OUTLINE................................................................................ 12 2.1 Efficient Program Production .................................... 13 2.1.1 DTPP .......................................................................................... 13 2.1.2 Image Expression Technology .............................. 14 2.1.3 Archive Retrieval and Preview System .......... 15 2.1.4 HDTV Optical Disk Camera...................................... 15 2.1.5 HDTV Hard Disk Recording System .................. 16 2.1.6 NTSC-HDTV Up Conversion Technology ...... 16 2.1.7 High-function Sound Pickup Technology .... 17 2.1.8 Translation Aid for English News ...................... 17 2.1.9 Transmission of Digital Program Material ..18 2.1.10 Color-temperature Compensating Filter ...... 18 2.2 Human-Friendly Broadcasting Services ...................... 19 2.2.1 Speech Recognition and Signal Processing for Broadcast News Transcription ...................... 19 2.2.2 Human-friendly Information Presentation ..20 2.2.3 Next Generation Human Interface .................... 20 2.2.4 Efficient Video Retrieval Based on Image Recognition .......................................................................... 21 2.2.5 User-centered Video Presentation Technique 21 OUTLINE .............................................................................. 22 3.1 Advanced Audio and Video Coding .......................... 23 3.1.1 Object-Based Video Coding .................................... 23 3.1.2 Generic Description for Contents Exchanges ..23 3.1.3 Digital Watermarking Technique for Copyright Protection ....................................24 3.1.4 Distributed Video and Audio Database ........ 24 3.1.5 Computer Network System for Broadcasting Stations .................................................................................. 25 3.1.6 Coding for 3-D Audio and Video Broadcasting Systems.................................................................................... 25 3.2 Super Reality Audio-Visual Systems ........................ 26 3.2.1 3-D Hi-Vision (HDTV) System .............................. 26 3.2.2 Future 3-D Television System .............................. 26 3.2.3 Depth Information Capturing and 3D Displays .................................................................................. 27 3.2.4 Ultra high-definition and wide-screen Display System.................................................................. 27 3.2.5 Sound Field Control ...................................................... 28 3.2.6 Advanced Image Acquisition System.............. 28 3.3 Research and Development of Frequency Resources .. 29 3.3.1 Revision of the Broadcasting-satellite Service Plan ........................................................................ 29 3.3.2 Multi-beam Receiving Flat Antenna for Satellite Broadcasting.................................................. 29 3.3.3 Future Satellite Broadcasting System ............ 30 3.3.4 Microwave/Millimeter-wave Transmission and Reception Technology ...................................... 31 3.3.5 Microwave Elements.................................................... 31 3.3.6 Optical Transmissions ................................................ 32 3.4 Recording and Mechatronics ...................................... 32 3.4.1 Future Home Server Technology ........................ 32 3.4.2 Tape Storage ...................................................................... 33 3.4.3 Optical Recording Device ........................................ 33 3.4.4 Perpendicular Magnetic Recording Head and Medium ........................................................................ 34 3.4.5 Micro-machining and Precision Measurement .................................................................... 34 3.5 Image Devices .............................................................. 35 3.5.1 Photoconductive Film.................................................. 35 3.5.2 Solid-state HARP Imaging Devices.................... 35 3.5.3 Field Emitter Image Sensor with HARP Film ..36 3.5.4 Next Generation Image Sensor ............................ 36 3.6 Display and Optical Devices .................................... 37 3.6.1 Plasma Display Panel (PDP) .................................... 37 3.6.2 Inorganic Electroluminescent (EL) Devices...... 38 3.6.3 Flexible Liquid Crystal (LC) Devices ................ 38 Studies toward New Broadcasting Services 1 OUTLINE ................................................................................ 5 1.1 Digital Satellite Broadcasting ...................................... 6 1.1.1 ISDB Data Broadcasting .............................................. 6 1.1.2 Digital Satellite Broadcasting Receiver ............ 7 1.1.3 ISDB Receiver System Architecture .................... 7 1.1.4 Source Coding for ISDB ................................................ 8 1.1.5 Home Server .......................................................................... 8 1.2 Digital Terrestrial Broadcasting .................................. 9 1.2.1 Digital Terrestrial Broadcasting Systems........ 9 1.2.2 SFN Broadcasting Radio Wave Relays .......... 10 1.2.3 Networks within and among Broadcast Stations .................................................................................... 10 1.2.4 Cable TV Transmission Technology ................ 11 Presentation of Research Results ........................................ 39 Application of Research Results .......................................... 42 Others.......................................................................................... 43 Foreword ........................................................................2 This Year: 1999 STRL ................................................4 1 NHK STRL ANNUAL REPORT 1999 | CONTENTS CONTENTS Research Results Research Data, etc. Studies for Improving Conventional Broadcasting Services 2 3 Studies for Opening Up Future Broadcasting Services
Transcript
OUTLINE ................................................................................122.1 Efficient Program Production ....................................13
2.1.1 DTPP ..........................................................................................132.1.2 Image Expression Technology ..............................142.1.3 Archive Retrieval and Preview System ..........152.1.4 HDTV Optical Disk Camera......................................152.1.5 HDTV Hard Disk Recording System ..................162.1.6 NTSC-HDTV Up Conversion Technology......162.1.7 High-function Sound Pickup Technology ....172.1.8 Translation Aid for English News ......................172.1.9 Transmission of Digital Program Material ..182.1.10 Color-temperature Compensating Filter ......18
2.2 Human-Friendly Broadcasting Services ......................192.2.1 Speech Recognition and Signal Processing
for Broadcast News Transcription......................192.2.2 Human-friendly Information Presentation ..202.2.3 Next Generation Human Interface ....................202.2.4 Efficient Video Retrieval Based on Image
Recognition ..........................................................................212.2.5 User-centered Video Presentation Technique 21
OUTLINE ..............................................................................223.1 Advanced Audio and Video Coding ..........................23
3.1.1 Object-Based Video Coding....................................233.1.2 Generic Description for Contents Exchanges ..233.1.3 Digital Watermarking Technique for
Copyright Protection....................................243.1.4 Distributed Video and Audio Database ........243.1.5 Computer Network System for Broadcasting
Stations ..................................................................................253.1.6 Coding for 3-D Audio and Video Broadcasting
3.2 Super Reality Audio-Visual Systems ........................263.2.1 3-D Hi-Vision (HDTV) System ..............................263.2.2 Future 3-D Television System ..............................263.2.3 Depth Information Capturing and 3D
Displays ..................................................................................273.2.4 Ultra high-definition and wide-screen
Display System..................................................................273.2.5 Sound Field Control ......................................................283.2.6 Advanced Image Acquisition System..............28
3.3 Research and Development of Frequency Resources ..293.3.1 Revision of the Broadcasting-satellite
Service Plan ........................................................................293.3.2 Multi-beam Receiving Flat Antenna for
Satellite Broadcasting..................................................293.3.3 Future Satellite Broadcasting System ............303.3.4 Microwave/Millimeter-wave Transmission
and Reception Technology......................................313.3.5 Microwave Elements....................................................313.3.6 Optical Transmissions ................................................32
3.4 Recording and Mechatronics ......................................323.4.1 Future Home Server Technology........................323.4.2 Tape Storage ......................................................................333.4.3 Optical Recording Device ........................................333.4.4 Perpendicular Magnetic Recording Head
and Medium ........................................................................343.4.5 Micro-machining and Precision
3.5 Image Devices ..............................................................353.5.1 Photoconductive Film..................................................353.5.2 Solid-state HARP Imaging Devices....................353.5.3 Field Emitter Image Sensor with HARP Film ..363.5.4 Next Generation Image Sensor ............................36
OUTLINE ................................................................................51.1 Digital Satellite Broadcasting ......................................6
1.1.1 ISDB Data Broadcasting ..............................................61.1.2 Digital Satellite Broadcasting Receiver ............71.1.3 ISDB Receiver System Architecture ....................71.1.4 Source Coding for ISDB ................................................81.1.5 Home Server..........................................................................8
1.2 Digital Terrestrial Broadcasting ..................................91.2.1 Digital Terrestrial Broadcasting Systems........91.2.2 SFN Broadcasting Radio Wave Relays ..........101.2.3 Networks within and among Broadcast
Stations....................................................................................101.2.4 Cable TV Transmission Technology................11
Presentation of Research Results ........................................39Application of Research Results ..........................................42Others ..........................................................................................43
Foreword ........................................................................2 This Year: 1999 STRL ................................................4
1NHK STRL ANNUAL REPORT 1999 | CONTENTS
CONTENTS
Research Results
Research Data, etc.
Studies for Improving ConventionalBroadcasting Services
2
3 Studies for Opening Up Future BroadcastingServices
2 NHK STRL ANNUAL REPORT 1999 | FOREWORD
Director-General Dr. Osamu YAMADA
Foreword
Dr. YAMADA joined NHK in 1967. Since
1971, he has been working at the NHK
Science and Technical Research Laboratories
on the research and development of
broadcasting system, including teletext, FM
multiplex broadcasting, and digital satellite
and terrestrial broadcasting systems.
He was appointed Director-General of the
NHK Science and Technical Research
Laboratories in June, 1999.
Biography
Toward the Creation of New Technologies to Support 21st Century Broadcasting
3NHK STRL ANNUAL REPORT 1999 | FOREWORD
Since its establishment in 1930, The NHK
Science & Technical Research Laboratories
has long contributed to the progress of
various broadcasting media, from the start of
radio, FM, television, and satellite broadcasting,
to Hi-Vision (HDTV) broadcasting. Fiscal 1999
has proven to be an exceptionally successful and
fruitful year, a year in which many of our past
digital broadcasting related research
accomplishments have come to fruition.
Digital satellite broadcasting will begin in
December 2000, realizing digital HDTV and data
broadcasting, both parts of the ISDB (Integrated
Services Digital Broadcasting) services that will fully
reveal the excellence of digital broadcasting. These
services have been long anticipated ever since
research was launched for such future broadcasting
system in 1982. Our laboratories performed the
world's first studies on the utilization of Plasma
Display Panels (PDP) for high definition displays,
and manufacturers have already made large PDP
displays available. As for HDTV, NHK has long
participated in the ITU (International
Telecommunication Union), making positive
contributions to advance the international
standardization of NHK's 1,125 scanning line HDTV
format. The many years of research and effort
finally bloomed with the approval of unification of
the 1125 HDTV studio standard as an ITU
Recommendation in March 2000.
NHK's flexible, large capacity digital satellite
broadcasting system (ISDB-S: Integrated Services
Digital Broadcasting-Satellite), including rain
attenuation countermeasures, was approved as an
ITU Recommendation in October 1999.
Regarding digital terrestrial broadcasting, the
Telecommunications Technology Council (TTC)
reported on the digital terrestrial television
broadcasting system to the Ministry of Posts and
Telecommunications (MPT) in May 1999. This was
followed by reports on the digital terrestrial sound
broadcasting system in November of the same year.
Our laboratories have promoted the development
and field tests of these systems and contributed to
the establishment of these digital terrestrial
broadcasting standards.
The advance of digitization and multimedia
broadcasting further enhances the importance of
efficient high-quality program production within a
limited budget. In this regard we also emphasize
the importance of ascertaining the exact needs of
the field, and promoting research to support
program production using the latest technologies.
In recent years, we have also pursued research on
"human-friendly broadcasting services" to ensure
that our elderly viewers and people with visual or
hearing impairments can enjoy our broadcasting
services. We believe the start of the news closed-
captioning service for news using the STRL's speech
recognition system on the March 27th edition of
"News at 7" was a significant accomplishment.
Furthermore, with the implementation of
practical digital broadcasting imminent, we are
continuing work on establishing new research
themes for the future. As the explosive diffusion of
the Internet and cellular phones continues, it is
clear that the fusion of broadcasting and
communications will advance rapidly. We will
continue to promote broadcasting research so that
desired programs and information can be obtained
anytime, from anywhere, by anyone during this
broadcasting era. Research will steadily progress on
super reality audio-visual systems such as 3-D
television and large ultrahigh-definition televisions
that are anticipated to be a part of the future
broadcasting services beyond HDTV.
The 21st century will finally start next year. In
October 2001, the STRL's new research facilities
will be completed. During this revolutionary period,
we will complete and announce to the public the
"STRL Mid-to Long-Term Vision" that maps the
future research endeavors of STRL policy based on
our foresight of broadcasting industry conditions
over the next 10 to 15 years. Even in this rapidly-
changing era, we consistently pursue the mission of
the public broadcaster: to provide better
broadcasting from the standpoint of our viewers
through our research and development efforts.
Osamu YAMADA
The digital terrestrial television broadcasting(DTTB) system was reported to the MPT bythe TTC in May 1999. This was followed bythe report on the digital terrestrial soundbroadcasting (DTSB) system after verificationtests at the ARIB in November 1999. Thesetwo reports cover digital terrestrialbroadcasting transmission systems.
DTTB and DTSB System Reported
Vehicle for the experiment pf DTSB system
Technological Developments
Data broadcasting system standards wereestablished in October 1999 by the ARIB. Withthis, all the standards for the start of digitalbroadcasting in December 2000 have beencompleted. The Digital Satellite BroadcastingReception System Test Center Consortium hasalso produced and distributed test signalsbased on the standards to accelerate thedevelopment of digital broadcasting receivers.Experimental broadcasting began on March15, 2000, allowing for the final adjustmentand operational confirmation on receivers.
Preparations for theCommencement of Digital SatelliteBroadcasting Advance Steadily
Experimental broadcasting commencementceremony
Technological Developments
4 NHK STRL ANNUAL REPORT 1999 | THIS YEAR
1999
2000
Apr.May
Aug.Oct.
Nov.
Jan. Feb.March
1999
2000
Apr.May
Aug.Oct.
Nov.
Jan. Feb.March
A ground-breaking ceremony is conducted for the new STRL research buildings.The digital terrestrial television broadcasting system is reported to the MPT (Ministry of Posts and Telecommunications) by the TTC (Telecommunication Technology Council).Their Imperial Highnesses, The Crown Prince and Princess visit the STRL. At the STRL open house, the entire NHK organization works together to present to the public the specific benefits that they can expect from digital satellite broadcasting. The following research results are presented to the public for the first time:- A 3-D virtual reality audio system- A multi-beam receiving antenna for satellite broadcasting- An image-based virtual studioThe TVML's first partial program application.A recommendation for a digital satellite broadcasting system (ISDB-S) is approved by the ITU (International Telecommunication Union).Data broadcasting system standards are established at the ARIB (Association of Radio Industries and Businesses) standards meeting.An "Insect microphone" that is so sensitive it can pick up the sound of a snail's heartbeat is developed. The digital terrestrial sound broadcasting system is reported to the MPT by the TTC.Mr. Yashiro, the Minister of Posts and Telecommunications, visits the STRL.Their Imperial Highnesses Hitachi visit the STRL.The 1,125 scanning line HDTV studio format is approved as the unification HDTV studio standard by the ITU.The "Axi-vision Camera" is developed, which in addition to shooting regular video images, can also detect the depth information of objects. The Digital BS Test Center Consortium launches experimental radio broadcasts.NHK starts live closed-captioned broadcasting using the STRL's broadcast news transcription system.
The Science & Technical Research Laboratories held their annual open house in May 1999,presenting to the public the results of their research and highlighting the specific benefitsthat will result from digital satellite broadcasting, scheduled to start in December 2000.Exhibitions related to future program contents production and human-friendlybroadcasting were on display. Due to the great interest shown in digital satellitebroadcasting, featuring Hi-Vision(HDTV) and convenient data broadcasting, the totalattendance reached 23,000, the highest ever in the history of the STRL Open House.
Digital Satellite Broadcasting Service Image Exhibited by NHKThe STRL Open House
Closed-captioning by a broadcast newstranscription system
Technological Developments
In order to meet the demand for a closed-captioning service on news programs, theSTRL has developed a broadcast newstranscription system. This systemautomatically recognizes an announcers'speech, allowing closed-captioning to becreated live. A partial live closed-captioningservice was started on the March 27thedition of "News at 7" using this system.
Development of Broadcast NewsTranscription System
The Crown Prince and Princess visited the STRLprior to the public Open House
The NHK Science and Technical Research Laboratories promoted the research anddevelopment of digital satellite and digital terrestrial broadcasting in order to realizeIntegrated Services Digital Broadcasting (ISDB), a system that provides various servicesutilizing video, audio, and data.
1.1 Digital Satellite BroadcastingDigital satellite broadcasting transmission scheme standardization was completed by the end of fiscal 1998. Our
main activities during 1999 involved participating in the data broadcasting standardization process, which had beenbehind schedule, supporting receiver development, and developing operational specifications in preparation for thestart of digital broadcasting in December 2000.
Specifically, in relation to data broadcasting systems, we contributed to the ARIB (Association of Radio Industriesand Businesses) standardization process, which resulted in the establishment of the ARIB standard STD-B24,approved in October 1999.
The BS-P (BS Platform Consortium), participated in by all digital satellite broadcasting providers, completedoperational specifications and corresponding receiver specifications for actual broadcasting. Our laboratoriesworked in conjunction with the BS-P's activities to ensure that these practical rules and receiver specificationswould appropriately reflect digital satellite broadcasting standards. Our activities also involved interaction with the"Digital Satellite Broadcasting Reception System Test Center Consortium" that was established to ensure theproduction of reliable and low-cost receivers through their collaboration with both broadcasters and receivermanufacturers. To ensure the smooth start of digital satellite broadcasting, we worked closely with the Consortiumin regard to test signal production and the launch of experimental radio broadcasts for receiver adjustments, inaddition to contributing to a wide range of preparatory work for the service.
The annual STRL Open House held in May 1999, proved to be a significant promotional event, revealing to ourviewers the new broadcasting services that upcoming digital satellite broadcasting will provide. In order for theattendees to experience these new services at the open house, we exhibited prototype data broadcasting services,digital satellite broadcast receivers, and home servers.
Concerning international standardization, the digital satellite broadcasting system, ISDB-S (ISDB-Satellite), wasapproved as an ITU-R Recommendation in October 1999. NHK was rewarded for its many years of globalenlightenment activity regarding HDTV systems which are now playing a major role in digital satellite broadcasting,when our 1125 HDTV format (1,125 scanning lines) was approved as the unification HDTV studio standard by theITU-R in March 2000.
1.2 Digital Terrestrial BroadcastingAs for digital terrestrial broadcasting, the system that reflects our laboratories' research results was established as
a Japanese national standard. Based on the results of large-scale field trials conducted by the ARIB, the digitalterrestrial television broadcasting (DTTB) system was reported to the Ministry of Posts and Telecommunications(MPT) by the Telecommunications Technology Council (TTC) in May 1999. With regard to digital terrestrial soundbroadcasting, a broadcasting system in common with DTTB was reported by the TTC in November 1999, the reportbeing based on ARIB verification tests conducted starting June 1999.
