Date Visited: October 4, 1991

Report Author: J. Larimer





Tadamichi Kawada

Executive Manager, Technology Enterprise Promotion Section,
NTT Interdisciplinary Research Laboratories

Shigenobu Sakai

Image Devices Research Group Leader, Senior Research
Engineer, Supervisor, Electron Devices Laboratory, NTT
Interdisciplinary Research Laboratories

Tadaaki Masumori

Senior Research Engineer, Supervisor, Image Devices Research
Group, Electron Devices Laboratory, NTT
Interdisciplinary Research Laboratories

We were met by Mr. Kawada and taken to a laboratory area, where we were joined by Mr. Sakai and Mr. Masumori. We were shown a teleconferencing system and a 15-inch diagonal high-resolution AMLCD and were given lunch in the company dining facilities. Discussions took place during the entire period.


The teleconferencing system consisted of several individual stations situated around a conference table. The system we saw had five stations, but the number of stations is not likely to be a limitation. An individual station consisted of a 9.5-inch diagonal AMLCD display and a data input device. The master station included a 9.5-inch diagonal AMLCD overhead projector and a computer keyboard and monitor. All stations were networked to the PC-type workstation at the master station. The system included voice teleconferencing. These components constitute a node on a teleconferencing network.

The system did not include direct televisual communications capable of conveying real- time images of the conference participants. Visuals such as graphs and pictures could be displayed on the individual station displays or on the overhead projector. During a teleconference, individuals at the various nodes can interactively manipulate a visual such as a graph or make database entries that are displayed visually. The nodes are connected over a 64 Kbits/second T1 line, the equivalent of 16 phone lines, with 56 Kbits/second devoted to voice and 8 Kbits/second to data.

The flat-panel displays used in the teleconferencing system were manufactured by Hosiden. The 9.5-inch diagonal display was a VGA (i.e., 640 x 480 color pixels) with a-Si TFTs and no gray scale. It has an 8-color palette formed by binary combinations of R, G, and B. The 9.5-inch diagonal overhead projector display was also made by Hosiden.

NTT showed us the teleconferencing system as an example of how it integrates technology to provide telecommunication services to customers. The research and development staff is used both to promote key enabling technologies such as flat- panel displays and to engineer systems such as the teleconferencing system. Mr. Sakai's group played a central role in developing the teleconferencing prototype system. Mr. Kawada's section, called the "Technology Enterprise Promotion Section," promotes the establishment of technology enterprises based upon NTT technologies.


The flat panel that was codeveloped contains NTT proprietary technologies that now are licensed to the codevelopment partner. We were given a paper describing the two 15- inch displays, which was presented at the last EURO Display Conference.

The NTT side of the display project was under Mr. Kawada's leadership. NTT provided three specific technologies:

  1. device fault tolerance or redundancy technology,
  2. high speed driver technology, and
  3. low-resistance bus line technology, primarily a material fabrication issue that affects addressing speed and therefore the ultimate size and pixel count of this and future displays.

As the manufacturing process becomes more stable and controlled, the need for the redundancy schemes will lessen and the redundancy technology (e.g., multiple TFTs per pixel) will be removed as the manufacturing yields increase. Driver technology on the AMLCD substrate is important, and compatibility with existing CRT-based systems is an important feature. The source bus lines are made in an ITO/Mo/Al three-layer stack configuration. Aluminum bus lines often form hillocks during heat processing, so NTT had to develop a hillock-free process. They believe their bus technology will support the development of future 30-inch diagonal flat-panel AMLCDs.

The high-resolution 15-inch display has 4 bits or 16 levels of gray per pixel and can address a color space of 2(superscript 12), or 4096 colors. We were shown no data on the discriminability of these colors, which of course depends upon the backlight, filters, pixel geometry and aperture ratio, viewing conditions, and the human visual system. We were shown the high-resolution version of the display, and it was subjectively the best large AMLCD display we saw on the entire trip, including all the LCD displays we saw at the Japan Electronics Show. I assume that the 15-inch display shown by Hosiden at the show was the VGA version.

The VGA version has 1920 x 480 dots arranged in a stripe configuration for 640 x 480 color pixels. The drivers for this display are analog and can display full color. We did not see or discuss this version of the 15-inch display, although I believe it was the 15-inch display that Hosiden showed at the Japan Electronics Show.

The high-resolution display has 1920 x 1600 dots that are arranged into RGB triads for 1280 x 800 color pixels. The gate-line drivers are divided between the right and left sides of the display, alternating two rows from one side and the next two rows from the opposite side. The source-line drivers are divided between the top and bottom of the display, with alternating columns driven from either the top or bottom. Since a pixel triad spans two rows, addressing one row of color pixels requires addressing two lines of dots. To do this a "1 line-2 scan interlaced-drive" scheme addresses first one line of dots and then the second line. All of the lines addressed from one side of the display are scanned before the lines addressed for the opposite side are scanned; this creates an interlace similar to conventional television, but based on line pairs rather than single lines.

There are two driving modes for this display. In Mode 1, 1120 x 750 color pixels are addressed at 40 frames/second (vertical sync rates of 80 Hz). In Mode 2, 4 triads or 12 dots (two contiguous gate-line driver pairs, one from each side of the display, and three contiguous source lines) are addressed as a single color pixel. In Mode 2 the vertical sync rate and frame rate are the same at 56.4 Hz or frames/second. In Mode 2 there are 640 x 400 addressable color pixels. It was not clear whether or not the driver scheme permits individual dot addressing in Mode 2. If individual dot addressing is possible, then it would be possible to implement dithering schemes to further enhance the gray scale and therefore the image quality performance of the display in this mode.

