FOR INFORMATION
September 4, 1998


NTT Develops One-Chip MPEG-2 Video Encoder LSI
with Expandability for HDTV



Nippon Telegraph and Telephone Corporation (NTT) and NTT Electronics Corp. (NEL) have jointly developed the SuperENC, a MPEG-2 video encoder large-scale-integration (LSI) chip that compresses large-volume video data. Sales will start September 21.

The SuperENC handles video compression with one chip, or HDTV (high-definition television) compression through connection of multiple chips. It also incorporates multiple processors to flexibly compress video signals in any of four NTSC/PAL*1 modes. It is widely applicable in video-related fields, such as digital video discs (DVD) equipment and video transmission, including video conferencing.

Moreover, it is the smallest MPEG-2 video encoder LSI and consumes the least amount of energy of any encoder in its class.


Background of Development

MPEG-2 is an international standard for video compression centered on standard and high-definition television, as well as an increasing range of multimedia fields, such as telecommunications and broadcasting, because of its high ratio of compression.

MPEG-2 uses an encoder to compress video signals and a decoder to then decompress the signals. Decoders, such as those used by viewers of direct broadcast satellite TV, for example, have already been reduced in size and cost, while encoders have remained relatively large and expensive. Small, low-cost MPEG-2 encoders are needed, however, for use in equipment utilizing DVD and for use in MPEG-2 encoders for terrestrial HDTV broadcasting, which is expected to expand soon.

NTT developed in 1996 the Reimay, the world's first low-latency*2 encoder board for high-quality MPEG-2 video suitable for interactive telecommunications. NEL sells the Reimay. The SuperENC incorporates advances in the Reimay's technology, and can be widely applied for video transmission and video recording.


Features of SuperENC Technology

1. HDTV-Compatible Compression Technology

High-volume video can be compressed and encoded by connecting many LSIs, which expands encoding capacity, but conventional encoder LSIs do not have this capability.

The SuperENC automatically adjusts the video compression process between LSIs when connected. In addition, connecting LSIs enables real-time encoding of high-volume video signals, such as HDTV, so the SuperENC can compress various types of video.

The SuperENC incorporates new one-pass variable bit rate*3 technology for recording with DVD recorders, and supports DVD-standard video packet output*4.

2. Compatibility with Four NTSC/PAL Modes

The SuperENC, in addition to the RISC processor*5 used in previous MPEG-2 encoder LSIs to control overall operation, has a built-in processor comprising 18 computation units. Parameters for the processors can be changed easily and the programs can be interchanged easily, so the SuperENC is suitable for four video compression encoding modes: the standard mode for communications satellite broadcasting, the low latency priority mode for TV conferences, the high compression priority mode for long-duration recording, and the high picture quality priority mode for professional broadcasting studios.

3. High Picture Quality and Low Energy Consumption with Original Motion Detection Unit

Improving video quality requires that the motion of the objects in a video be detected with high precision, but this requires high-volume computations that work against efforts to achieve a high picture quality and low-volume energy consumption.

In order to solve this problem, the SuperENC uses a motion detection unit that combines two methods: the Look Neighbor First Search (LNF) method, which selectively uses a wide-area search or proximity-area search, and area hopping*6, which improves tracking of high-speed motion. Together, these methods enable motion to be detected accurately and efficiently with minimum searching, which vastly reduces computations for movement detection, while also producing excellent picture quality with low consumption of energy.


Products

1)Standard Mode Model (Standard Mode, Low Latency Priority Mode, High Compression Priority Mode)
2)High Picture Quality Priority Mode Model (Standard Mode, Low Latency Priority Mode, High Compression Priority Mode, High Picture Quality Priority Mode)
3)HDTV Model (Standard Mode, Low Latency Priority Mode, High Picture Quality Priority Mode)


Sales

Products compatible with the standard mode will go on sale on September 21. Other products will become available at later dates.


Prices

Standard-mode product samples will cost ¥30,000, excluding consumption tax. Prices for other products are not yet decided.


Specifications

Motion Estimation Search Area: 211.5 pixels (horizontal) by 113.5 pixels
(vertical)
Voltage:Internal-2.5 volts; Input/output 3.3 volts
Power Dissipation:1.5 watts
Package:208 Pin Plastic QFP
Attached SDRAM:NTSC/PAL-16Mx2; 4:2:2P-64Mx1
Encoding Standards:MP@ML, SP@ML, Half D1, 4:2:2P@ML, and MP@HL (connected chips)


Notes:

*1: NTSC/PAL
These are formats for transmitting color video signals. NTSC, employed in the United States and Japan, has 525 scanning lines, the capacity to send 30 frames per second, and a vertical/horizontal ratio of 3:4. PAL, the format employed in Western Europe, except France, has 625 scanning lines, the capacity to send 25 frames per second, and a vertical/horizontal ratio 3:4.

*2: Low Latency
Low latency refers to the time from image input to compressed signal output. Smooth telecommunications with interactive video require the sender to deliver images to the receiver on a real-time basis as much as possible, so signals must be encoded in minimal time.

*3: One-Pass Variable Bit Rate
This system maintains video quality by changing encoding compression rates in real time according to the conditions of the signals to be recorded, when recording is done with a recording medium with a restrictive capacity. It allows maximum use of a video recording medium's capacity.

*4: Video Packet Output
This is a basic data format for writing to DVD. It incorporates streamed, 2Kb-long bundles of data, called packets.

*5: RISC (Reduced Instruction Set Computer) Processor
These processors employ only simple instructions, which simplifies the control system, frees hardware design and increases processing speed.

*6: Area Hopping
This technology adjusts the search area by estimating motion in the entire screen.


-
MPEG-2 video encoder large-scale-integration (LSI) chip“Super ENC”



News Release Mark
NTT NEWS RELEASE