Research and development to forward practical digital terrestrial broadcasting was also advanced. We continuedresearch for improving the reliability of a coupling canceler necessary for SFNs (Single Frequency Networks) viabroadcast radio relays.
We are also involved in the international promotion of digital terrestrial broadcasting systems to support theirinternational standardization and global diffusion. At the February 2000 meeting of the ITU-R SG11, a draftrecommendation was adopted for ISDB-T (ISDB-Terrestrial), a DTTB system. We also provided instructions forexperiments and held seminars in Asia and Brazil in cooperation with DiBEG (Digital Broadcasting Expert Group),an organization that promotes the diffusion of enlightenment information about ISDB-T to other countries.
Research Results 1
Studies toward NewBroadcasting Services
OUTLINE
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services6
Preparing for the start of digital satellitebroadcasting in December 2000, NHK has contributedto the standardization process of data broadcastingsystem at the meetings of the TelecommunicationsTechnology Council (TTC) and the Association ofRadio Industries and Businesses (ARIB).
As experts in the field, and in cooperation withthe NHK Engineering Administration Department,we assisted the BS-P (BS Platform Consortium)Receiver Specification Subcommittee inestablishing operational specifications. Thesespecifications were discussed by the organizationsshown in Table 1, and enacted as standards or operationalguidelines in October 1999.
Data Broadcasting Coding Scheme and ProtocolUpon the completion of data broadcasting system standards
by the ARIB, our laboratories proposed and contributed aprototype system model and protocol stack (Figure 1) for use asa framework of standards for a data broadcasting codingscheme.
Regarding the multimedia coding scheme, thestandardization of which had been behind schedule, it wasdecided at the TTC Digital Broadcasting System Committeemeeting held in July 1999, that the standardization should bebased on the XML coding scheme, not on MHEG, as promotedearly on. Prior to this decision, the ARIB had been promoting thedevelopment of an XML-based multimedia-coding schemesince March, 1999. Along the same lines, our laboratoriesconducted examinations and verification tests on the standardin cooperation with receiver manufacturers. These contributedto the development of an XML application language, BML(Broadcast Markup Language), and XML system, BXML(Broadcast XML), that enables the presentation of an XML fileon BML browsers.
These ARIB data broadcasting system standards wereenacted as STD-B24 in October 1999.
Conditional Access SystemConcerning the conditional Access System (CAS) of digital
satellite broadcasting, we contributed to the ARIB and BS-Pstandardization process by proposing a messaging function.This was accomplished in cooperation with related NHKdepartments, including the Digital Broadcast DevelopmentBureau and the Audience Services Department, together withother broadcasters and manufacturers.
The ARIB standards related to CAS were enacted as STD-B25
in October 1999.We also researched the problems and technical system of
CAS for cable TV networks, and participated in discussions withthe Japan Cable Television Engineering Association.
STRL Open House ExhibitionsDeveloped prototype contents and receivers based on the
services listed below were exhibited at the STRL Open House. - Weather information- Anytime news- Earthquake information- Program associated data services- Electronic Program Guide (EPG)At the open house, the prototype multimedia contents
produced by the NHK Multimedia Development Departmentwere encoded and transmitted in the form of multiplexed TSs(Transport Streams). The receiver exhibition highlightedmanufacturers' set-top type systems currently underdevelopment that feature TS input interface for reception. Apicture from the open house is shown in Figure 2. Thisexhibition not only verified the performance of prototype databroadcasting services, but also clearly presented to visitors theevolution of the TV medium from watching to using.
Advanced Data Broadcasting System Working GroupXML System Working GroupMultiplexing Working GroupConditional Access System Working Group
Digital Satellite Receiver Development PartyDigital Receiver System Working GroupHigh-speed Digital Interface Working Group
Receiver Specification Subcommittee (Download, receiver, data broadcasting, SI/EPG, CAS,
interactive communication, transmission)
ARIB STD-B24: "Data coding and transmission specification for digital broadcasting"
ARIB STD-B25: " Conditional Access Specification for digital satellite broadcasting services using broadcasting satellites"
ARIB STD-B21: "Digital receiver for digital satellite broadcasting services using broadcasting satellites (desirable specifications)"ARIB TR-B15 "Operational guidelines for digital satellite broadcasting services using broadcasting satellites"
Organization that conducted the discussions
ARIB
Name of established standard or operational guidelines
BS-P
Digital Satellite Broadcasting System Development PartyDigital Satellite/Terrestrial Broadcasting Common Technology Subcommittee
Table 1: Deliberative organizations and standards for digital satellite broadcasting in Japan
ISDB Data Broadcasting1.1.1
Figure 2: Data broadcasting exhibition at the NHK STRL Open House
Multimedia Services
PES PESSection Section
PSI/SI, event index
Video/audio
Closed-caption/
superimposition
Data stream
Multimedia coding
Mono-media coding
Carousel transmission
scheme
MPEG2-TS (TS packet)
Various communication
protocols
Interactive networks (PSTN/mobile
network)
indicates specifications covered by ARIB STD-B24
TV/Sound services
ExtensionFigure 1: Protocol stack
We contributed to the establishment of operationalguidelines for digital satellite broadcasting downloadschemes. The ARIB specification for software downloadingwas determined based on an NHK proposal. Thisspecification is designed to distribute receiver systemsoftware through the broadcasting channel to modifyand/or update receiver system software, and to add the newfeatures. The specification enables "push" downloadingrequested by the broadcasters or by receiver manufacturers.The specification is designed to be highly reliable, highlyefficient, and to result in energy savings. The specificationcan also adapt to a variety of downloading situations : theupdating target can be a specific receiver model or all
receivers, a part of the installed software or all the software,and the download can be automatic or user selectable.
Research also continued on next generation receiversystem architecture. In fiscal 1999, extensive investigationswere performed on next generation receivers, especially onAPIs (Application Program Interface) to establish a commonplatform. Software independence from any OS (operatingsystem) and any hardware is achieved by a virtual machine,and wide connectivity of the receiver is achieved byinterfaces with next generation infrastructures includingIMT-2000 and home network systems. For the 1999 STRLOpen House, we developed a prototype next generationreceiver equipped with a Java virtual machine (JavaVM)
(inserted figure). Thisreceiver makes nextgeneration downloadingpossible, allowing theupdate of all receiversutilizing a commoncode. This prototypereceiver presents thepossibility of newbroadcasting servicesthrough the addition ofnew functions such asan Internet browser.
Support for the Test Center ConsortiumTo develop digital satellite broadcasting receivers, test signals
are necessary to confirm a system's conformance to establishedstandards. For this purpose, the "Digital Satellite BroadcastingReception System Test Center Consortium" was established inApril 1999, to produce and distribute test signals for digitalsatellite broadcasting receivers. We have also participated in theproduction of such test signals to draw up test items andproduce test signal specifications based on our research anddevelopment accomplishments. The test signals for digitaltelevision services were produced using a variety of video
coding, multiplexing, and channel coding equipment developedby the STRL. We were also involved in test signal production fordata broadcasting, SI (Service Information), and interactiveservices. We cooperated further with the Consortium in theestablishment of test items and transmission facilities for thelaunch of experimental radio broadcasts.
Evaluation of Hierarchical Transmission-modulation ServiceSubjective evaluation tests were conducted to quantitatively
measure the effects of a hierarchical transmission-modulationservice. The experiment proceeded as follows: We askedviewers to select their favorite channel from among thosepresented. Certain channels were switched to low layer (audioonly, low frame-rate picture, reduced-size picture, or lowresolution picture) to look into the percentage of viewers thatwould continue viewing the same channel (channel holdingrate). As a result, it was found that a low layer service with lowresolution images obtained a channel holding rate of 90% orhigher (inserted figure).
We compared this result with the channel holding rate forservices just below the analog satellite service availabilitylimit (a reception CN ratio below 6dB that results in soundinterruption and a discontinuation of service, which we callthe service availability limit). This examination verified thatthe hierarchical transmission-modulation service would beable to maintain a channel holding rate equal to or higherthan that of present analog satellite broadcasting.
ISDB Receiver System Architecture1.1.3
Prototype next generation receiver (first stage prototype)
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services 7
Characteristics- Equipped with a Java virtual machine.- Common downloads for all receiver
models.- Downloads from communication
networks in addition to broadcastchannel.
Digital Satellite Broadcasting Receiver1.1.2
SportsDramaInformation program
Audio only Reduced-size picture video
Low frame-rate picture
Low resolution picture
Analog satellite
(C/N=6dB, Audio is
available.)
0
10
20
30
40
50
60
70
80
90
100
Ch
ann
el h
old
ing
rate
Low layer services
(%)
Hierarchical transmission evaluation results
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services
Video CodingFor MPEG-2 picture coding, research has been
promoted on improving picture quality bymodifying the coding schemes while maintainingcompatibility with existing decoders. In the MPEG-2 picture coding; motion vectors are transmittedusing differential coding. Continuous macro-block(MB) suppresses the dispersion of detected motionvectors and field/frame modes, reducing theoverhead bitrate needed to transmit motion vectorinformation. This spare bitrate can be allotted toDCT data, thereby improving the picture quality.Computer simulation confirmed an approximately 2-2.5 Mbpsdecrease in overhead information in regard to video data withhigh coding criticality levels (see the figure).
Picture quality assessment tests were conducted onMPEG-2 coded cascaded connection video. These testsrevealed that, among the various coding parameters rangingfrom the first codec to the second codec, prediction types(called I,P and B pictures) and motion vectors influencepicture quality significantly.
The criticality (the number of generated bits at a fixedquantization characteristic) and S/N of HDTV broadcasts at22Mbps were measured over a 2 month period for predictionkey application to establish the relationship between codingpicture quality and the digital broadcasting time rate. As aresult, the time rate distribution was clarified.
Audio CodingWe have continued research on encoding techniques to
improve the sound quality within the range of MPEG-2 AAC(Advanced Audio Coding) system standards. Examinationswere performed on sound reproduction bandwidths andencoded sound quality. We developed an encoding methodto control the DCT coefficient of the bandwidth, which isoperated based on the Perceptual Entropy informationconcerning the bound on bitrate for transparent audiocoding. It is confirmed that this method improves encodedsound quality at the bit rate of 112-144kbps/stereo.
The AAC subjective listening tests results for IntermediateAudio Quality showed that the improvement of AM stereoquality could be attained at approximately 32kbps/stereo bitrate.
Information quantity distribution on each encoded video frame (video at "market")
Source Coding for ISDB1.1.4
8
Home server research and development continues toproduce more advanced digital broadcasting services (insertedfigure). In fiscal 1999, we constructed a prototype home serverthat can playback any one program while recording two otherdigital HDTV programs and verified the operation of the homesever. This accomplishment was based on a newly developedengineering technique for video hard disks.
New functions were incorporated into this prototype homeserver, such as the "Top function" that enables completeviewing from the beginning of a program already inprogress, and "Pause function" that can suspend a programon the air for later viewing. This home server was developed
for watching a set of scenes automatically rearranged fromrecorded programs using program index informationaccording to the viewer's choices. This function allows theviewer to watch highlights of a long news program orportions of a sports program in which the viewer’s favoriteathlete is participating.
Easy recording reservation was made possible through theuse of EPGs (Electronic Program Guides) distributed by digitalbroadcasting. The system displays the viewer’s mostfrequently watched programs in each category to assist inrecording reservation, and the system is also capable ofseries recording of the same programs on a weekly basis.
Storage AgentH i e r a r c h i c a l S t o r a g e M a n a g e m e n t S y s t e m
Programs to be constantly updated, Delayed viewing
Program bank, Digest viewing
Digital Broadcasting Automatic selection of programs the viewer wants to watch.Simultaneous recording/playback of several programs.
News Weather Forecast
Program Guide
NHK Special Drama Movie
Selecting Viewing Anytime.Interactive Program Viewing.
Home server functions
Home Server1.1.5
Digital Terrestrial Television Broadcasting (DTTB)A report on the digital terrestrial television
broadcasting system (ISDB-T: Integrated Services DigitalBroadcasting-Terrestrial) was presented to the Minister ofPost and Telecommunication by the TelecommunicationsTechnology Council (TTC) in May 1999 after the large-scale field trials conducted by the Association of RadioIndustries and Businesses (ARIB). At the ITU-R StudyGroup 11 Meeting in February 2000, ISDB-T wasapproved as one of its international standards. Ourlaboratories have consistently contributed to the system'sdevelopment from system proposal through field trials.
Using the segmented BST (Band SegmentedTransmission)-OFDM scheme, the ISDB-T is capable ofproviding a variety of services from Hi-Vision (HDTV) for fixedreception to mobile-multimedia broadcasting. The outline andtransmission parameters of the DTTB are listed in Table 1.
The large-scale field trials utilizing Tokyo Tower and UHFChannel 15 with a 100W output showed that to realize areception ratio of 99% or more, the minimum field strengthof fixed reception must be approximately 57dB V/m in64QAM modulation. SFN experiments using the Sawararelay station yielded fundamental knowledge on SFN, andthe relationships between bit-error rates and active multi-path signals within an SFN service area.
To clarify technical requirements for the introduction ofDTTB stations, protection-ratios between current analog TVand DTTB were measured using the prototype hardware.The measured interference data showed that the co-channelDU ratio from DTTB to analog TV was 45dB, and that the DUratio of lower adjacent-channel interference was 0dB andthat of upper-adjacent channel interference was 10dB.These results were adopted as the interference DU ratios of"Technical Requirements for the Establishment of DigitalTerrestrial Television Broadcasting Stations" by the TTC. Afield strength of 60dB V/m or greater was established todefine the DTTB service area of fixed reception.
In addition to the efforts of the DTTB system-standardization in Japan, the STRL provided instruction forand cooperated in ISDB-T experiments in Singapore, HongKong, China, and Brazil which were carried out by theDiBEG (Digital Broadcasting Expert Group).
Digital Terrestrial Sound Broadcasting (DTSB)As for DTSB, narrow-bandwidth ISDB-T was standardized
in November 1999 after the large-scale field trials and the
indoor interference experiments carried out by ARIB fromJune to September 1999.
The audio source coding scheme, AAC (Advance AudioCoding), is the same as that of DTTB, though the three input-sampling frequencies are supplemented with their half-ratesto expand the quality range of audio services for mobilereception. The digital signal multiplex scheme is what MPEG-2 systems have in common with DTTB.
The transmission scheme is the BST-OFDM, which is thesame as used for ISDB-T except for the bandwidth. Narrow-bandwidth ISDB-T provides two kinds of transmissionbandwidths, 430 kHz of one segment and about 1.29 MHz ofthree such segments to accommodate larger bit-rateservices. One-segment receivers can receive a segmentwithin a TV signal of 13 segments as shown in Figure 1.
In addition to system specifications, the TTC report alsoincluded the technical requirements for a connectedsegments transmission where multiple segments keepingthe carrier-orthogonality are transmitted without guardbands between them. The connected transmission is a verypromising method to greatly improve frequency-utilizationefficiency through the inessential guard band.
Data on mobile reception was also collected during large-scale field trials that took place using the Shibaexperimental station and the Sawara relay station on VHFchannel 7. A passenger car-style STRL experimental vehiclewas used for the mobile reception tests.
The experimental vehicle traveled on expressways and inthe city to verify high-pass mobile reception. The resultsrevealed that a minimum field strength of 38dB V/m withDQPSK modulation, and 44dB V/m with 16QAMmodulation, is required to obtain correct-reception timeratios higher than 99%. In a four city metropolitan areatravel test, including Matsudo city, the electromagnetic fieldstrength necessary to obtain a reception ratio above 99%using DQPSK modulation was 35dB V/m and greater.
Regarding interference protection ratios, detailed mutualinterference characteristics were gauged utilizing parameterssuch as the frequency differences between analog televisionand sound channel locations. The experiment's results werecompiled in "Technical Requirements for the Establishmentof Digital Terrestrial Sound Broadcasting Stations."Furthermore, the STRL held demonstrations in Australia incooperation with DiBEG, and seminars examining theMalaysian system as part of the overseas diffusion promotionactivities for narrow-bandwidth ISDB-T.
Frequency spectrum
13 segments
13 segment receiver 3 segment receiver
Partial reception
Partial reception
Digital Terrestrial Sound Broadcasting
1-segment transmission
3-segment transmission
1 segment receiver
430kHz
Digital Terrestrial Television Broadcasting
Figure 1: Digital terrestrial broadcasting transmission signals and receivers
Digital Terrestrial Broadcasting Systems1.2.1
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services 9
Information source codingMultiplexingTransmission line coding
Transmission modeCarrier spacingNumber of carriersEffective symbol lengthGuard intervals
Modulation schemeInterleaveError correction
Control signalHierarchical transmission
14 kHz1405252 s
22 kHz2809504 s
31 kHz56171.008ms
1/32, 1/16, 1/8, 1/4 of effective symbol lengthDQPSK, QPSK, 16QAM, 64QAMFrequency and timeInner coding : Convolutional code
(1/2, 2/3, 3/4, 5/6, 7/8 )Outer coding : RS (204, 188 )TMCC and pilot signalsA maximum of 3 parameter-sets are possible on one channel
Table 1: Digital terrestrial television broadcasting (DTTB) system
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services
In order to realize a digital terrestrialbroadcasting single frequency broadcastingnetwork (SFN: Single Frequency Networks) via abroadcasting signal relay system, a SFN technicalproblem needs to be overcome. The problemconsists of distortions caused by couplingbetween transmitting and receiving antennas,resulting in characteristic transmission distortionand induced amplifier oscillations (insertedfigure). As a countermeasure, research is beingpromoted on a coupling wave canceler thateliminates such distortions using electric circuits.
We have obtained coupling characteristics using existinganalog relay stations and confirmed the cancellation theorythrough computer simulations and indoor preliminaryexperiments. During fiscal 1999, we developed a cancellationalgorithm with high tracking performance for coupling wavefluctuation. Additionally, thorough consideration was given tothe influence of various errors and distortion in the practicalhardware. These results were incorporated into a preliminaryexperimental device and indoor experiments were performedthrough which we confirmed that the cancellation algorithmwas capable of the desired performance under the followingconditions: the existence of three coupling waves, the syntheticelectric power of these waves being nearly equal to the electricpower of the master broadcasting station wave(D/U 0dB), anda coupling wave variation of 2.5Hz.
