFS): A TVdelivery service by line-of-sight microwave that theFCC licenses for use by educational institutions.
Integrated Services Digital Network (ISDN): A fullydigital telephone network now being implemented.This system makes use of the existing copper wireinfrastructure but adds improved electronics whichallow much higher data rates to be carried.
Interlaced Scan: A technique which first shows all theeven lines of the TV picture or frame, and then showsall the odd lines of the frame. Each set of linescorresponds to one field. This allows the picture to beshown without flicker while reducing the total band-width necessary to transmit the picture.
International Telecommunications Union (ITU): Anintergovernmental organization with 164 membercountries, whose purpose is to develop regulations andvoluntary recommendations, provide coordination oftelecommunication development, and foster technicalassistance fo developing countries. The CCIR is one ofthe organizations under the IT’U.
Kilohertz (kHz): One thousand cycles per secondKu-Band: The range of frequencies between 11 to 14
GHz. See figure 3-1.Liquid Crystal Display (LCD): The type of display used
on calculators and watches.Low-Power TV (LPTV): Stations licensed by the FCC
to use low transmitter power, usually in areas notlocally sewed by full-power stations.
Megahertz (MHz): One million cycles per second.Multichannel Multipoint Distribution Service (MMDS):
A TV delivery system using line-of-sight microwavewith four or more channels operated by a singlecompany. MMDS is often called wireless cable’ andis similar to I’I’FS in operation.
Multi-media Terminals (MMT): Computer terminalsthat can combine normal text and graphics withnear-real-time video images or other forms of visualdisplay.
Multiple System Operator (MSO): A company thatoperates more than one cable TV system.
Multiplex: See box 3-1.Multiple Sub-Nyquist Sampling Encoding (MUSE):
The bandwidth compression technique developed byJapan’s NHK to allow delivery of an HDTV qualitysignal over a DBS system.
National Telecommunications and Information Ad-ministration (NTIA): A U.S. Government agencyunder the Deparrnent of Commerce.
National Television Systems Committee (NTSC): Theindustry group that defiied the current U.S. B&Wandthen color TV standards. The NTSC system is used inthe United States, Cana@ Japan, and elsewhere.
Nippon Hoso Kyokai (NHK): The national radio andtelevision broadcasting organization for Japan. Hasextensively funded and coordinated HDTV develop-ment in Japan.
Optical Disks: Recording media including CDs that storeinformation in patterns of microscopic pits on thesurface of the disk, which can then be detected by asolid state laser and detector system and reproduced assound, images, or data.
Pay Per View: Program services purchased by sub-scribers on a per-program rather than per-month basis
Phase Alternation by Line (PAL): The type of TVsystem used in most European countries (with thenotable exception of France), The People’s Republicof China, Australia, and elsewhere.
Progressive Scan: A TV picture that is shown in a singlescan—the way we read a book-rather than byalternately sending all even and all odd Lines as for aninterlaced scan.
Resolution: A measure of a picture’s detail.Satellite Master Antenna Television (SMATV): Or
‘‘private cable”; a miniature cable system that receivesprogramming by satellite and serves a housing com-plex or hotel.
Sequential Encoded Color Amplitude Modulation(SECAM): The TV system used today in France, theSoviet Union, and elsewhere.
Taboo Channel: A TV channel left unused in order toprevent interference on adjacent active TV channels inthe same geographic area.
Television Receive Only (TVRO): A satellite receivingantema, also known as a downlink or a bac~ard dish.
Transponder: A satellite component that receives andretransmits a TV signal or perhaps many narrower-band data channels.
Ultra High Frequency (UHF): The band including TVchannels 14 through 83. See figure 3-1.
Uplink: The transmission or corresponding equipment toa satellite for relay.
Very High Frequency (VHF): The band including TVchannels 2-13, which are more powerful than UHFchannels. See figure 3-1.
Video Cassette Recorder (VCR): Piece of equipmentused for recording and replaying TV broadcasts orprerecorded video materhd at home.
Video Home System (VHS): The most common formatfor today’s VCRs.
Video Random Access Memory (VRAM): A type ofmemory chip similar to a DRAM, that is optimized forhigh-speed handling of video images.
Sources include: “Behind the Buzzwords,” Channels/Field Guide,1989; and “Everyone’s Thlking About HDTV, But What &e TheFacts?” Ampex, Redwood City, CA, 1989.