The display we saw was connected to a PC and could accommodate several frame buffer sizes (e.g., 1024 x 760, 1120 x 750, and 640 x 400). We saw only static images, but the Hosiden 15-inch display we saw at the Japan Electronics Show subjectively seemed to have excellent temporal performance. If these are the same displays, then one would expect the motion-rendering capabilities of the high-resolution display to also be good-- but we saw no direct supporting evidence for this conclusion. Subjectively, the spatial image quality was outstanding. The display had excellent uniformity; although some pixels were out, they were difficult to spot in images with high information content. The viewing angle performance seemed very good. NTT reports a contrast loss of approximately 3 db over approximately 15 degrees of solid angle and good performance--approximately an order of magnitude contrast loss--for 50 degrees right or left and 40 degrees up and 20 degrees down.

The backlight was 26W fluorescent @ 3000 cd/m2. NTT claimed a screen luminance of approximately 50 ft-lamberts or 5% efficiency, but it seemed dimmer--more like 10 or 15 ft-lamberts, which would put its efficiency at 1% or 2%. Again, there were no data on this, so this is a subjective estimate. The high-resolution display is a well- engineered device that can be easily designed into a variety of systems. It appears to have excellent compatibility with many existing frame size standards.


NTT's research role is to promote the development of enabling technologies essential to NTT's future needs. It selects technologies that Japanese companies are unwilling to develop because of the risks and difficulties these technologies impose upon the developer. NTT is often able to enter into codevelopment projects with private companies, thus ensuring that essential technologies will be available when needed.

The completion of the 15-inch displays in July of 1991 marked the end of NTT's research efforts on direct-view LCDs. Now that there are many Japanese companies working on direct-view AMLCD technology, NTT's efforts are no longer required. The next topic to be addressed by several of the individuals who worked on the 15-inch display project will be a video telephone. The goal is to develop a system capable of producing the subjective experience of presence between two speakers at different locations. It is believed that the system must provide eye contact between the speakers.

NTT enters into joint research and development efforts with other companies to develop critical technologies. One of the mechanisms that NTT used in the past to support and encourage joint efforts has been the royalty system. Royalties from codeveloped technologies were returned to the NTT technical staff to support continued R&D.

It was estimated that there are several thousand people working on and researching LCD technology within Japan. Approximately 20 manufacturing companies and an equal number of university and government labs are also working on this technology.

Mr. Kawada speculated that direct-view flat-panel displays based on a-Si TFT AMLCDs, which have now achieved 15 inches in size, will achieve 30-inch diagonal sizes in the next 10 years. Direct-view displays of 50 inches will require the development of a plasma display panel. He also speculated that for projectors, poly-Si TFTs will dominate, and they will span a projected image size of 30-200 inches. CRT projectors will also continue to be viable in this range. Near-term solutions to the image brightness problem will include dual projector systems.

We were also given two viewgraphs, one of which showed the American contributions to AMLCDs:

  1. Graphic Control Specification, VGA, EGA,...
  2. Poly-Si TFT LCD Technologies
    (David Sarnoff Research Center)
  3. Glass Substrate Fabrication (7059)
  4. Step and Repeat Lithography System

The second viewgraph, entitled "Future Prospects of LCD Technologies," stated that "active matrix LCD will come into wide use as a high information density display within five years, after the following improvements:


Yield enhancement/process reduction


Larger than 10 inch, better than VGA,
wider viewing angle


Larger backlight life

Main fields:

A-Si TFT LCD for direct-view displays
(3-15 inches); poly-Si TFT LCD for
projection displays (3-5 inches with circuitry)"


A brochure described the activities of the four NTT R&D centers, with 12 laboratories distributed across the centers. The major laboratories are as follows:

  1. Telecommunication Networks Laboratories
  2. Network Information Systems Laboratories
  3. Human Interface Laboratories
  4. Communication Switching Laboratories
  5. Transmission Systems Laboratories
  6. Radio Communication Systems Laboratories
  7. Software Laboratories
  8. LSI Laboratories
  9. Optoelectronics Laboratories
  10. Interdisciplinary Research Laboratories
  11. Basic Research Laboratories
  12. Communication Science Laboratories

NTT developed an Integrated Services Digital Network, ISDN, which has been in service since 1985. NTT's broad goals are to:

  1. provide easy access to all types of information and communication equipment,
  2. provide access from any location, and
  3. define the systems characteristics on the basis of patterns of usage.

They envision a triple-faced communication system that encompasses service technologies, network technologies, and basic research to support future communication systems. The following list represents R&D topics and products within these three categories.

Service Technologies

  1. Speech recognition and production
  2. Fast data encipherment
  3. Knowledge base management systems
  4. Hardware-independent network technologies
  5. Office automation technologies

Network Technologies

  1. Photonic switching technologies
  2. High-speed coherent optical communication technologies for chromatic dispension compensation
  3. Optical frequency division multiplexing technologies
  4. Uniplanar monolithic microwave circuit technologies
  5. Multibeam satellite communications
  6. Network planning support systems
  7. The intelligent network
  8. A synchronous transfer mode-switching technology

Basic Research

  1. High-speed optical disk memory
  2. Synchrotron orbital radiation x-ray lithography
  3. Integrated laser array technology for frequency division multiplexing
  4. LSI optical devices
  5. Ultra-high-speed GaAs integrated circuits for optical transmission systems
  6. Underground radar systems
  7. Quantitative software management
  8. Materials characterization
  9. Velocity modulation transistors
  10. Cell culture studies of neural networks

Published: June 1992; WTEC Hyper- Librarian