Secondly, moving forward by field experiment verifications,we proceeded with the trial manufacture of an advancedcoupling cancellation device. The device conformed to ISDB-T, the digital terrestrial broadcasting transmission standards,and could also be applied to a D/U inversion with strongerreception signals from the coupling wave than those of thedesired wave. Furthermore, we developed a high-speedcoupling characteristic measuring device conformant to theISDB-T system to study coupling characteristics at a digitalbroadcasting relay station. Combining these devices with aplanar reception antenna with improved side lobecharacteristics, we advanced preparation for a field trial usingthe TAO's (Telecommunications Advancement Organizationof Japan) collaborative facilities for R&D on digital terrestrialbroadcasting.
SFN Broadcasting Radio Wave Relays1.2.2
Networks within a Broadcast StationAn examination of general-purpose networks within a
broadcast station using computer network technology isbeing promoted. The characteristics of video/audio datatraffic occurring within a broadcast station can be practicallycategorized as follows:- Burst transmissions that perform relatively short, repeated
transfers, such as for editing work.- Transmissions with extended transfer times, such as for
the storage/transmission of program material.In 1999, with the aim of analyzing the video stream traffic
involved with editing work, we measured a VCR's On(action)/Off (stop) information during off-line editing work.Based on the results, we found that the probability densityfunctions of both the On and Off duration lengths areapproximated by the sum of two exponential functions, andwe modeled an editing terminal's transfer demand based onthis standard (inserted figure). Various simulations on datatraffic were performed using this model for fundamentalexamination regarding network design within abroadcasting station.
Networks among Broadcast StationsTo achieve smooth digital terrestrial broadcasting
program-switching, a transport stream splicing process is ofprimary importance. During fiscal 1999, we confirmed abasic splicing procedure. Additionally, investigations also
began on more advanced splicing techniques that rarelyinfluence receivers, and multiplexed signal transmissioncontrol techniques between broadcast stations.
Networks within and among Broadcast Stations1.2.3
10
On duration length (second)
0 20 40
1
10-1
10-2
10-3
10-4
Measured valueApproximation formula
Pro
babi
lity
den
sity
Material transfer demand model
Master station
Relay broadcast station
Receiver (viewer)
Transmission antenna
Reception antenna
Master station wave
Coupling
f
fCoupling f
f
Signal coupling from transmission antenna to reception antenna
To make ISDB services available to cable TV subscribers,research and development for digital transmissions overcable TV networks is being promoted.
Re-transmission of Digital Satellite BroadcastingDigital satellite broadcasting's information rate per
transponder is approximately 52 Mbps, meaning that no singlecable TV channel is capable of handling such an amount ofdata. Digital satellite broadcasting also adopted a systemwhich can transmit multiple (a maximum of 8) TSs (TransportStreams). We therefore developed a multiple TS transmissionsystem to efficiently re-transmit digital satellite broadcastsover cable TV networks. This system utilizes the samemodulation scheme as the existing digital cable televisionbroadcasting system (ITU-T Rec. J.83 Annex C) andtransmits framed multiple TSs (see figure). Following anexhibition at the STRL open house, the STRL proposedthe system to the Japan Cable Television EngineeringAssociation Standardization Committee, and contributedto its development and field trials. The STRL alsocontributed to the Telecommunications TechnologyCouncil (TTC).
Re-transmission of Digital Terrestrial TelevisionBroadcasting (DTTB)
Regarding the re-transmission of digital terrestrialtelevision broadcasting (DTTB) over cable TV networks,
continuously since 1998, we have been contributing toexperiments and discussions concerning the technicalrequirements of a pass-through method which performs re-transmission without modulation scheme changes. This wasreported on by the TTC in January 2000.
Interactive TransmissionWe also developed an ATM layer transmission interface
for cable TV transmission channels. Transmissionexperiments to connect a terminal and multiple servers with64QAM (downstream) and QPSK (upstream) modems wereperformed.
Structure of frame
Framing
Existing transmission scheme
Frame headerSync byte
TS-1 TS-2 TS-3 TS
Multiple TSs (developed scheme)
Single TS (existing scheme)
Common stream form
Re-transmission over cable TV networks
Multiple TS transmission scheme
Cable TV Transmission Technology1.2.4
NHK STRL ANNUAL REPORT 1999 | Studies toward New Broadcasting Services 11
12 NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services
For the improvement of conventional broadcasting services, with a view to their futuredigitization, studies continued on technologies that support efficient and more expressiveprogram production, and human-friendly services that make possible the provision ofprograms that are easy for all viewers to watch, listen to, and understand.
2.1 Efficient Program ProductionOur research on efficient program production technology has promoted technologies to support the program
production process from the planning to the editing stages. This has involved the construction of a database comprisedof video materials and related information such as scripts or coverage memos that are generated at each stage of theprogram production process. In order to extract desired programs or data promptly from a massive broadcasting stationdatabase, we strove to develop a retrieval system that extracts and attaches indexing information to video data andother material, and then retrieves desired information based on the indexing. We also participated in and contributed tothe global production of agent TV specifications, that allow for the customized selection and presentation of programsbased on EPGs (Electronic Program Guides) and the viewer's personal preferences and interests.
Regarding new video expression technology, studies continued on an intelligent robotic camera system that canautomatically shoot images by emulating cameramen's shooting techniques. We constructed a prototype systemwhereby two robot cameras shoot a talk show program by exchanging information between them. Research alsoprogressed on an image-based virtual studio where actual video images can be compartmentalized and freelysynthesized to create a studio set. We further improved the system to allow for easier manipulation and morenatural synthetic video images. Enhanced functions related to picture quality and CG expression were alsoincorporated into TVML (TV program Making Language), a simple computer language used to describe TV programscripts, and the system was actually used for TV programs.
Efforts were also made to promote technological developments that correspond to the diversifying and advancingHDTV program production process which is expected to further expand with the implementation of digital satellitebroadcasting. Aiming to achieve high-speed and high-density recording with an HDTV optical disk camera, researchcontinued on optical disks, optical heads, and signal processing. Research was also conducted on an HDTV harddisk recording system for application to HDTV non-linear editing. Additionally, an HDTV OBL (Outside BroadcastingLink) was also developed for mobile relay broadcast use, such as for marathons.
Additional program production related research included the conversion of an "ice-zone microphone," originallycreated for use at the Nagano Olympic skating competition, into an "insect microphone," that is now capable ofpicking up micro-sounds from an insect. This insect microphone was used on scientific programs and receivedfavorable comments. In order to contribute to the efficient use of stored programs, emphasis was placed on thedevelopment of a video archive retrieval system that can be simultaneously accessed by multiple users over stationnetworks. Research also proceeded on computer technology to support the Japanese-English translation ofbroadcasting materials, especially for the efficient translation of newscasts.
2.2 Human-Friendly Broadcasting ServicesAccelerating research on speech recognition systems that can automatically generate closed-captioning from a
newscasters' speech resulted in the achievement of a performance goal; a speech recognition accuracy of 95% wasobtained when an anchor studio announcer read a manuscript. On the basis of this accomplishment, we constructeda practical system for experimental use in live news programs, and a broadcast news transcription system was putinto service for the March 27th, 2000, edition of "News at 7."
To make highly operable digital broadcasting available to all viewers, we also studied human-friendly interfaces such asremote control devices and operation screens. We also improved the performance of tactile communication devices whichdeliver digital broadcasting textual information to the visually-impaired or to individuals with both visual and hearingimpairments. Regarding a speech rate converter capable of slowing down an announcers' rapid speech to a morecomfortable speed for elderly viewers, we developed PC software that can perform real time speech rate conversion.
Studies for Improving ConventionalBroadcasting Services
Research Results 2
OUTLINE
Desk Top Program Production (DTPP)Research has continued on a prototype hierarchical TV
program editing system based on TV program plan sheets tostore video data together with various types of textualinformation such as TV program scripts that are generated ateach stage of the program production process.
During fiscal 1999, enhancements were made to the videooperating functions of a workstation version of this systemand other functions were added including a network-basedcollaborative work environment. We started developing a datainput camera with an efficient information input process afteran application experiment conducted in 1998, using the TVprogram "Science Eye", revealed the need for increasedefficiency. The data input camera is capable of videomanagement based on plan information of DTPP at the time ofshooting, and automates the video material classificationprocess, thus improving editorial efficiency. It is also equippedwith a gyro and a GPS to automate the assignment ofkeywords such as camera operations and shot locations.
We developed a prototype program generation system foreducational English conversation programs, which is basedon a study of program information description methodsusing XML (eXtensible Markup Language) for secondary
material use and materialretrieval. Experiments onthe production ofconversational-style TVMLprograms and Internet webpages using basic Englishconversation informationitems in a template format(keywords of theconversation, application
examples, etc.) were conducted. With the aim of digest viewing or video retrieval in digital
broadcasting, we carried out indexing production experimentson a sport (soccer) video, and produced a summary video ofthe game using the spectators' cheering.
TVML (TV program Making Language)TVML is a computer language that describes TV programs
using text-based scripts. A script written in TVML isinterpreted by computer software called the TVML Player,and a real-time program is automatically produced. Duringfiscal 1999, we completed a Windows version of the TVMLPlayer and distributed it as freeware. This enabled therealization of a program production environment on anindividual personal computer. As a professional version ofthe program with broadcasting quality output, we developedTVML Player PRO by making improvements in the picturequality, sound quality, CG expressions, and others. TVMLPlayer PRO was actually used for three NHK broadcastprograms (Figure 1). Other research projects related toTVML included the automated production of a newsprogram and talk show, a camera simulator that allows formanual manipulation of a TVML camera by connectingTVML to a camera tripod, and automatic TVML productionusing a DTPP TV program plan sheet. TVML presentationsand demonstrations were performed nationally andinternationally to high acclaim.
Broadcasting AgentsJoint research on a European project, FACTS (FIPA Agent
Communication Technologies and Services), advancedresearch and development on agent TV that utilizes agenttechnology (software which manages/retrieves/processesmassive amounts of information for users) to realize programselection functions for viewers. The final agent TV version wascompleted during fiscal 1999.
This completed agent TV is a system which selects andpresents programs for individual viewers employing userprofile information, EPGs (Electronic Program Guides), andpersonal preferences and interests (Figure 2). The systemconsists of 20 different types of agents and it providesadvanced functions such as automated user profilecompilation, Natural Language Interfaces (NLI), andcommunication support between different Agent Platforms.Among these functions, NHK contributed to the developmentof a Natural Language Agent (NLA) for the provision of NLIsand a Provider Agent (PA) for the distribution of programinformation including EPGs.
This system is based on a FIPA (Foundation for IntelligentPhysical Agents) international specification standard, and wehave greatly contributed to the improvement of thisspecification by reporting on the actual operatingcharacteristics of this new system.
Program IndexingConcerning broadcasting station databases, we are
conducting research on methods for program and dataretrieval and browsing by utilizing indexing informationattached to video and textual data.
As for the indexing support technique, a prototype systemwas employed that utilized a topic detection method inventedduring fiscal 1998 for TV-news manuscript databases. Theprototype system was exhibited at the STRL open house in fiscal1999, and was highly evaluated. Additionally, we applied thistechnique to an individual topic detection method that selectstopics according to a user's personal interests and needs.
Furthermore, by providing system suggestions based onknowledge we had acquired through our DTPP research, wecontributed to the establishment of an international MPEG-7standard that will realize the common utilization of dataindexing stored in different databases by means of one searchengine.
Figure 1: Data input camera
DTPP2.1.1
DisplayHome server
Home
InternetBroadcaster
Program description data(EPG)
Agent platformProgram flow
Viewer
Program information flowMessage flow
Broker agent
Interface agent
Provider agent
User profile agent
Figure 2: FACTS Agent TV
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services 13
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services
Intelligent Robot CameraWe have continued the study of intelligent
robotic camera systems that allow onecameraman to operate multiple cameras usingthe same shooting techniques as severalexperienced operators. In fiscal 1999, weclarified each camera's individual role, the typesof shooting cuts, and the technical director'scommands to the cameramen by analyzing theshooting techniques used with multiplecameras in a talk show program. Based on thisdata, we constructed a prototype two camerarobotic operating system that can be used for atalk show involving two performers using a flipchart (Figure 1). Two sensor cameras (fixed)measure an object's position and capture acomplete studio image for the real-timeestimation of each object's 3-D coordinates.The robot cameras compose their shots using the followinginformation along with the object's 3-D coordinate data:
- Information concerning how and which object the otherrobot camera is shooting at any given moment.
- Sound information concerning which person is speakingand for how long.
We also test manufactured a man-machine interface thatallows a cameraman to operate the system with voicecommands and a pan pod.
Additionally, in order to apply the system to live sportsprograms, basic experiments were performed regardingsensing techniques for the detection of a small object movingabout in a large area using a sensor camera's pan, tilt, andzooming functions.
Image-based Virtual StudioAiming to construct a high quality and functional virtual
studio, research has been progressing on an image-based virtualstudio that can compose electronic studio props based on actualshot images. We integrated past research accomplishmentssuch as high-resolution omnidirectional imaging (environmentalvideo components)for the production of the surrounding scenery,and an element technology that breaks down video images ofsmall objects, such as a vase or a desk, into computer graphicsparts (three-dimensional video components), and developed aprototype Hi-Vision (HDTV) image-based virtual studio. Theprototype system was comprehensively evaluated, identifyingfuture problem that need to be overcome.
Regarding environmental video components, ultra-high-resolution omnidirectional images were multi-layered to
express locationalr e l a t i o n s h i p sbetween objectsat the time ofs y n t h e s i s ,realizing theillusion of asubject appearingand hiding behindan object.Concerning threed i m e n s i o n a l
video components, we developed a technique to estimateobject texture and specular-reflected coefficients by eliminatingspecific illumination information. As a result, it becamepossible to alter the illuminating conditions of three-dimensional video components as if they were in theenvironmental video components filmed under differentlighting conditions.
In a prototype virtual studio, a virtual camera manipulatorenabling the adjustment of camera position, direction(anywhere within 360 ), and viewing angles, realizes a videoproduction environment with a high degree of flexibility.Moreover, a graphical user interface (GUI) was constructed tomake changes in the selection and arrangement of three-dimensional video components feasible, in addition to changesin illumination conditions. This system creates CG actors andactresses that replace actual performer images, and utilizes GUIto control the movement and speech of the characters. Throughexperiments on these tools, we confirmed they could be appliedto HDTV program production, although camerawork andillumination conditions entail certain restrictions.
Electronic Video Correction on HDTV Drama ProductionIn the production of HDTV dramas, images are shown in high-
definition, making incongruities and imperfections in studio setsand costumes more obvious, such as the borderline of wigsworn in Japanese historical dramas. Although efforts had beenmade to correct the problem by using special wigs designed forHDTV programming, this solution resulted in higher costs andlonger dressing times. Therefore, we developed an electronicprocessing technology for after-shooting of HDTV images,specifically targeting the problem with actors' wigs. By specifyingthe portion of the video that needs correcting, this technologyautomatically processes the massive amount of informationcontained in all of the picture frames.
In fiscal 1999, we realized high-speed processing byoptimizing the developed algorithm. The processing time forone-second of HDTV material is less than 10 minutes. Incooperation with the Engineering Administration Departmentand the Broadcast Engineering Department of NHK, we furtherdeveloped a practical verification system employing thistechnology and confirmed its effectiveness and simplicity byutilizing the system on a historical drama.
Hi-Vision sensor camera
Wireless microphone input
3-D motion analysis system
2-D motion analysis systemHost computer
Cameraman's head
Cameraman's hand
Condition transition chart
Robot camera 1
Robot camera 2
Cameraman's eyesDetection by color extraction
Combined feedback typeFeed forward control
Figure 1: Robot camera system
Image Expression Technology2.1.2
14
Figure 2: Image-based virtual studiooutput example
Person,etc. (CG)
Environmentalvideo
components
Three-dimensional
videocomponents
The NHK Archives are scheduled to open in 2003. Theyrequire program retrieval and previewing functions on anetwork level using a PC. We constructed a prototype videoretrieval and preview system for the archive, employing ahierarchical recording system to evaluate the functionality.
For the basic prototype system manufactured in fiscal1998, we proposed a hierarchical storage managementsystem which combined a high-speed accessible hard diskwith a large capacity, low cost tape streamer to store largevolumes of highly compressed reference video programs.This basic prototype showed that the realization of acompact, low cost, high-speed, large capacity, and highlyreliable system is possible.
In fiscal 1999, we applied this hierarchical storage
management system to the reference video server for aretrieval and preview system. Advances were made in thedevelopment of contents management methods for multipleaccess as well as in the connection to a network (insertedfigure). As a result, we demonstrated that multiple userscould retrieve and preview programs from their own PCs ona network in this experimental system. This retrieval systemis capable of referencing program-related information inaddition to the instant previewing of key frames and videoprograms.
For program retrieval, the system uses a key word search,which is enhanced by the narrow down and synonymreference functions.
In order to improve the preview operability, a general-purpose program playbackconsole can be connected to apersonal terminal. It realizeseasy variable forward andbackward playback functionsby its jog dial. The system canalso display down-convertedHi-Vision (HDTV) for referencevideo programs in addition tostandard definition television.
Prototype retrieval and preview system
Archive Retrieval and Preview System2.1.3
The optical disk camera's advantageous functions, suchas quick field editing, make it more suitable as a camcorderfor program coverage than conventional tape-type cameras.
An HDTV optical disk camera cannot be realized withoutattaining a high transfer rate for large amounts of high-density data. Therefore, we have promoted research onoptical disks, optical heads, and signal-processing schemes.
Regarding high data bit rates, by making both the opticalhead, using red laserdiode (wavelength 635nm), and thesignal processing system faster, we developed a phase-change optical disk evaluation system (inserted picture) thatcan measure a bit error rate (BER) of 100Mbps and higher,and conducted high-speed recording experiments.
Multiple beam parallel recording was effective inimproving the data transfer rate, and we confirmed thepossibility of near-infrared wavelength (780nm) use throughthe research of multi-beam optical sources. We also deviseda control method with a feed-forward system that couldfollow the high-speed disk revolutions with high accuracy.
Regarding high-density characteristics, we started the trialmanufacture of both optical heads and optical disks for bluelaserdiode, and made recording experiments using them.This was possible due to the rapid progress being made onthe development of blue laserdiodes (wavelength 405nm).Moreover, we began research and characteristic evaluationson an optical element that is highly sensitive to the blue lightwavelength. Additionally, we developed simulation software
for the reproduction process, and other various software forreproduced signal processing. This made examinationspossible in regard to the relationship between the bit errorrate (BER) and necessary S/N ratios, and also in regard to theoptimum signal-processing method.
We participated in the working group of the ARIB (Associationof Radio Industries and Businesses), discussing requirements foran broadcasting-use optical disk. We also initiated studies on asuitable HDTV optical disk format for broadcasting.
HDTV Optical Disk Camera2.1.4
High-speed recording experiment on a phase-change optical disk
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services 15
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services
To realize an HDTV nonlinear editing system, we startedresearch on an HDTV video hard disk system in fiscal 1997.That same year, we proposed a hard disk evaluation tool forvideo applications, and we further proposed, in fiscal 1999, atechnique to describe characteristics of hard disk drives withsmall number of parameters, using the tool. The parametersnecessary for this technique are the known disk rotationalspeed, the physical location of the recorded data, and theseek time (the time required for the head to move to thedesired track) obtained by means of the evaluation tool(inserted figure). Additionally, the creationof a virtual hard disk drive by parameteradjustments led to the establishment ofguidelines for future performanceimprovements of conventional hard diskdrives. Applying these accomplishments,we provided technical cooperation tomanufacturers, thus contributing to themanufacture of a practical hard diskrecording system capable of simultaneousplayback and recording/playback, andsynchronous simultaneous recording andplayback for two HDTV systems.
Aiming to improve the reliability of videohard disks, we have also developed anultrasonic vibration spindle that vibratesthe disk during start and stop functions to
reduce the friction between the head and disk. By ultrasonichead levitation experiments, we analyzed the optimumvibrating mode. As a result, the friction coefficient wasreduced more than 75%, thus confirming the system'seffectiveness in reducing head friction. We also discoveredthe possibility of miniaturizing the spindle by increasingsound levels with a reflector plate. Furthermore, wedeveloped a sound sensor to estimate system hard diskproblems caused by head crashes, and disk damagedetection experiments were carried out.
HDTV Hard Disk Recording System2.1.5
Digital satellite broadcasting which is to start inDecember, 2000, will provide digital HDTV broadcasting. Toprepare for this advance, we are pursuing research toachieve high quality images for material converted by NTSC-HDTV up-converters (devices used to convert NTSC signalsto digital HDTV signals) in order to utilize existing andarchive material in an HDTV format.
This system, differently from conventional linearinterpolation, first learns the corresponding relationshipbetween the waveform edge of HDTV images and that ofconverted NTSC data signals. Applying this information, thesystem estimates the HDTV image contours that are to beproduced based on the NTSC signal edge data, thus realizinga high-resolution image. The effectiveness of the system on
still pictures had already been confirmed.In fiscal 1999, we examined the problems related to still
image processing and extracted future research agendas,with the aim of applying the results to video images. Weproposed an estimation technique utilizing high-resolutionelements that correspond to motion information and atechnique involving the addition of high-resolution elementsto the converted image, and confirmed the basicperformance of the system(inserted figure).
NTSC-HDTV Up Conversion Technology2.1.6
16
Motion information detection
Scanning line conversion
Edge detection/analysis
High resolution components
NTSC HDTV
High-definition edge signal studied in advance
NTSC-HDTV up conversion system structure
Disk rotation
Physical data location
Readout data
Data readout
Rotational latency
SeekHead position at the moment
Hard disk readout function image
Selective Sound Receiving Microphone SystemResearch is ongoing on a new sound pickup system that
can isolate a desired sound source even when the soundsource and other noise sources are located in the samedirection. Simulations have shown that approximately 40dBof noise can be suppressed when the sound sources arelocated within 30 of the microphone's main axis and thedesired sound and noise have no correlation. Digital filterprocessing was adopted to correct the characteristicdifferences between the two microphones used for thesystem, one of the factors that decreased the system'sseparation performance, and this improved thecharacteristic differences, maintaining them under 1% of thewaveform signal amplitude. Additionally, wemanufactured a dedicated signal processor forobtaining the separation coefficient, and fieldtested the device with its possiblecommercialization in view.
Insect MicrophoneWith the aim of picking up small sounds, such as
those generated by insects, we constructed amicrophone consisting of a dynamic electroacoustictransducer attached to a contact pin. Although itsoperating principle is similar to that of a recordneedle, we fabricated a contact pin supportingmechanism with the high compliance and realized
a high-sensitivity and wide-band operation. Its sensitivity is-33dBV/0.1G, and it successfully picked up the sound of asnail's heartbeat, the sound of a snail eating a carrot, and thesound of an ant's footsteps, etc. This new microphone wasintroduced in TV programs (inserted picture).
IC MicrophoneThe development of an ultra-small and high-performance
condenser microphone using silicon chip processingtechnology is now underway. In fiscal 1999, we fabricated a 5
m thick silicon diaphragm using a photoresist pattern designbased on parameters obtained from simulations done in 1998.We also began fabrication of the microphone structure.
High-function Sound Pickup Technology2.1.7
Machine Aided Translation System: From Japanese toEnglish
We are investigating a machine aided translation systemfrom Japanese to English (Translators' Workbench) whichsupports such translation work. This system includes variousfunctions such as (a) a translation example browser, (b) a termretrieval function, and (c) a bilingual Web retrieval function. Inthe future and as a supporting function, this system will beintegrated with Japanese-English machine translation, whichis also being promoted. In fiscal 1999, we realized the above-mentioned (a) and (b) functions, as well as part of the (c)function, and applied them to the construction of a prototypebilingual editor which helps a translator to work on a project.
Especially, a "translation example browser" which includes the(a) and (b) functions was unified in the English article writingsystem that the International Planning & BroadcastingDepartment and the Multimedia Development Department ofNHK have been developing (inserted figure).
Machine Translation from Japanese to EnglishRegarding machine translation from Japanese to English,
research continued on a rule-based translation system. Thissystem performs translation in three distinct phases: theanalysis of Japanese; the transfer of Japanese structures intoEnglish ones; and the generation of English script. RegardingJapanese-English transfer, we constructed a pre-transfer
function that links a Japanese bunsetsu dependencyanalysis to the Japanese-English transfer process. Toachieve high-speed transfer, advances have been madeon the processing procedure, including the limiting ofapplicable rule selections. We produced a bilingualdictionary consisting of 300 function words forJapanese-English transfer. The vocabulary of contentwords has been increased from 2,000 words in 1998 to5,000 words presently. Regarding the generation ofEnglish script, we constructed a prototype of thegeneration engine which determines the English wordorder. As a result, the English translation of Japanesenews can be accomplished, although the system is stillsomewhat limited in its expressions.
Translation Aid for English News2.1.8
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services 17
Contact pinDynamic microphone unit
Pedestal
Insect microphone
Translation browser
Presentation of corresponding sentences in both language articles
Japanese manuscript How was this expression
translated in the past?
Translator
Expression retrieval
Desk
Japanese article English article Japanese article English article
Japanese-English bilingual article database
Separation of words
Morphological analysis
Retrieval of articles containing this expression or words
Search engine
Translation example browser
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services
HDTV OBL (Outside broadcasting link)OFDM-OBLs (Orthogonal Frequency Division Multiplexing
-Outside Broadcasting Link) for standard televisionbroadcasts that use the 800 MHz-band and HDTV servicelinks in the 6-7 GHz-band have been developed with the aimof utilizing them for live relay broadcasts of races, such asmarathons.
Regarding the latter HDTV system, we have researched asynchronous demodulation scheme using pilot carrierswhich is expected to have superior characteristics incomparison to the differential demodulation scheme.Optimizing the configuration of pilot carriers resulted inpreferable mobile transmission characteristics in themicrowave band with a high Doppler shift.
On the other hand, in the case of road race relay
broadcasts, including non line-of-sight propagation,utilization of the 800 MHz-band is more advantageous,requiring fewer receiving stations for transmission. Thus, weinvestigated a transmission scheme which secures anecessary transmission capacity for digital HDTV materialthrough the application of two channels and multilevel QAMformats, jointly promoting the development of thetransmitter with the NHK Engineering AdministrationDepartment.
SNGIn order to realize a portable SNG RF terminal, research
on a low-profile active antenna with which active devicessuch as power amplifiers can be integrated is beingpromoted. In fiscal 1999, we calculated the influence on thephase and amplitude of the radiated waves due to mutualcoupling between antenna elements and confirmed that thebeam pointing error will be at an allowable level for ourportable SNG RF terminal.
We studied the structure of transmitting active antennasusing EM simulations and found that the electromagneticallycoupled slot antenna is a promising antenna element for aportable SNG RF terminal.
We also studied a feeding method which can reduceactive devices by appropriate coupling between antennaelements.
6-7 GHz-band16QAM-OFDM64QAM-OFDM15.5MHz75Mbps
25Mbps(16QAM, r=1/2)57Mbps(64QAM, r=3/4)Fixed to mobile relay within the range of sight
800 MHz-band16QAM-OFDM
8.5MHz 287Mbps
29Mbps(16QAM, r=1/2)
Mobile relay including non line-of-sight propagation
r: Coding rate
Frequency bandMajor modulation schemesOccupied bandwidthMaximum transmission capacityBit rate
Application
OFDM-OBL under development
Transmission of Digital Program Material2.1.9
In TV program shooting, a color glass filter is inserted intothe camera optics to compensate color temperature whenthe light source is changed. However, it is difficult to inserta color filter during shooting, and the amount ofcompensation can not be changed for the glass filter.Therefore, in cooperation with NHK Nagoya BroadcastingStation, we developed a liquid crystal (LC) color-temperature compensating filter that can continuouslyadjust color temperature at high speed. The filter consists oftwo glass substrates coated with a transparent electrodeand nematic LCs mixed with dichroic dyes imbeddedbetween the substrates. The dyes' light absorption spectrachange depending on the direction of the molecularalignment. In fiscal 1998, we realized an absorptionspectrum characteristic that converts solar light intoincandescent by the use of two types of dyes withabsorption peaks at 398nm and 556nm. We designed a filterusing two LC cells so that the quantity of color temperaturecompensation is independent of the polarization of theincident light.
In fiscal 1999, we optimized the dye concentration ratio.In addition, we changed the LC orientation fromhomogeneous to homeotropic alignment. As a result, thefilter can continuously control the color temperature oftransmitted light between 3700K and 5600K with an appliedvoltage of up to 10V. Response time is approximately 80ms.A camera equipped with the filter (inserted picture) was
used in the live program "Lunchtime Nippon" in Novemberand December of 1999, and accomplished the shooting ofnatural video images under varied lighting conditions withdifferent color temperatures. For automatic operation, wemade a prototype system that drives the filter according tothe spectral change in lighting, detected by a small sensorinstalled on the camera. The basic operation of the systemwas verified, however; the correction error increases if thelight source is out of sight of the sensor. Furtherimprovement in the system is under way.
Continued research and development of a broadcast newsspeech transcription system has been promoted.Universities and researchers working both inside andoutside of research institutions in Japan have cooperatedtogether toward the development of an automatic closed-captioning system. Such a system would be used to displayan announcer's speech in captions during a news program.To accelerate the development of such a system, we arecreating a news speech database of NHK's main newsprograms, such as "News at 7" and "Good Morning Japan."Starting on April 1, 1999, the contents of each day's newshave been added to the database. We have also made effortsto accumulate language data by creating transcriptions ofsuch programs as "Sunday Debate" and "Sunday Sports."
In relation to an Acoustic Model, extensive examinationswere conducted on a model production method used for anews speech database to improve its recognitionperformance. As a first stage toward speaker dependentspeech recognition from speaker independent speechrecognition, we performed a clustering of announcers'speech data and developed a technique to select anappropriate cluster Acoustic Model at the time ofrecognition. As a result, we obtained a 10 to 20 % reductionin the recognition processing time.
Regarding the Language Model, we tested a method toutilize a manuscript produced by a reporter. This manuscriptis usually submitted right before the news broadcast.Further, we constructed a technique to automaticallyestimate a word based on its context. This involvesmanually inputting words that were not included in thereporter's manuscript. In another recognition method wedeveloped, we register in advance where a word appears ina reporter's manuscript, and the system performsrecognition prioritizing of word selection based on thisregistered word order. These improvements resulted in arecognition accuracy approximately 4 to 5 % higher.
Regarding the decoder and related technology, the formerrecognition system had used information from a wholesentence to output a subtitle. We switched to a new systemwhich finalizes a recognition result sequentially as itrecognizes words. This reduced the average time to output arecognition result from 7.2 down to 0.6 second.
We integrated these research results and used an updatedrecognition system on the September 30th, 1999, editions of"News at 7," "News at Noon," and "Good Morning Japan."
The system provided an average recognition accuracy of86% on the portion containing a studio announcer's speech(492 sentences)(this includes 97% of the anchor's speechcontaining 122 sentences, and 88% of the 213 sports andweather related sentences). Recognition delay on theportion containing the anchor was 3 seconds. Especiallynotable is the recognition accuracy of 95% and higher for thestudio anchor, thus achieving our target performance.
We also pursued development of a system whichmanually detects and corrects mistakes in the recognitionresult instantly. This system has 2 stages in its correctionprocess. The first is "error detection" and the second is"error correction." In order to enhance the accuracy of thiscorrection process, we also constructed a prototype systemcapable of synchronous presentation of a sound and itstextual manuscript to an error detector and corrector. This isaccomplished through the application of a speech rateconversion system.
The recognition performance during the anchor's portionof a broadcast has reached a level for practical use. It wasdecided to launch this closed-captioning service on the NHKnews program "News at 7" on the evening of March 27,2000. In preparation for this occasion, a practical speechrecognition system and a practical recognition errorcorrection system were developed.
Figure 1: Broadcast news transcription system
Speech Recognition and Signal Processing for Broadcast News Transcription2.2.1
Simultaneous subtitling production system
Speech input
Speech recognition system
Database for speech recognition
Recognition error correction
Electronic general scenario
Closed-captioning
Final check
Figure 2: Automatic news closed-caption production system
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services 19
Binocular viewing of stereo 3-D images is known to giverise to a conflict between vergence and accommodation. Inorder to clarify the extent to which the accommodationfunction is impaired in elderly subjects, we studiedaccommodation responses to binocular 3-D images andperformed a subjective evaluation of the results. The resultsshowed that the accommodation response induced inelderly subjects by movement in depth of 3-D images wasless than 1/3 of that induced in younger subjects, indicatingthat older people respond less to 3-D images.
With the aim of improving TV color reproduction takinginto consideration the changes in chromatic visionassociated with aging, we attempted to ascertain mostpreferred TV color temperature conditions for both youngand elderly viewers. The results revealed that the mostpreferred color temperature for the elderly is in the vicinityof 16,000K, compared to around 9,300K for young subjects(see figure). We also clarified visual psychological effectsunder conditions of most preferred color temperature bymeans of principal component analysis.
With a view to improving access for visually impaired usersto ISDB menu screens etc., we continued work on a multi-modal interface which uses both Braille and sounds, includingauditory warning signals, in place of the usual graphical userinterface (GUI). Basic research was conducted on presentingthe GUI environment through tactile modalities such as shapeand surface texture, as well as sound.
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services
Elderly group (average age 68)Young group (average age 23)
Eva
luat
ion
rat
ing
Image: ITE skin color chart (female face)
1.0
0.5
0.0
-0.5
-1.0
5,000 10,000 15,000 20,000
Color temperature (K)
Result of optimal color temperature experiment
Human-friendly Information Presentation2.2.2
Digital satellite broadcasting will bring a variety of colorfulnew services. In order to determine factors which will makethe digital receiver remote controls easy for anybody to use,we test-manufactured various prototype remote controlsand evaluated the coordination between the actual remoteoperation and its display on a computer-simulated screen.Three types of remote control were used: a button type withthe smallest number of buttons possible, a trackball typesimilar to the widely-used PC pointing devices, and a voicerecognition type operated by vocal commands (see picture).
Research and development is underway on a receptionterminal that will provide data broadcasting and electronictextual data to visually-impaired persons or persons with
both visual and hearing impairments. Evaluation tests wereconducted on the accessibility and the operability of teletext,ISDB information, and six-finger Braille display systems.Improvements were made based on the test results.Additionally, with a view to future interactive services, westudied a remote communication procedure designed forpeople with both visual and hearing impairments. It wasverified that a newly developed conversation protocol canmake such remote communication possible. We alsoconstructed a prototype wearable six-finger Braille terminal,and confirmed the basic characteristics of its input andoutput functions.
To make broadcasting enjoyable for elderly viewers whofind TV speech too rapid, we developed a speech rateconversion system. This technology can slow down speechto a speed that is easier to understand. In fiscal 1999, wedeveloped this technology into a software application, andconfirmed that it is capable of outputting vocal data througha real-time speech conversion process on a PC. We alsotest-manufactured a non-linear editing device with avariable-speed reproduction function. This allows speech tobe synchronized with the picture to contribute to efficientbroadcast scene editing work. We also investigated a soundsignal processing method which can maintain theintelligibility of output speech at up to 5 times the normalspeed.
Next Generation Human Interface2.2.3
20
Speech recognition type
Trackball type Button type Conventional type
Prototype remote controls for digital television
With a view to applications such as automatic videoindexing and editing support, research has been pursuedcentering on automatic face recognition.
We made improvements to a prototype recognition system(see figure) which detects, tracks and recognizes people'sfaces in video material. The system is able to identify of theorder of tens of individuals, a recognition accuracy sufficientfor practical use. Improved facial feature selection hasprovided increased robustness to dilations and rotations ofthe image caused by three-dimensional movement. We alsostudied a facial pose angle estimation method which may berequired in order to automate the process of databaseregistration.
In order to examine the use of face recognition to supportvideo editing, we constructed an index of face recognitionresults from a video clip, and test-manufactured a graphicaluser interface (GUI) for a video retrieval system. The GUIaccepts search keys such as a person's name or thecomposition of the scene, and the system retrievesappropriate video sequences by accessing the index.
A feature extraction technique that makes use of colorinformation, composition, and background information wasinvestigated in an attempt to create a flexible video retrievalsystem with human-like capabilities. Working with the ATRHuman Information Processing Research Laboratories, weinvestigated the extraction of optimal color statistics for usein the flexible retrieval system. We also proposed an image
retrieval technique which uses the image composition andcomplexity of the background as keys for recognition.Retrieval tests were conducted using still images of varioustypes.
Concerning possible effects on viewers' health from visualeffects such as are used in some animations, researchcooperation continued with external research institutionssuch as the Tokyo Women's Medical University and theMedical School of Gifu University. Surveys were alsoconducted concerning research trends.
Video
Recognition result
Face recognition system
Face image database
Outline of identity recognition system
Efficient Video Retrieval Based on Image Recognition2.2.4
In order to clarify the basic structural elements for user-friendly and intelligible EPGs (Electronic Program Guide), westudied TV programming categorical structure hierarchies.In the test, we used a top-to-bottom method of hierarchy forcompulsory classification experiments, and a bottom-to-toptechnique on subjective assessment regarding listedprogram name similarity. The experimental subjects were 24
persons in their 20s and up, and 295 NHK programs wereused. The average number of genres using compulsoryclassifications was 7.8. Eighty percent of the subjects used10 categories to classify the programs. When the similarityof the program names was analyzed with a cluster analysismethod, the results indicated that the most popular genreson EPGs (Electronic Program Guide)should totalapproximately 10.
Additionally, we investigated the degree that a person'sunderstanding is affected by the asynchronous presentationof multiple stimuli. Short term memory behavior was usedas an index. Especially examined were the results whenusing video and sound, or text and sound, since these arethe presentation methods used with TV. In the experiment,we presented a word in the order of sound and text, or textand sound, to test whether a subject would recognize if itwas the same word or not. The results are shown in theinserted figures, indicating that remarkably more accuraterecognition was observed in the case of the sound and textorder, and 80% and higher recognition accuracy wasobtained up to a delay of 2 seconds. Based on these results,we further enhanced our knowledge regarding anintelligible presentation of information involving interfacedevice display delays using both a "text" image and "sound"information.
Influence of text/sound data stimulus onset asynchrony on a subject's recognition
NHK STRL ANNUAL REPORT 1999 | Studies for Improving Conventional Broadcasting Services 21
22 NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services
Regarding studies for opening up future broadcasting services following the diffusion of Hi-Vision (HDTV) and digital broadcasting, work has proceeded on advanced audio and videocoding, super reality audio-visual systems, research and development of frequency resources,and other materials and devices.
3.1 Advanced Audio and Video CodingIn the area of high efficiency object-based video coding with flexible image manipulation, we conducted research on
basic technologies related to object shape coding, scene description, and object extraction. We also studied adistributed video database that retrieves video data containing desired scenes from broadcast station networkdatabases. Regarding digital video and audio watermarking techniques designed for program copyright protection,research continued on a technique that produces watermarking with robustness against signal processing. Regarding3-D video and audio coding, research proceeded on elementary technologies.
3.2 Super Reality Audio-Visual SystemsIntelation to 3-D television systems, research was carried out on eye strain factors related to binocular type 3-D
HDTV, and on integral photography (IP) capable of producing multi-viewpoint autostereoscopic effects observablewithout special glasses. We constructed a prototype 3-D camera that can determine the distance from the subjectwhile shooting a regular TV image. For the development of an advanced image sensing system, we focused ourstudies on visual and psychological effects on wide-screen system and ultrahigh-definition cameras with aresolution of 2,000 scanning lines. We also constructed a virtual reality (VR) 3-D audio system that reproducesappropriate sound fields that correspond to positions within a virtual CG studio.
3.3 Research and Development of Frequency ResourcesWe have been studying an advanced satellite broadcasting system that reduces service interruptions due to rain
attenuation. Emphasis was placed on research related to the satellite's phased array antenna that adjusts beampatterns according to rain conditions in each area being served. Additionally, we constructed a prototype multi-beamreceiving flat antenna that can also receive TV programs from broadcasting satellites in different orbits using a singleonboard flat antenna. We have been studying future millimeter-wave broadcasting systems suitable for large-capacityshort-range transmission and dense wavelength division multiplexing (DWDM) technology, which is a basic technology forthe construction of large-capacity optical networks within broadcasting stations.
3.4 Recording and MechatronicsTo achieve ultrahigh-density recording, we conducted a wide range of studies involving disk mediums and heads for
perpendicular recording, ultrahigh-density and high-speed optical recording, magneto resistive (MR) heads, and anultra-thin magnetic layer tape for large-capacity tape storage systems. As part of our micro-mechatronics research, wemanufactured a prototype actuator to realize a head with precise tracking characteristics.
3.5 Imaging DevicesResearch continued on improving the photoconductive efficiency of HARP film for use in a new Super HARP
HDTV camera. Aiming towards the realization of ultrahigh-definition and ultrahigh-sensitivity image pickupdevices, we proceeded with studies of a field emitter image sensor with a HARP target and a solid-state HARPimaging device.
3.6 Display and Optical DevicesAs for research on high performance plasma display to forward the diffusion of PDPs, we continued studies for the
improvement of luminous efficiency. We examined electroluminescent (EL) devices in anticipation of their utilization inthe development of next generation, paper-thin, flexible flat displays, in addition to liquid crystal devices advantageous forlarge displays with low power consumption.
Studies for Opening up FutureBroadcasting Services
Research Results 3
OUTLINE
We have made further research advances regarding highefficiency and operable object-based coding in terms of objectshape coding, scene description and object extraction.
For shape coding of arbitrarily shaped objects, weproposed a new coding method utilizing Wavelettransformation, and verified that this technique is superiorto the MPEG-4 method with respect to the coding efficiencyand manipulation capability (enlargement, reduction, andrevolution of objects).
For scene description that is used to synthesize visualobjects, we clarified the broadcasters' functionalrequirements. To meet these needs, we proposed expandingthe capability of BIFS (BInary Format for Scenes) of theMPEG-4 to specify each objects' reproduction time. As anapplication of the scene description, we proposed an editinginformation descriptor capable of expanding video effects,including the wiping of arbitrary shapes.
We contributed to MPEG-4 to produce studio profile(studio use) standards.
In regard to object extraction methods, improvements weremade to an object extraction algorithm so that the objectboundary moved smoother. This was accomplished byutilizing a background image subtraction method thatperforms such extractions based on the supposition thatbackground images do not change (inserted pictures). As forobject extraction based on motion estimation, we improvedthe motion estimation accuracy of the gradient-based method
with an eigenvalue algorithm, by introducing robustestimation techniques and eigenvector interpolationtechniques. We also continued studies on motion inferring ofa background image using high-precision camera operationparameters. In addition to the panning and tilting parametersutilized during the previous year, we have now made possiblethe use of a camera's zooming parameters, and accomplishedthe extraction of a three-dimensional background image by 3-D information obtained from a camera dolly.
To achieve general-purpose object extraction, we alsoproposed a knowledge-based algorithm that selects asuitable algorithm, from among multiple extractionalgorithms, after learning the relationship between thefeature characteristics of an image and its extraction result.In fiscal 1999, we developed and confirmed the effectivenessof such a knowledge-based algorithm that utilizes codingefficiency as an evaluation standard.
Object extraction by background image subtraction methods
Object-Based Video Coding3.1.1
Presently, each broadcaster uses a variety of codingformats for contents storage, transmission, and editing. If aprogram production environment could be constructed inwhich program producers did not have to be conscious ofthe formats they are dealing with, an improvement could beexpected in contents operability and interoperability,enabling more efficient program production andtransmission. We have continued research on such anenvironment described in terms of a network, its connectedequipment (module objects), and contents (media objects).Without a knowledge of each coding format, the Description
Method can utilize these respective objects simply byconnecting them in a logical manner.
In fiscal 1999, we specified the API (Application ProgramInterface) for connection between each object and the datainput-and-output API for constructing a functionalverification system (inserted figure) that arranges simulatedvideo players, monitors, etc. on a network. Advancementwas also made on the development of a module selectionalgorithm and broker module object for object management.This object is necessary for the realization of desiredfunctions through the combination of multiple modules. All
the equipment on thenetwork is implementedusing the common APIthat this researchproposes. On the premiseof this verification system,an environment wasrealized in which videodata could be reproducedwith the automaticconnection of necessaryequipment according tothe contents in use,without the need forspecial operations.
Generic Description for Contents Exchanges3.1.2
Contents B (format B)
Contents A (format A)
Decoder B
Monitor
Decoder A
Editor
Network
Player A
Player B
Module object
Media objects
Each device's (module object) protocol stack
Unique function of the device(reproducer, decoder,---)
Generic description (middle-ware)
ATMCORBA
TCP/IP
Composition of a functional verification system
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 23
Original image Resultant image
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services
In preparation for thestart of digital satellitebroadcasting in December2000, we are proceedingwith studies on a digitalwatermarking techniquethat embeds copyrightinformation into videoand audio data in aformat undetectable toviewers for the protectionof intellectual propertyrights. In fiscal 1999,examinations were madein regard to copyright information contents and digitalwatermarking technique requirements.
In regard to the digital watermarking of video data, wedeveloped a watermarking method for baseband HDTVvideo signals in line with system requirements. Theseadopted embedding and detection methods do not utilizethe original images for detection, and system secrecy wasalso considered in order to prevent easy tampering with thedigital watermarking. While striving to improve techniquesthat prevent picture quality deterioration due to theembedded digital watermarking, we are also studyingrobustness of watermarkings and evaluation methods forother processing techniques such as MPEG-2 (MP@HL)
encoding and decoding, image contraction, and partialcropping.
Additional new developments include audio digitalwatermarking embedding and detection methods that donot utilize the original audio material. Hearing testsinvolving various sound source signals were conducted onthis new method, and it was confirmed that the humanauditory system can not distinguish between the originalmaterial and sound data with embedded watermarking. Weverified that this embedded sound data has satisfactoryrobustness for signal processing including MPEG-2 AACcoding/decoding, noise addition-equalizing, and the partialcropping of audio data.
Original video/Original sound
Embedding device
Detection device
Watermark embeddingVideo/sound
Audience
Illegal copy production
Copyright NHK
Copyright NHK
Broadcasting
Digital watermarking for broadcasting
Digital Watermarking Technique for Copyright Protection3.1.3
24
In fiscal 1998, we proposed an object-based system foruse as a video/audio database for broadcasters. This systemis designed to realize the retrieval of data distributed on anetwork based on program contents such as "a goal scenein a soccer game." During fiscal 1999, research was focusedon input signal processing and video retrieval techniques fordatabase construction.
For data input processing, color and texture informationwas utilized since it is relatively easy to obtain at the time ofprogram production. As a result, it was found concerningsports programs such as soccer games that a player can be
extracted without being affected by the background images.Research was also advanced on video retrieval based onmotion specifications such as "a goalkeeper diving to theright side of the goal." For this type of retrieval, we proposeda new method that consists of two stages. The first classifiesmotion using the degree of stochastic motion characteristicssuch as centroid movement values. The second is a retrievaltechnique based on movement intervals (the time period ofan action performed in a TV video). The effectiveness of themethod was confirmed.
Research also progressed on retrieval technology thatmakes good use of multimedia's audio andtextual characteristics in addition to its videoretrieval functions. This technology utilizesaudio signals for video retrieval. Audio signalsare analyzed in fixed time intervals and videodata are automatically classified intoprogram categories such as speech, music,etc on the basis of the degree of stochasticcharacteristics.
We also devised verification systemspecifications for the proposed dispersedvideo/audio database (inserted figure). Theeffectiveness of the proposed system will beverified using the prototype database system.
100Base-TX LAN
Database serverUnix
Video dataMeta-data
Workstation ATM622M
Application serverUnixCORBA,Java,C++
Workstation ATM622M
Client PCWindows NT4.0CORBA,Java,JMF2.0
Personal Computer
622MbpsATM (Asynchronous Transfer Mode)
JDBC ORB ORB
Retrieval command
Retrieval result (video data)Video dataMeta-data
JMK (Java Media Function)ORB (Object Request Broker)
Aiming to clarify coding efficiency requirements for entire3-D audio and video systems, we continued to promote thedevelopment of element technologies for capturing, coding,and display for 3-D systems.
Regarding capturing systems, we researched the statisticalcharacteristics of 3-D HDTV standard still pictures, andconstructed a 3-D video model based on the research results.Using this model, optimum camera arrangements wereestablished for 3 or 4 viewpoints. As for matchingcorresponding pixels to acquire depth information, wedefined a new matching cost (correlative function and itskurtosis) of between 3 and 4 viewpoint images and showed
that the introduction of matching cost distribution improvesdepth data accuracy. Using this depth information and thevideo, 3-D video images from any viewpoint were reproducedin order to evaluate picture quality. High quality pictures wereobtained from the experiment (see inserted figure).
Multi-viewpoint 3-D video coding research led toexaminations of scalable coding for multi-viewpoint images.Fundamental estimations, such as the relationship betweenthe number of viewpoints (number of cameras) and depthinformation estimation error, were investigated.
For the display studies, we proposed a method to smoothout depth data by introducing an evaluation function wedeveloped in order to interpolate a video from any viewpoint.Improvement of the interpolated image quality was confirmed.
As for 3-D sound presentations, using an array-speakersystem for subjective assessment tests, we verified adifference limen for distance (near-to-far) and direction (left-to-right) between two sound images synthesized in front ofa listener. Also verified was the noticeable ratio of sourceseparation (energy ratio between two sound source), a valueneeded to secure good separation quality (front-to-backdirections). Moreover, subjective assessment tests werecarried out on the ratio of direct to indirect sound energy(energy ratio of direct sound and indirect/reflected sound)with the aim of establishing a parameter for the sense ofdistance regarding sound images, and these confirmed theparameters' effectiveness.
Depth information acquired by 4-viewpoint cameras
220cm
207cmCenter viewpoint image
Distance from a camera
Reproduction image when a viewpoint is moved to the right
Reproduction image when a viewpoint is moved to the left
Reproduced images from any viewpoint using depthinformation acquired by 4-viewpoint cameras
Coding for 3-D Audio and Video Broadcasting Systems3.1.6
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 25
During fiscal 1999, we constructed a prototype networkmodel based on the protocol stack examined during theprevious year. Using this model, we investigated thepossibility and the technical problems associated withinteroperability and Plug and Play functions, as well as videodata stream transmission protocols on the middle layer ofbroadcasting networks. For the verification of technicalproblems (inserted figure), we also established a videostream server and terminals on Gigabit Ethernet equippedwith a middle-ware CORBA (Common Object RequestBroker Architecture): a common specification for distributed
object technology standardized by the nonprofitorganization OMG (Object Management Group).
As part of the surveys taken on the actual condition ofbroadcasting station network use, we collected video editorlog data at the NHK Broadcasting Center (in Shibuya). Byanalyzing these records, we obtained basic data needed forvideo editor traffic estimation.
Computer Network System for Broadcasting Stations3.1.5
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services
Stereoscopic HDTV, or 3-D Hi-Vision is stereoscopictelevision based on binocular disparity. It provides extremelyrealistic 3-D pictures with high technical realizationpotential by taking full advantage of the high resolution andwide-screen features of HDTV. In order to utilize this systemas a broadcasting medium that delivers services for manyunspecified people, the construction of a system thatproduces natural, easy to view images with less eye strainfactors is necessary. To achieve these goals, we have beenconducting a comprehensive research on 3-D shooting,display systems, and human vision characteristics.
One of the major factors of stereoscopic picture qualitydeterioration caused by a display device is cross talk (doubleimage disturbances induced by incomplete separation of theright and left image). We conducted experiments on crosstalk to determine the detection limits using stereoscopicimages (geometric test patterns). The experiments revealedthat the cross talk visibility threshold, although dependingon the degree of contrast, was less than 0.2-0.3%. Subjectiveevaluation tests using ordinary sample images were carriedout and fundamental data were obtained. We furthercompared the disturbance effects caused by the cross talkand similar 2-D image ghost disturbances, and verified that3-D cross talk disturbances tend to exceed those of 2-Dimages.
Using a newly developed 3-D HDTV camera system(inserted figure), subjective evaluation tests and factor
analysis were performed on 3-D HDTV videos producedunder various shooting and viewing conditions. As a result,two "sensation of reality" and "ease of viewing" factors wereextracted. Taking these factors into consideration, theanalysis of a comparative survey between 2-D and 3-Dimages revealed that 3-D images are superior to 2-D imageswith respect to the sensation of reality. However, factorscores for 3-D picture ease of viewing varied greatlydepending on the image patterns.
New 3-D HDTV camera system (developed in fiscal 1998) usedfor the production of evaluation image materials
3-D Hi-Vision (HDTV) System3.2.1
Basic researches on multi-viewpoint, integral photography(IP), and holographic image acquisition techniques areunderway to realize an advanced 3-D TV system that canproduce 3-D pictures that can be viewed without specialglasses from different viewpoints.
A new concept was adopted for this multi-viewpointsystem, "multi-viewpoint sampling of 3-D real space." As partof this concept, we numerically expressed 3-D imagecharacteristics and also established a method to determineoptimal design parameters for image acquisition, such as
multi-viewpoint camera pitch and convergence distance.Although the shooting of multi-viewpoint 3-D images normallyrequires multiple cameras that are equivalent in electrical andgeometrical characteristics, we designed a field-sequentialstereoscopic camera which can singly shoot 3-D images. Thistheoretical system was trial manufactured and its performanceverified.
With regard to integral photography (IP), we confirmed thatthe system provides real-time multi-viewpoint 3-D images thatcorrespond to different views without the need of special
glasses. It was possible to transfer a viewpoint in avertical or lateral direction and furthermore to watchthe 3-D system from a lying position.
Basic holographic research continued to realize amoving holographic picture using a liquid crystaldisplay (LCD) panel. The pixel interval ofconventional LCD panels is too wide for holographicdisplays, limiting the area in which a three-dimensional reconstructed image can be observed(viewing zone), and making it difficult to perceivethe images with both eyes as autostereoscopicimages. To solve these problems, we verified thatdiffraction beams, originally removed due tointerference with reconstructed images, can actuallybe used as object beams to enlarge the viewingzone.
Future 3-D Television System3.2.2
26
Right view of 3-D image
Left view of 3-D image
Top view of 3-D image
Bottom view of 3-D image
Shooting
Image subject
Integral photography (IP) display exampleIt can display 3-D images corresponding to different viewpoints.
3.2.3
3.2.4
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 27
Research is being promoted on 3D camerasand displays for future 3D TV that providenatural images with little eye strain.Previously, there had been no means toobtain depth information of objects at videorate. However, during fiscal 1998, wedeveloped a 3D camera (Axi-vision camera)equipped with infrared illumination and anultra-fast speed shutter to detect the pixel-by-pixel depth information of objects at a speedof 15 frames per second.
In fiscal 1999, with the aim of improvingresolution and expanding the shooting areaup to the size of a person, we developed anew bright high-speed modulation LED arrayillumination device. As a result, speckle noiseassociated with laser diodes was eliminated, and theshooting angle was widened. Noise caused by insufficientinfrared illumination was also reduced and the depthresolution was improved to 1.8 cm.
Furthermore, we analyzed the noise characteristics thatdetermine the resolution, and clarified the relationshipbetween measured depth, illumination intensity and noiselevel. Based on these results, we proposed a method foradjusting noise reduction schemes of a spatial filteraccording to the intensity of the reflected light. Whenapplied to depth-key image extraction, the method was able
to reduce the fluctuations in the extraction boundary.We also constructed a prototype 3D display system of the
depth sampling method by combining eight CRTs and halfmirrors. We confirmed that real-time 3D image pickup anddisplay was possible with 8-stage sampling of the depthinformation from the 3D camera.
Modulated infrared illuminator
Distance image
Color image
Color camera
Modulated illumination light
Noise filterCCD
Signal processor
Image intensifier
Diagram of the 3D camera (Axi-vision camera)
Depth Information Capturing and 3D Displays
Visual and Psychological Effects onWide-screen System
"Potency" and "Comfortableness", whichindicate wide-screen video psychologicaleffects, were investigated with regard toscreen size, viewing angle, and materialsfor flat and concave screens.
On the flat screens, still and panningimages were evaluated using an ultra high-definition slide projector. The results ofevaluations show that the degree of"Potency" on the flat screens increases inproportion to viewing angle, but that the degree of"Comfortableness" decreases in the case of extremely largeviewing angles (inserted figures). It was also confirmed that "themost preferred viewing position" is where both reach amaximum. On the concave screens, evaluations wereconducted at a fixed viewing distance, but with imagepresentation angles up to 180 in the horizontal direction. Theresults showed that as the viewing presentation angleincreased, so did the "Potency" and the induced effects inviewers' body sway corresponding to the screen image.
We also evaluated the relationship between the influenceof surrounding images and the preferred viewing distancefor a wide-triple-screen system using still pictures withvarious surrounding angles. The results show that the mostpreferred viewing distance tends to decrease when there is
influence from surrounding images.
Visual Wide-field Display For the development of an ultra high-definition display,
we started developing an experimental multi-screen displayproduced by combining four LC projectors seamlessly.
We also reduced the cross-talk between right and leftimages on the PDP stereoscopic display systems, by improvingthe cross-talk canceler, and confirmed the improvement of thevideo image quality. Additionally, signal processing thatcancels PDP phosphor persistence was examined, resulting inan improvement of the video picture quality.
The simulations of the field-sequential-color displaymethod using a special CRT showed the need to improvepicture quality for motion pictures.
Ultra high-definition and wide-screen Display System
(a) Degree of "Potency" evaluation (b) Degree of "Comfortableness" evaluation
Pote
ncy
degr
ee
Com
fort
able
ness
deg
ree
3
2
1
0
-1
-2
-3
30 40 50 60 70 80 90 100
3
2
1
0
-1
-2
-3
30 40 50 60 70 80 90 100
300cm240cm180cm120cm
Screen width
Horizontal angle ( )
Horizontal angle ( )
Examples of evaluation results on degree of "Potency" and degree of "Comfortableness" (image "building," shooting angle of 66 )
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services
3.2.5
28
We have been conducting the sound-field control studies for next-generationsuper reality audio system. This systemreconstruct sound field using acousticalroom impulse responses based on theroom's shape. A computer simulationmethod was researched in order toestimate the acoustical room impulseresponses accurately. Furthermore, inorder to create the "surround sound"feeling with as small a system aspossible, we examined the loudspeakerlayout and the number of necessaryreproduction channels.
As an application of acoustical roomimpulse responses in sound fieldreconstruction, an experimental 3-Daudio system was manufactured thatuses VRML (Virtual Reality Modeling Language) and allowsfor real-time sound field changes (inserted figure). Thissystem allows a user to enjoy interactive sound fields thatreflect the locations linked with CG room images as the usermoves around a room. This system reconstructs the soundfield at the listening location by constantly adjustingparameters concerning reverberation time and the locationalrelationship between the sound source and listening point.
We continued our research on surround loudspeakers for
the 3-D audio system carried over from fiscal 1998, toimprove the "loudspeaker-wall" in regard to size, striving fora more compact and narrower system. We developed asmall prototype "sound pillar" loudspeaker aiming for itspracticality. The sound pillar is structured as a columnconsisting of an upper main loudspeaker array and a lowersupplementary loudspeaker array. A wide listening rangecan be obtained with a small number of Sound Pillars bycontrolling their directivity and level.
Sound Field Control
Signal flow
Control flow
Audio serverVirtual sound source head search control
ControllerLAN
Space data server
Image production and operation terminal
CG image
Mouse operation
8 channel signal
3-D audio signal generatorRoom shape and
sound source location VRML dataReflected sound data (room impulse response)
Dry sources
Experimental virtual reality 3-D audio system
Ultra High-definition Image Acquisition TechnologyStudies are underway concerning an ultrahigh-definition
image acquisition device that exceeds the performance ofHi-Vision (HDTV). Such a device is essential to future superreality broadcasting. Applying the 8-million-pixel CCDimager developed in fiscal 1998, we manufactured anexperimental monochrome ultrahigh-definition camera(inserted figure). We obtained the theoretical value of theexperimental imaging element which consists of a limitingresolution of 2000 TV lines horizontally and vertically. Thiselement forms a divided-area parallel-processing systemstructure for huge-pixel-number image acquisition. Acharacteristic channel signal multiformity results in picturequality deterioration in this system, so we started examiningvarious methods to resolve this problem. One of thepreliminary experiments regarding the system's programproduction application was the trial manufacture of a device
which converts a 2070 (vertical) 3840 (horizontal) pixelultrahigh-definition image to HDTV or NTSC format byreducing or clipping the original image. Another exampleprogram production application was demonstrated using avideo effect technique that synthesizes a reduced completeimage with an optional enlarged partial image.
Subject-adaptive Pick-up TechnologyIn order to effectively improve the theoretical limits of
ultrahigh-definition and high-speed imaging(such asresolution, sensitivity, S/N, dynamic-range), we continuedresearch on a subject-adaptive pick-up method thatautomatically optimizes pick-up characteristics for eachsubject. First of all, an evaluation was made of the 1998prototype wide dynamic-range element that countssaturation frequency on every pixel and its basic operationalperformance was confirmed. We also started new research
on an image sensor thatdigitizes each pixel signaland investigated a newCCD image sensorcomposition capable offreely transferring variouscharacteristics amongspace resolution, timeresolution, S/N, sensitivity,and so on.
Digital signal processing1 Pixel arrange inversion Image division
Digital signal processing2 Edge enhancement Image reintegration
Analog signal processing DC control Gain control CDS
8-million-pixel CCD
16 ADCs
8 outputs
8 outputs
DAC
DAC2kx2k
2kx2k
SDIx24
Monitor
Experimental ultrahigh-definition image pick-up system blocks
Advanced Image Acquisition System3.2.6
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 29
Satellite broadcasting services in the 12GHz band arecarried out in compliance with the Broadcasting-satelliteService Plan, which is described in the Radio Regulations(RR). A resolution requesting the distribution of morechannels for each country was adopted at the WorldRadiocommunication Conference (WRC) in 1997. Followingthis resolution, the ITU-R (International TelecommunicationUnion-Radiocommunication Sector) established the IRG(Inter-conference Representative Group) and the GTE (aGroup of Technical Experts) to carry out technicalinvestigations in preparation for replanning proceduresscheduled to take place at a WRC meeting to be held in May2000.
Our laboratories, in cooperation with related NHKdepartments, drafted a new Plan based on the current one,which will give each Region 3 (Asia/Oceania) country a 12
channel assignment (four additional channels for Japan)based on interference calculation on each satellite system.The new draft Plan was presented to an expert groupmeeting of the AGTE of APT (Asia PacificTelecommunication) which finalized discussions regardingthe Region 3 Plan. Further points were presented by Region3 experts, and an APT draft Plan was completed which wasthe consensus of the Region 3 countries, including Japan.This Plan was proposed to the GTE and IRG. Specialists fromour laboratories also participated in the GTE and IRG, andproceeded with discussions involving experts from Region 1(Europe/Arabia/Africa), who used the APT draft Plan as abase Plan during the ITU-R replanning process. Channelassignments for Japan are shown in the figure below.
New channels under discussionCurrently assigned channels
Right-hand circular polarized wave
38.36MHz
1 3 5 7 9 11 13 15 17 19 21 23
APT draft Plan of channel assignments for Japan
Revision of the Broadcasting-satellite Service Plan3.3.1
We constructed an experimental multi-beam receiving flatantenna (inserted picture) for satellite broadcasting. Withthis model we conducted an experiment on thesimultaneous reception of signals transmitted from twodirections. This confirmed that the system can detectseparate signals, obtaining isolation higher than 26 dB fromtwo different beams 40 degrees apart (the maximum anglebetween the present broadcasting satellites observed fromthe ground).
The efficiency of a flat antenna is for the most partdetermined by the feeder loss between an antenna elementand a first low noise amplifier. An effective method toreduce this feeder loss, and to further simplify the structurefor future miniaturization, is to integrate the low noiseamplifiers with antenna elements. We tried this with a newprototype antenna using 8 4 elements and a low noiseamplifier (HEMT : high electron mobility transistor), andconfirmed that it will perform as desired.
With the aim of making possible the reception ofbroadcasts with different polarization, basic studies toreceive arbitrary polarization waves were performed. Weconstructed an experimental polarization matching circuitwhich only selects the desired linear polarization byelectronically controlling the signals received by dual-fedsquare microstrip patch antenna elements. As a result, itwas observed that the antenna could maintain receivingcharacteristics of 20dB and greater cross polarization
discrimination over a wide bandwidth.
Multi-beam Receiving Flat Antenna for Satellite Broadcasting3.3.2
Prototype multi-beam receiving flatantenna for satellite broadcasting
Advanced Satellite Broadcasting SystemResearch on an advanced satellite broadcasting
system has been continuing since the 1998 fiscal year.This system will make it possible to reduce theinterruption of broadcasting service due to rainattenuation. The key component of the system is thephased array antenna used for the satellite. This systemis designed to adjust how the beam is formed based onthe weather conditions in each area, therebyconcentrating the beam pattern on the area of heavyrainfall to compensate for rain attenuation (Figure 1).
In order to better understand the characteristics of rainattenuation, we have studied the probability of excessiveprecipitation in Japan to a certain level and the related rainattenuation estimates. As a result of a thoroughinvestigation of the 10-minute-rainfall-ranking data from40 nationwide observation points during the past 58 years(Japan Meteorological Agency), it has been determined thata period of precipitation, defined as a 10-minute rainfall of morethan 20mm, occurs at one or more of the observation pointsevery year. Further, using the precipitation data from a groundstation in Okinawa that measures rain attenuation by thecurrent DBS down-link(established in fiscal 1996), an analysis ofthe distribution of the peak one-minute rainfall in a 10-minuteperiod was conducted, and showed that the distribution can beapproximated as being lognormal. These results have made itpossible to predict the degree of general rain attenuation basedon the AMeDAS (Automated Meteorological Data AcquisitionSystem) 10-minute rainfall data from 1,300 nationwide points. Itis estimated that the peak amount of rain attenuation in theTokyo area for a 20mm 10-minute rainfall will be approximately35dB for the 12GHz microwave band.
Moreover, we approximated the satellite transmission powernecessary to compensate for rain attenuation by referring to theAmeDAS 10-minute rainfall data gathered for approximatelyone year. This result showed that using a 100km-diameterradiated beam for compensation in the 12GHz-band, most rainattenuation can be compensated for with a total satellitetransmission power of 1kW.
In researching the use of onboard phased array antennas forsatellites, we conducted basic radiation pattern characteristicstudies of an array antenna configuration which combined aTWT feed array with a reflector.
The possibility exists that this broadcasting satellite systemmay cause interference outside of its service area, near Tsushima
islands, for example, if the EIRP (Equivalent IsotropicallyRadiated Power) increases at the fringe of the service area. Forthis reason, an electric field distribution on an equivalentaperture, which we supposed, was calculated using theallowable limits of PFD (Power Flux Density) outside a servicearea as determined by the Radio Regulations (RR). We studiedthe maximum values to be applied to rain attenuationcompensation using six types of window functions as an electricfield distribution for the equivalent aperture. This showed that byselecting an appropriate window function to be used in an areawith past experience of heavy rain, it may have possible tocompensate for rain attenuation even at the fringe of a servicearea without exceeding the PFD limits outside it.
The TWT for a phased array antenna that performs precisebeam control requires an antenna element interval of less thanone wavelength. For this, an experimental thin-type TWT(25mm diameter) with a compact electron gun and collectorsection was manufactured experimentally.
21 GHz-band Satellite BroadcastingResearch on the super-long-period interleave transmission
technique was continued. Severe rain attenuation of greaterthan 100dB is observed in a high frequency environment (abandwidth of 21GHz or higher). As for the aforementionedcountermeasure, even EIRP amplification using a phasedarray antenna is not enough to deal with such severe rainattenuation. This creates the need to develop a more
advanced transmission scheme withthe possibility of using storage-typereception.
We have continued an investigationof the following two new transmissionschemes: 1) a scheme whichinterleaves the transmission programsfor a very long time-span (i.e. 24-hour)and uses a receiver to accumulate theinterleaved data and properly decodeit (Figure 2) ; and 2) a hierarchicalinterleave scheme which changes theinterleave period on each program bymultiplexing the transmission controlsignals.
Nah
a
Fuku
oka
Hiro
shim
a
Osa
ka
Nag
oya
Toky
o
Send
ai
Sapp
oro
70
60
50
EIRP
(dBW)
TWT array
Large reflector
Figure 1: Advanced satellite broadcasting rain attenuationcompensation system
Future Satellite Broadcasting System3.3.3
Live broadcast (news, etc.)
Storage-type reception
(drama, etc.)
Transmission data Interleave
24 hours
Interruption
Time
Transmission signal
Reception data
Error correction
(Super-long-period interleave transmission)
12GHz 21GHz,42GHz
Rain attenuation
Example :
Figure 2: Outline of super-long-period interleave transmission
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services30
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 31
Millimeter-waves have propagation characteristics suitablefor large capacity short-range transmission. This enablesindividual viewers to enjoy interactive services, communityservices, and mobile communications. We have been studyinga future millimeter-wave broadcasting system since 1997. Infiscal 1999, we analyzed millimeter-wave propagationcharacteristics in residential areas.
A computer simulation was executed to analyze coverage,namely the percentage of locations in which millimeter-wavecan be received, using a three-dimensional residential mapcontaining information on building shape and height. Weassumed multiple millimeter-wave transmitting stations andthen calculated their coverage. Coverage for household was
calculated as the feasible percentage of households withappropriate receiving locations on their walls or roofs.Coverage on the road was also computed for mobilereception. In the Kinuta area of Setagaya-ku, Tokyo (which is atypical residential area), for example, coverage for householdsand coverage on the road were approximately 95 % and 50 %,respectively, using 18 millimeter-wave transmitting stations(inserted figure). The number of the stations is the same asthat of PHS (Personal Handy-phone System) base stations inthe area. The coverage for households and on the road can beimproved by the reception of reflected waves.
We obtained an experimental license for a 21-GHz-bandradio station in December 1999. Using this station, weproceeded with experiments on non line-of-sight millimeter-wave propagation on the road.
Microwave/Millimeter-wave Transmission and Reception Technology3.3.4
0
0
20
40
60
80
100
10 2018
30
(%)
Co
vera
ge
Number of transmitting stations per 1km2
With reflected wave
Without reflected wave
Coverage for household
Coverage on the road
Example of calculated coverage
Noting that the permittivity of liquidcrystal changes greatly with externalbias voltage, we have been developinga microwave device which enables usto electronically control transmissioncharacteristics by using liquid crystal.In fiscal 1999, an experimentalmicrowave variable delay line wasconstructed. The delay line iscomposed of a microstrip line usingliquid crystal as a dielectricsubstrate(inserted figure). Theexperiment showed that the phaseshift at 20 GHz was 370 , and thedelay time change was 54 ps, in a casewhere the strip line length was 193mm. It also revealed a non-dispersivedelay characteristic over a wide rangeof microwave frequencies, indicating the possiblerealization of a significant wide-band microwave variabledelay line.
The insertion loss of the delay line was also studied. Theinsertion loss was approximately 20 dB at 20 GHz, mainlycaused by conductor loss. In order to reduce this conductorloss, the strip line width has to be widened, which also thencreates the need for a thicker liquid crystal. On the otherhand, as the thickness of the liquid crystal increases, the
response time of the liquid crystal alignment becomesslower. It was found that it may be possible to improve theresponse time by using a new type of liquid crystal, such aspolymer-dispersed liquid crystal or dual-frequency switchingmode liquid crystal.
Microwave Elements3.3.5
Strip line (length=193mm)
Bias vo
ltage so
urce
Alignment film
Ground planeMicrowave
Glass substrate
Liquid crystal 50 m
Initial alignment direction
Microwave variable delay line structure with liquid crystal
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services
Dense Wavelength Division Multiplexing (DWDM)We have promoted a dense wavelength division
multiplexing technology, which is the basic technology thatwill be used for future optical video routing networks inbroadcasting stations. With the aim of realizing high-densityand high-speed switching between channels, researchcentering around an optical filter has been promoted.
The examined optical filter is called a Fabry-Perot opticalwaveguide filter. Both sides of the optical waveguide arecoated with a reflective film, allowing it to retain a narrow-band transmission characteristic.
In order to reduce the voltage needed to switch channels,
we proceeded with electrode gap optimization andexaminations of both the electrode structure and theproccessing technique. This allowed us to more efficientlyimpress an electric field on the optical waveguide. Weconfirmed that application of the study results reduced theapplied voltage by 30%. We also developed a voltage controlmethod that makes the transmission characteristic stableagainst temperature changes, and devised a technique toreduce transmission loss at the time of voltage impression.
Additionally, an optical filter in which two Fabry-Perotoptical waveguide filters are cascaded was designed toeffectively repress unnecessary transmission peaks. Themaximum amount of multiplexed wavelengths under thisrepression was determined through numerical analysis. Theanalysis clarified that optimizing the structure will allowwavelength multiplexing of up to 128 channels.
Optical-fiber Transmission of a Millimeter-wave SignalWe advanced studies of an optical modulation scheme to
transmit a millimeter-wave signal through an optical fiber.In 1999, in order to transmit a millimeter-wave signalwithout influence from optical fiber chromatic dispersion,we designed an optical SSB (Single Sideband) with a carriermodulation scheme which represses a single side of thesideband. Using this method, an optical fiber transmissionexperiment on a millimeter-wave (13.6 GHz) signal wascarried out.
Optical Transmissions3.3.6
32
Wavelength multiplexedlights
A selected wave
Electrodes
Reflector films
Optical waveguide
Fabry-Perot optical waveguide filter
We have been developing a network-type home servercapable of being shared by multiple receivers. A networkinterface is based on IEEE1394 (400Mbps). The network-typehome server can record/playback four digital HDTVprograms. Each of the viewers can remotely operate thehome server through the home network from the receivers.In addition to conventional VCR operations, new functionswill be available such as catch-up viewing of programalready in progress or of a portion of a program missed by aviewer.
New multimedia services utilized by large capacity homestorage devices that use both broadcasting and the Internet
and the systems that enable these services are called "TVAnytime" (inserted figure). The standardization of "TVAnytime" had been promoted by the DAVIC (Digital Audio-Visual Council) until September of 1999 when this functionwas succeeded to the TV Anytime Forum. A system standardfor temporary practical use is scheduled for completion inFebruary 2001. NHK has proposed a score of technicalsuggestions based on its home server research results,including the concept of "anytime service," and the majorityof these proposals have been adopted as standards.
Domestically, in order to standardize next generationmultimedia services and support digital broadcasting
systems, ARIB established a WorkingGroup for Broadcasting Systems Based onHome Servers in October 1999. We areworking toward efficient contentscirculation accomplished by a broadcastingand communication fusion-type service,tackling technical problems on both thereceiving and transmission sides.
Future Home Server Technology3.4.1
Contents provider Service consumer
Broadcasting
Internet
Program exchange network Distribution network
Home network
Homeserver
Service provider
TV Anytime system
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 33
Research of High-capacity Tape Storage UnitTo achieve high-density recording, we have been
advancing the development of a highly sensitive MR(Magneto resistive) head for rotary drums and an efficienthigh-performance ultra-thin magnetic layer tape. Weoptimized design of the head by means of computersimulations and fabricated a prototype of the device.Evaluation of the read performance of this prototype MRhead revealed sensitivity twenty times that of conventionalinductive heads. We also test manufactured an ultra-thinmagnetic layer metal evaporated tape adjusted to thesensitivity characteristics of the MR head with a magneticlayer thickness of less than 0.1 m. The reproduction signalSN ratios achieved by combining the MR head and ultra-thinmagnetic layer tape indicated a possible area recordingdensity of 1Gbit/inch2 (1.55Mbit/mm2) or higher. Thisrecording density is 40 times higher than that achievedwhen using D-3VTR.
We performed video recording and reproductionexperiments utilizing a VCR equipped with the prototype MRhead, and realized stable video reproduction with a byteerror rate of 10-5 (inserted picture).
Development of Stable Reproduction VCRs for LongTerm Storage Tapes
For stable reproduction of deteriorated 1-inch videotapedue to long-term storage, we developed an ultrasonic
vibration drum which reduces friction between the tape andthe drum. To optimize drum vibration, a measurementsystem of modal analysis was constructed. As a result, wewere able to reduce friction by 30%.
Restoration and Improvement of JIS Standard Type IVCRs
We restored and improved a JIS standard type I VCR setenabling high-picture-quality reproduction of valuable videodata owned by individuals, data that is not stored inbroadcasters' archives.
Highly sensitive MR head recording/reproductionexperimental device
Tape Storage3.4.2
Research has continued on new recording materials for thedevelopment of optical disk for HDTV cameras, future programproduction devices, and ultrahigh-density/high-speed opticalrecording systems for archives. With the aim of high-speedrecording utilizing phase-change recording materials, weapplied the film stress method to the development of a high-speed crystallization method. This new process was evaluatedby means of an atomic force microscope, and guidelines fordesigning this new high-speed medium were obtained. Utilizingthis technology, we succeeded in preparing an instrument toproduce film for the manufacturing of optical disks. Along withother related experiments, evaluation of the system's opticalproperties and micro-recording marks began through use of ascanning near-field optical microscope. Research on high-speed
magnet-optical media is also being performed.Regarding high-density technology, we continued research
on a super-resolution film that reads out micro-marks at ahigh CN ratio. Such performance is difficult to reproduce withconventional optical systems. For this super-resolution film,we fabricated a transparent film doped metal and found that asuper-resolution phenomenon occurs when we perform anoptical evaluation using blue light (441nm wavelength).
In regard to high-speed recording/reproduction technology,research advances were made on multi-beam light sources andparallel readout. Although development of a light source withnarrow optical spot pitch is important to multi-beam recording,existing semiconductor lasers are limited in their capability ofnarrowing such spot pitch due to heat interference. Toovercome this obstacle, we studied a new multi-beam lightsource that uses optical waveguides and constructed an opticalwaveguide with a 0.45% relative index difference (insertedfigure). With a 780nm input wavelength, the device achieved awave-front aberration of less than /20 and an optical spotpitch of 11.5 m with a 0.5mW to 10mW output. This confirmedthat multi-beam light sources with optical waveguides areapplicable to optical recording. Research and development ofblue multi-beam optical sources was also started.
Aiming for future ultrahigh-speed readouts utilizing parallelreadout techniques, we proposed algorithms for high-speedtracking and signal reproduction using an intelligentphotodetector array.
Optical Recording Device3.4.3
Multi-beam light source with optical waveguides
Optic fiber Optical waveguide
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services34
Perpendicular magnetic recording heads and recordingmedia have been researched successively.
For our perpendicular magnetic disk research, we haveoptimized the microcrystalline structure of cobalt-chromiumalloy recording film. The carrier to noise ratio (C/N) has beenimproved by as much as 8 dB at a linear recording density of240 kFCI (recording bit-length of 0.1 m). Furthermore, we haveestimated the recorded profiles of media using simulationsoftware in order to attain high C/N at the ultra high density. As
a result, we foundthat the grain size ofa cobalt-chromiumalloy film should bereduced byapproximately 10 nmin order to acquirethe high recordingdensity of 100Gbit/inch2. It wasproved that anamorphous medium,for example terbium-iron-cobalt alloy, hasa higher recordingdensity limit fromthermal relaxationthan that of
conventional cobalt-chromium alloy media, enabling ultra highdensity recording of 350 Gbit/inch2 and higher.
Read/write characteristics of the hard disk systems havebeen evaluated. A byte error rate of below 10-7 was attainedat the recording density of 10 Gbit/inch2 by the side erasingof recorded bits and by optimizing read/write conditions.
We have estimated the relationship between the headconfigurations and output characteristics by computersimulation. It was found that for ultra high recording densityabove 100 Gbit/inch2, the S/N of the yoke-type MR head ishigher than that of conventional shielded MR heads.Furthermore, it was proved that the waveform distortionscould be reduced by the addition of a flux-guided shield tothe yoke-type MR head, as shown in the figure.
Soft magnetic thin films with a large saturation magnetic fluxdensity were required for recording head materials in ultra high-density applications. We could deposit excellent iron nitride thinfilms with a saturation magnetic flux density of 2.1T, coercivityof 0.5 Oe and relative permeability of 2620 at 10 MHz.
TMR (tunneling magnetoresistance) devices are stronglyrequired for future ultra-high density recording systems,such as read-out heads and solid-state magnetic memories.We were able to establish the key deposition technique ofthe stoichiometric tantalum pentoxide thin films with flatand sharp surface structures applicable for insulation layersin TMR devices.
Upper yokeLower yoke
MR element
Flux-guided shield
Magnetic flux from medium
Medium
Structure of a yoke-type MR head withflux-guided shield
Perpendicular Magnetic Recording Head and Medium3.4.4
Future high-density recording will require a head actuatorwith precise positioning capability to accurately track arecording head. For this, a second stage (piggyback)actuator at the tip of the primary actuator is needed toperform such precise positioning.
Our laboratories have test manufactured a gimbal-typeactuator using piezoelectric material meeting the requiredspecifications for the second stage actuator which are: a100mg onboard mass, a 3kHz or higher resonancefrequency, and a 1 m or higher displacement (insertedfigure). With this actuator, we obtained a 3.5kHz resonancefrequency and a 3 m displacement. The head actuator in amonolithic piezoelectric material has application potential
for future disk recording devices. Experimental manufactureand drive experiments on an attitude controllingpiezoelectric actuator are continuing with the aim oflooming control.
Additionally, an actuator is desirable to gain a greatdisplacement and to perform precise first stage positioning.Continuing progress is being made on the construction of aprototype laminated piezoelectric actuator to realize suchcharacteristics.
Also promoted was the manufacture of precise measuringjig devices and related experiments for the development of anew disk clamp mechanisms that adjust to disk eccentricityand surface deflection.
Micro-machining and Precision Measurement3.4.5
Recording track
Slider (head on board)
First stage tracking (rough action)
Rough action pivot
Second stage tracking (fine action)
Disk
Positioning actuator for disk device
Piezoelectric material
Research has continued on improving the photodetectionefficiency of the ultrahigh sensitivity photoconductive film(HARP: High-gain Avalanche Rushing amorphousPhotoconductor) developed by our laboratories with the aimof further improving picture quality. In fiscal 1999, we
elevated theconversion efficiencyof green light, up toapproximately twicethat of theconventional method,by optimizing thephotosensitive film( t e l l u r i u m - d o p e dfilm) structure. Weapplied this newHARP film to a colorcamera's greenchannel pickup tubeand evaluated itsperformance. Theresults confirmed thatthe signal to shotnoise ratio wasimproved byapproximately 3dB,and more vivid color
imaging was made feasible, even for dark subjects. This newphotosensitive film will be applied to a 2/3 inch New SuperHARP image pick-up tube (electromagnetic focusing andelectromagnetic deflection type) for Hi-Vision (HDTV).
Amorphous silicon photoconductive films have beeninvestigated toward the development of next-generationavalanche multiplication photoconductive film. We have foundthat the film's avalanche multiplication phenomena occursonly in a device on a silicon substrate with certain impurityconcentrations. During fiscal 1999, optimum use was made ofa device simulation method to analyze the multiplicationphenomenon. This enabled us to explain the characteristics ofthe test device, giving consideration to localized states in theamorphous silicon. Promising prospects were also obtained inregard to the future production of devices with even highermultiplication efficiencies.
Research has been continued on nanometer sized silicondot film (inserted figure) to develop a photoconductive filmwith a wavelength selective function, depending on the sizeof the silicon dots. During fiscal 1999, in continuedcooperation with Hiroshima University, further technologicaladvancement was made on the production of thickphotoconductive films. We established a film manufacturingtechnique to deposit an adequate number of layers forphotoconduction by repeating a low pressure chemicalvapor deposition and an oxidation process on a siliconsubstrate.
Photoconductive Film3.5.1
Nanometer scale silicon dot film:atomic force microscope image
and its schematic structure
Nanometer scale silicon dot (size: several nanometer)
Substrate
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 35
Work is continuing on the development of asolid-state HARP imaging device that combinesa HARP film with an MOS type solid-statescanner in order to produce a future ultrahighsensitivity camera with low power consumptionthat is compact, lightweight, and highly mobile.
Solid-state Image SensorRegarding MOS transistors with high-
breakdown voltages that read HARP filmsignals, we had already accomplished a breakdown voltageof 60V by fiscal 1998*. This year, we developed a voltage-resistant technique by controlling the impurity concentrationdistribution perpendicularly throughout an element in orderto improve the MOS transistor's breakdown voltage by anadditional 20V, guaranteeing stable operation. Numeroussimulations utilizing a low-concentration junction with agraded drain structure indicate a prospective 20Vimprovement in the drain breakdown voltage.
Application of the above results advanced thedevelopment of a new diffusion process to realize a deepjunction with low-impurity concentrations.
HARP Film Lamination TechnologyRegarding a technique to connect a HARP film and CMOS
image sensor, although there are some problems withstability and uniformity, two laminating techniques have
been developed. The first is a bump connecting techniquethat is relatively easy to realize at the present technologicallevel. The second is a direct connecting technique thataffixes a HARP film directly to the image sensor. As formicro-bump connections, a polyimide-core bump formationprocess was recently developed. In comparison with theconventional indium micro-bump electrodes, this is a moreuniform method and is suitable for microminiaturization,enabling a reduction of fixed pattern noise (inserted figures).
To support the direct connection technique, progress wasmade on chemical and mechanical polishing for theoxidized and metallic film surfaces for planarization, as wellas the development of a new lamination technique todirectly connect HARP film and a CMOS readout circuit.
*Breakdown voltage of ordinary MOS device isapproximately 15V.
Solid-state HARP Imaging Devices3.5.2
Glass substrate
Polyimide-core bump
Solid-state image sensor
Solid-state image sensor with HARP film SEM photograph of polyimide-core bumps
HARP film
Polyimide-core bump formation process
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services36
To realize high-performance TV cameras for future super-realistic broadcasting, we have been promoting research ona new type of image sensor featuring ultrahigh-resolutionand superhigh-sensitivity. The new image sensor consists ofan X-Y matrix FEA (Field Emitter Array) and a HARP film.
Design and fabrication techniques Our main goals for fiscal 1999 were the implementation of
simulation techniques necessary for structure design andthe development of fabrication methods indispensable forpractical use.
For the former, we pursued studies on fundamentaltechniques for the numerical analysis of electron beamtrajectories to verify the influence of electrode structure onelectron beam characteristics. As a result, the simulation ofelectron beam trajectories emitted from a single field emitterwas made possible. Advances were also made onfabrication methods. The possibility of manufacturing acompact FEA image sensor was confirmed by the inventionof a package model (inserted figure). As the figureillustrates, a glass ring supports a mesh electrode,connecting with the HARP film and FEA at both ends. Thesensor is evacuated after the components are assembled.
In addition to engineering FEA structures to realize high-resolution image sensors, progress was made on otherimportant developments such as the dividing of HARP filmsfor drastically improving S/N and dynamic-range
characteristics on the parallel signal readout system.
Development of High-efficiency field emitter arrayWe have promoted the advancement of fabrication
techniques for high-efficiency FEA that can emit high-density electron beams applying low-voltages. Since 1998,advances have been made on a coating technique for DLC(Diamond-Like Carbon) film as a highly efficient and stablefield emitter material. We also proposed a new, highlyefficient FEA, made possible by anodization.
Vacuum sealing
Glass ring
Mesh electrode HARP film
Field emitter array
Cross section structure of field emitter image sensor withHARP film
Field Emitter Image Sensor with HARP Film3.5.3
A next generation image sensor with superior functionssimilar to biological visual systems is under development. Infiscal 1999, we referred to biological mechanisms such aslight adaptation (the perception ability of the human eyewhile gradually adapting to a different light environment,such as entering a bright place from dark place) and colorconstancy (the process by which the human eye canrecognize the same color as lighting changes) etc., toinvestigate the possibility of improving camera performancein regard to dynamic-range and white balance etc.
With regard to improving dynamic-range, there are twoproblems: low noise signal detection under low-light levelconditions, and photoelement saturation control under high-light level conditions. As an advanced wide dynamic-range
imaging scheme to obtain linear output characteristicsunder low-to-high light environments, we invented a newimaging system by combining the ordinary accumulatedsignal readout with a photodiode saturation data detectionprocess. Conducted simulations revealed the performance ofthis new system was equivalent to conventional imagingdevices under low-light conditions, with the possibility ofmore than 100 times the wide dynamic range. Based on thisproposal's pixel structure (inserted figure), a layout design ofan image sensor consisting of 32 32 pixels using 0.5mCMOS technique was completed. This research has beenpromoted in cooperation with outside research institutions(ATR, Waseda University, U.S. Georgia Institute ofTechnology).
Next Generation Image Sensor3.5.4
Reset
Saturation signal
Control block
Photoelement
Light
Storage charge
Threshold
Comparator
Wide dynamic-range image sensor pixel structure
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services 37
High Performance Plasma DisplayWith the purpose of improving the luminous efficiency of
PDPs (Plasma Display Panels), we promoted research onpanel structure, driving schemes, and panel manufacturingtechnology. To improve the luminous efficiency, a panel hasto be operated with a low discharge current. Therefore, wedeveloped a "DC memory drive method" that provides astable discharge even with a low discharge current, andapplied this method to the positive column (a dischargemode of high-efficiency ultraviolet ray emission applied tofluorescent lamps, etc.) panel. A luminous efficiency of 1.4lm/W was acquired, which is approximately 1.8 timeshigher than that obtained during fiscal 1998. Regarding thepanel structure, we developed a 3-electrode cell structuresuitable for the DC memory drive method reduced dischargecurrent variations, and degradation of image quality due tothe interference between pixels. We constructed anexperimental 6 inch (15cm) diagonal panel and obtained afavorable TV image by utilizing the DC memory drive andoptimizing the drive conditions.
Element TechnologiesTo realize highly luminescent and efficient PDPs by
improving phosphor characteristics, we investigated thedeterioration mechanism of the blue phosphor(BaMgAl10O17: Eu) in the fabrication process. It wasobserved that the deterioration phenomenon differs forexcitation by ultraviolet of 254nm and 147nm. It wasconcluded that the deterioration resulted from Eu(europium) valence changes, from a divalent to trivalentstate, on the surface of the phosphor particle.
Regarding manufacturing technology related to DCpositive column memory panels, we pursued theoptimization of the anode structure and fluorescent surfaceforming process to forward researchon high luminescent PDPs withstable operation.
Fundamental Research onDischarge Physics
Basic research regarding dischargephysics was promoted to clarify theelectron emission mechanisms of acathode and establish a dischargesimulation technique to improve PDPefficiency.
We conducted a theoreticalexamination of the insulator cathodesecondary electron yield (averagenumber of secondary electronsemitted per ion incidence with theelectrode, or per excited particle),and clarified the relationshipbetween band parameters (physicalvalues include electron affinity andforbidden bandwidth) and the value. We also calculated thetheoretical value of MgO, one ofthe AC-PDP insulator cathode
materials, against various rare gases, obtaining the value forfor the Kr and Xe ions of 0. This became a significant
guideline in the search for new cathode materials. Regarding PDPs, the effective secondary electron yield '
is lower than the value in a vacuum environment due tothe back diffusion effect whereby some of the secondaryelectrons collide with gas molecules and are returned to thecathode. A detailed examination of this back diffusion effectwas performed using the Monte-Carlo simulation procedure,and a ' value for typical conditions was calculated.Additionally, we started discharge simulations using this 'value.
As the first step to clarify the discharge mechanism for thepractical cell forms, we applied a newly developed 3-Dsimulation technique to a surface discharge AC-PDP andshowed the influences of the address electrode on thedischarge process.
A world first side-view observation of the dischargeprocess was performed for a surface discharge AC-PDP. Thecell's visible and infrared emissions were detected at a 10nsresolution for an experimental cell placed on a demountablegas chamber using an IICCD (Image Intensified CCD) camera(inserted figure). The influence of cell height and filled gaspressure on the discharge region shape was clarified.
Time lapse after disch
arge pulse applicatio
n
Ne-Xe4%, 53.3kPa, Vs=200V
180
220
260
300(ns)
Observation direction
One PDP cell
Emission
Anode Cathode
Anode Cathode
Side-view observation of surface discharge AC-PDP emissions
Plasma Display Panel (PDP)3.6.1
NHK STRL ANNUAL REPORT 1999 | Studies for Opening up Future Broadcasting Services38
In order to realize light-weight flexible large displays in thefuture, we continued to advance basic research on liquidcrystal (LC) composite films consisting of ferroelectric liquidcrystals (FLCs) and a polymer. The films contain a polymerfiber network structure (inserted figure), which is produced byirradiating the mixture of acrylate monomers (20%) andFLCs(80%) with ultraviolet rays to polymerize the monomers.In addition to the physical support for the substrates, thepolymer network acts as an agent to fluctuate the thresholdvoltage of LC microdomains in the vicinity, thus enabling gray-scale capability for the otherwise bistable FLC device.
During fiscal 1999, the polymerization process was analyzedto improve the device characteristics. Analysis by liquidchromatography revealed that less than 1% of monomersremained in the LC, indicating satisfactory separation betweenthe polymers and the LC. However, the analysis also detected
a LC photolysis byproduct induced by ultraviolet radiation, sofurther improvements in the polymerization process will benecessary. In addition it was found that the molecular weightof the polymer was in the range of 2 thousand to 2 million andthe average polymerization degree was approximately 100,which was not strongly influenced by ultraviolet intensity.
Fabricated LC film devices have a short response time (lessthan 100 s) due to the fast-switching FLCs. Also the devicesexhibit a wide field of view that exceeds 100 degrees, asuitable characteristic for picture displays. A problem toovercome is that the transmittance and contrast depreciate asthe polymer concentration increases. Device fabrication withflexible plastic films (200 m thick, polycarbonate) in place ofconventional glass substrates was also examined.
Ferroelectric liquid crystal
Transparent electrodes
Polymer
Plastic film
Plastic film
Flexible Liquid Crystal (LC) Devices3.6.3
Structure of liquid crystal film device
High-Luminance Blue EL MaterialsWith the aim of developing high-resolution full-color EL
displays, we continued our research on blue-emittingcerium-activated strontium thiogallate (SrGa2S4:Ce) as an ELmaterial. We have grown SrGa2S4:Ce thin films by an ultra-high vacuum multi-source deposition process, usingmolecular beam epitaxy (MBE) equipment.
During fiscal 1999, in order to establish guidelines toimprove the luminance, we investigated the chemical statesof SrGa2S4:Ce constituent elements in detail by means of X-ray photoelectron spectroscopy (XPS). Especially, we foundthat negatively divalent oxygen ion O2- (impurity) occupieda neighboring site of trivalent cerium ion Ce3+ (luminescentcenter). We concluded such atomic arrangements give riseto the enhancement of energy transfer efficiency fromelectrons to centers. Furthermore, it was revealed that a thinfilm material of a different composition, Sr2Ga2S5:Ce,provides blue photoluminescence with good color purity[chromaticity coordinates of (x, y)=(0.15,0.12)].
In order to improve high-resolution full-color displaytechnology, we fabricated a prototype EL device in whichthe luminous elements for the primary colors (red, green,blue) are stacked layer by layer (inserted figure). We choseSrGa2S4:Ce for the blue emitting layer, SrS:Ce for the greenemitting layer, and ZnS:Mn for the orange layer. Redluminescence was extracted from the orange luminescenceby means of a red filter. The respective layers' measured
chromaticity coordinates were R (0.64,0.36), G (0.19,0.37),and B (0.15,0.10). White chromaticity coordinates obtainedby combining the primary colors were (0.32,0.25). Theprototype color EL panel has 64 64 pixels, a pitch of0.5mm, and a pixel area of 0.3 0.3mm2.
Glass substrateGlass
substrateInsulating
layer
Aluminum electrode
Black panel
Protective layer
Transparent electrode
Transparent electrode
Transparent electrode
Blue emitting layer
Green emitting layer Orange emitting layerRed filter
Display side
4mm
Insulating layer
Insulating layer
Layered structure of a color EL device
Inorganic Electroluminescent (EL) Devices3.6.2
39NHK STRL ANNUAL REPORT 1999 | Presentation of Research Results
To facilitate the wide overseas distribution of STRLresearch activities and accomplishments, the Englishjournal "Broadcast Technology" was recently inaugurated(to be published quarterly). We also publish the Englishtreatise journal "Laboratories Note" which is designed toallow STRL researchers and professional engineers toannounce the latest research results on a global scale.
Total 64*1: International Conference on Computer Communication*2: International Display Workshops*3: International Conference on Image Processing*4: The Brazilian Society of Television Engineering*5: Montreux Symposium, HDTV Workshop, ABU, SMPTE, etc.
The two day 1999United Nations WorldTelevision Forum washeld at the UnitedNations headquarters inNew York on November18 and 19. NHKpresented exhibitions on digital satellite broadcastingservices that are to begin in the year 2000 and futurehome server technology.
Overseas Exhibitions
TELECOM99 was heldat Palexpo in Geneva,Switzerland, fromOctober 10 throughOctober 17, 1999. NHKparticipated in the eventand introduced the ISDBconcept based on the theme: "Digital BroadcastingSystems in the 21st Century." Demonstrations of NHK'sresearch accomplishments were also presented such ashome server systems and a visual video retrieval system.
The STRL also maintains an English version of its homepage to present detailed information concerning thelaboratories on a global scale. This homepage includes anoutline of the laboratories, research results, eventinformation such as the STRL Open House, press releases,and journals.
Homepage
UN Forum TELECOM99
Overseas academic meetings and international conferences
Exhibitions, Address, and Presentation
STRL Open House 1999
Image-based virtual studio We are developing technology thataims at creating an image-based virtualstudio set based on image componentsfrom real pictures instead of CG. Thepresentation was made of an image-based virtual studio set which allows auser to freely arrange images of smallobjects such as desks or vases (3-Dimage component) on an ambientbackground scene (environmentalimage component).
40 NHK STRL ANNUAL REPORT 1999 | Presentation of Research Results
Exhibition Outline
STRL Open House
Wednesday, May 26 Open house ceremony
Thursday, May 27 By invitation
Friday, May 28Open to the general public
Saturday, May 29 Open to the general public
Sunday, May 30 Open to the general public
Open House Schedule
0 5,000 10,000 15,000 20,000 25,000
199919981997
Attendance
Exhibition: 26 sites/33 itemsAttendance: 22,905
1. Digital BS-The Future of Television
Hi-Vision (HDTV) newsAn Open Studio was established at thesite to show visitors the entire processof Hi-Vision (HDTV) programming, fromthe editing/ production of HDTV newsof the day to its broadcast on HDTV setsaround the Laboratories at thirty-minuteintervals.
From watching to using TVDigital satellite broadcasting will provide anew service, data broadcasting. This newservice includes: EPGs (Electronic ProgramGuides), news headlines and theirsummaries, weather forecasts, and flashreports of important information such asfor earthquakes and tsunami waves. Someof the services were presented at the OpenHouse.
2. The Latest DigitalBroadcasting Technology
Home server of the futureA home server is a home storagesystem of the digital era for realizing an"anytime function" which records aviewer's favorite programs orinformation for subsequent viewing atany time. Two types of home servers,one built into a receiver and the other ahome network-type, which can beutilized from multiple TVs in the house,were exhibited.
Cable transmission of BS digitalbroadcasting We have been conducting research ona digital satellite broadcast re-transmission system in order to makethe programming available over cableTV networks. The following threeschemes were demonstrated at theopen house: 1) transmission with twocable 6MHz TV channels, 2) new wide-band channel transmission, and 3)frequency-converted transmission.
3. Technology for the EnhancedCreation of Program Material
Video retrieval system for archivesWe have been advancing with researchon high-speed video retrieval systemsfor massive video archives. A prototyperetrieval system and high-densityrecording (20 times of D-3 VCR)technology for this system weredisplayed.
Broadcast news transcription system As a part of human-friendlybroadcasting services, there is anincreasing demand for the number ofTV programs with closed-captioning.We exhibited a "speech recognitionsystem" which automaticallyrecognizes a newscaster's speech andcreates a script for closed-captioning.
41NHK STRL ANNUAL REPORT 1999 | Presentation of Research Results
Multi-beam receiving flat antenna forsatellite broadcastingIn the future, it is expected that manysatellites will provide broadcastingservices. Therefore, we are pursuingresearch on a flat antenna which cansimultaneously receive program signalsfrom multiple satellites. At the openhouse, we exhibited a prototype flatantenna. This antenna has a range of
20 east-west, which covers BS(110 E) and the furthest CS (144 E) .
TV remote controllers for the digitalbroadcasting eraThe various services that will bedelivered to the home by digitalbroadcasting should be readilyavailable to anyone. We introducedevaluation test results on TV remotecontrollers and program guide menuscreen designs. Several remotecontrollers designed on the basis ofthese test results were also presented.
6. Hi-Vision (HDTV) Wall-mountedTV
Hi-Vision PDP TV sets have beendeveloped and manufactured. The HDTVdisplays are very thin, less than 10cm indepth, making it possible to enjoyimpressive natural color images andsharp text on large HDTV screens thatutilize a minimal amount of space. Itsdiffusion is anticipated as a wall-mountedHDTV for the home.
Field emitter image sensor withHARP targetA compact camera capable of capturingclear images under dark or dim lightingconditions is in demand for specialbroadcasts such as emergency reports.We are pursuing studies on a new flatimage sensor consisting of a high-sensitivity HARP target used inconjunction with micro-sized fieldemitters.
Symposium
Keynote Address
PresentationsAttendance: 615
Attendance: 692
Symposium
Keynote Address
Presentations 15:00-15:30
15:30-16:00
16:00-16:20
16:20-16:40
16:40-17:00
"Search for the Possibility of New Broadcasting: A Discussion Concerning the Appeal of Digital Satellite Broadcasting"Participated in by the presidents of digital satellite broadcasters and an NHK Managing Director
"Basic Concepts and Technologies Underlying Effective Production in the Multimedia Era"Mr. Shigeru SHIMODA (Multimedia Services, NHK Science and Technical Research Laboratories)
10:30-12:30
14:00-15:00
15:00-17:00
Digital Terrestrial BroadcastingMr. Makoto SASAKI (Digital Broadcasting Networks, NHK Science and Technical Research Laboratories)Subjective Listening Tests of MPEG-2 AACMr. Joji URANO (Nippon Television Network Corporation)Speech Recognition System for Closed-captioning Broadcast News Mr. Akio KOBAYASHI (Human Science, NHK Science and Technical Research Laboratories)Video Retrieval System for ArchivesMr. Takashi KATO (Recording Technology & Mechanical Engineering, NHK Science and Technical Research Laboratories)Basic Study on Integral 3-D TelevisionMr. Haruo HOSHINO (Three Dimensional Audio-Visual Systems, NHK Science and Technical Research Laboratories)
Presentations (Friday, May 28) Symposium
3-D virtual reality audio systemWe are advancing with fundamentaltechnologies for next generation superreality broadcasting, such as a 3-Dvirtual reality audio system with soundlinks to CG images. This system iscapable of reconstructing dynamic real-time 3-D sound fields that correspondto 3-D CG images.
4. Technology for the Realization ofHuman-Friendly Broadcasting
5. Fundamental Technologies forFuture Broadcasting Services
On Friday, May 28, 1999, a symposiumwas held and participated in by thepresidents of f ive digital satell itebroadcasters and an NHK ManagingDirector. The event was a great successwith both the main location, theauditorium of the NHK Science & TechnicalResearch Laboratories (300 seats), and theprovisional monitor location (300 seats)being completely filled. Announcer Sadatomo MATSUDAIRA ofNHK was the symposium mediator fordiscussions that concerned new services,diffusion strategies, and the future image ofdigital satellite broadcasting.
42 NHK STRL ANNUAL REPORT 1999 | Application of Research Results
Total:27
Program Cooperation
Application ofResearch Results
General TV, November 3rd,1999
The "insect microphone" that the STRL developed wasintroduced in the November 3rd edition of the NHK GeneralTV program: "Tameshite Gatten." This microphone can pickup special sounds, such as an insect's footsteps, byconverting micro-vibrations into electrical signals. Anotherapplication of this principle was seen in an "ice-zonemicrophone" used during the ice skating events at theNagano Olympics in February 1998, to capture clear skatingsounds from within the ice.
Examples of programs that applied STRL's research results
Returning Research Resultsto the Public
It is anticipated that a new Super HARP HDTV camera withoutstanding high-sensitivity characteristics developed by ourlaboratories for night-time shooting will have applicationsoutside the broadcasting field. For example, due to inadequatefilm sensitivity, the current retina examination procedurerequires the injection of a large quantity of fluorescent dye intothe patient's vein. This procedure is hard on the body, eveninducing nausea in many patients.However, in November 1999 at the MIYAKE ophthalmic hospitalin Nagoya City, the first retinal examination utilizing the newSuper HARP HDTV camera was performed. The camera has asensitivity about 100 times greater than a conventional one, andthe experiment confirmed its capability to evaluate the retina'scondition with a minimal amount of dye. The experiment alsorevealed that the video images produced by the camera weresuitable for determining the circulatory condition of bloodvessels.
A CG character produced by TVML appeared for the firsttime in NHK programs during the 1999 summer vacation.In the BS2 program "New Midnight Kingdom," a high-quality CG character produced by TVML PRO (TVML high-quality picture version)appeared on the screen as ananchorperson of the program. In the program "Let'sAccess Digital Wonder Box: We want to know more aboutJapanese Rice," which was aired on Educational TV,children's ideas were gathered during the live studioprogram and immediately broadcast as a mini-programthat they produced using TVML.
High picture quality TVML CG
character ("New Midnight Kingdom")
TVML's activity in summer special programs
Appearance of an "insect microphone"
New Super HARP camera application to medical fieldA powerful tool for retina examinations
Research Data, etc.
Type New Total179 1188 28 199207 1387
DomesticOverseasTotal
Patent and utility model applications submitted
Patent and Utility Models
ContractsLicenses Patents Know-how
TotalType New 11 135 13 276 10 237 3 39
Patents and utility models in use
Type New Total118 887 88 506206 1393
DomesticOverseasTotal
Patents and utility models granted
Technical CooperationTotal: 29
BS-2, August 13,1999
General TV, August 20, 1999
Educational TV, August 21,1999
as of end of March, 2000
Program production using the
"insect microphone"
Ophthalmic examination experiment (From NHK news)
Others
43NHK STRL ANNUAL REPORT 1999 | Others
ITU-R SG10ITU-R SG11
ITU-R IRG*1
ITU-R GTE*2
ITU-T SG9
ITU
WP-A
WP-B
ABU
DAVIC
TV Anytime
MPEG-4
MPEG-7
FIPA
AGTE*
Contributions from NHK Labs.Report on narrow band ISDB-T(WP10B).Unification of the 1125 HDTV studio standard (WP11A).Addition of ISDB-T to the DTTB Recommendation (WP11A).Revision of the HDTV digital interface Recommendation (WP11B).New ISDB-S Recommendation ISDB-S.Re-planning of BSS (JWP10-11S).New recommendation on the subjective assessment of stereoscopic television pictures, including test materials (JWP10-11Q).New test materials for television picture (JWP10-11Q) subjective assessments.XML-based multimedia coding scheme (JTG10-11).Examination of possible Region 3 plan review and revision based on NHK's ATGE draft plan.Region 3 re-planning technical examination based on NHK's ATGE draft plan.New recommendation on user requirements for statistical multiplexing.
Narrow band ISDB-T report.800MHz-band OFDM Outside Broadcast Link study.Production of FM multiplex (DARC system) handbook.
A TV Anytime concept.A storage-type broadcasting service.Home server functions.TV Anytime system requirements.Technical tools such as APIs.A program indexing system for the realization of program script viewing. A metadata personal filter.A TV Anytime system reference model.
A home server architecture.Demonstrations of NHK's home server.
TV Anytime service model requirements.
TV Anytime system requirements.TV Anytime metadata requirements.TV Anytime security requirements.
Studio profile specification, software implementation, verification by bit stream exchangesCamera parameter transmission scheme and software implementation/standardization. Requirements and specification for scene description broadcasting applications.
A proposal and verification experiments on a description scheme from the viewpoint of video editing/production.A proposal for a description scheme for structured video annotation.
Contribution of the followings to FIPA via the FACTS* project.Validation of FIPA specifications through field trials.Provision feedback for the improvement of FIPA specifications.
Contribution to the IRG and GTE through the AGTE plan compilation as a Region 3 member based on NHK's draft plan.
APT
Chairman, System Design AHGT. Kurioka
Vice-Chairman, WP11BY.Nishida
Vice-Chairman, JWP10-11Q K.Watanabe
Vice-Chairman, JTG10-11 T.Yoshimura
Business Models-WGTechnology-WG
Application-TC
CAMS-TC
System-TC
Terms of referenceSound broadcastingTelevision broadcasting
Review and revision study of the BSS PlansRe-planning feasibility studyTelevision and sound transmission
New systems and services
Sound broadcasting
Application, service, and requirements
Content representation, APIs, metadata, and security
System design and integration
Application, service models, and scenariosSystem, metadata, contents referencing, and security
Coding of audio-visual objects
Multimedia contents description interface
Standardization of agent technology
Region 3 re-planning feasibility study
Chairman H.Miyazawa
Vice-Chairman T. Kurioka
1 Inter-conference Representative Group 2 Group of Technical Experts
FIPA Agent Communication Technologies and Services
APT Group of Technical Experts
Contributions to International Standardization Organizations
44 NHK STRL ANNUAL REPORT 1999 | OTHERS
Planning & Coordination
Science & Technical Research Laboratories Construction Administration Office
NHK Science & Technical Research Laboratories
Multimedia Services
Digital Broadcasting Networks
Digital Satellite Broadcasting Systems
Advanced Audio & Video Coding
Human Science
Advanced Imaging Devices
Display and Optical Devices
General Affairs
Executive Research Engineer Yozo UTSUMIYuichi NINOMIYAKatsumi NAKABAYASHI
Planning and management of research work, International correspondence, Public relations on research issues
Three Dimensional Audio-Visual Systems
Recording Technology & Mechanical Engineering
Human-friendly broadcasting, Image and speech recognition technology, Human interface, ...
Organization
In order to meet the need for more advanced research in thedrastically changing media environment, work is underwayby the NHK Science & Technical Research Laboratories toconstruct a new research center. The core function of thisnew center will be broadcasting technology research for the21st century. The new structure, with a total floor space of22,600 square meters and 77m in height, will have 14stories aboveground and 2 stories below. The new STRLbuilding complex is scheduled to be completed in October2001. Full-scale construction started in April 1999. Piling work, including foundation piling, was followed by basementexcavation work. Starting in September, we proceededwith work to counteract the radio interference inneighboring areas that will be caused by the new building.Foundation and structural work began in December, 1999.
Doctorates acquired in fiscal 1999 4
STRL doctorate holders as of March 31, 2000 58
(Doctors of Engineering)
In fiscal 1999, a researcher from Canada was invited tothe STRL. A researcher from the Republic of South Africahas also being conducting research at the laboratories(since last fiscal year). The STRL sent two additionalresearchers to the United States and two otherresearchers will continue their work in the U.S. which wasstarted in fiscal 1998.
