Fujitsu to Release H.264 Format Video-Processing LSI Chip Supporting High-Definition and Low Power Consumption

Tokyo, November 30, 2006

Fujitsu Limited today announced the development of a new large-scale integrated circuit (LSI) chip capable of real-time compression and decompression of high-definition TV-level video, in H.264⁽¹⁾ format. The new LSI chip, to be released in March 2007, is based on a H.264 processing algorithm developed by Fujitsu Laboratories Limited and integrates Fujitsu's high-performance audio and video processing technology with its 90nm process technology. This is the industry's first such chip to support H.264 High Profile, Level 4.0⁽²⁾ standard used in next-generation DVDs. This LSI enables high-resolution recording, playback, and transmission of high-definition video on audio-visual products such as portable AV products, hard disk recorders, and home network equipments.

The new chip will be exhibited and demonstrated at the International Telecommunications Union (ITU) “Telecom World 2006” scheduled to be held December 4-8 in Hong Kong.

In recent years, the need has grown for higher functionalities of audio visual products, such as recording and playback of high-resolution high-definition video. Products that can handle high-resolution video need to be able to compress a large volume of data, and the use of H.264 format - featuring superior compression performance compared to previous formats such as MPEG-2⁽³⁾, is growing rapidly.

While H.264 offers superior compression performance compared to previous formats, it also entails over 10 times the computation load that must be processed at the time of compression. Furthermore, the High Profile, Level 4.0 version of H.264 used in next-generation DVDs requires even higher data processing volumes to enable greater image quality, necessitating a high-speed LSI chip to handle real-time H.264 compression.

Building on its work of developing LSI chips for MPEG formats, Fujitsu have been involved in standard-setting organizations related to H.264 and has accumulated relevant technological knowledge. By using proprietary compression technology developed by Fujitsu Laboratories to reduce the processing load required, and incorporating Fujitsu's embedded memory technology and 90nm manufacturing process technology for compact size and low-power operation, the new chip is the industry's first LSI chip of its kind that can handle H.264 High Profile applications.

The chip has a wide potential range of applications for equipment requiring high-resolution video processing, such as battery-powered portable AV products, desktop-type hard disk recorders, and home network equipment for video transmission.

Fujitsu will continue to enhance further the high functionality of this new chip, to be full high-definition (maximum up to 1920 dots x 1080 lines) supportive. Fujitsu will enrich its line of video processing LSI chips, including products such as a "multi-decoder" chip that can handle multiple formats, including both H.264 and MPEG-2.

Sales Target

5 million chips (inclusive of all H.264-supportive chips from Fujitsu) within fiscal 2007 (April 2007 - March 2008).

Product Features

1. Real-time compression and decompression of HDTV video and sound
Capable of real-time compression and decompression of high-definition video (maximum up to 1440 dots x 1080 lines) using H.264 High Profile Level 4.0 format, with simultaneous compression and decompression of sound in Dolby Digital⁽⁴⁾ format. This is the first time in the industry that these processing features have been integrated onto a single chip for general use (according to internal research as of November 2006).

2. Embedded memory enables smaller size and lower power consumption
This is also the industry's first H.264-supportive LSI featuring two pieces of embedded 256Mbit (MB) FCRAM , with optimized design that combines logic and memory on one LSI chip, enabling a more compact size and lower power consumption.

3. Proprietary compression and image-enhancement technology
The new chip has a proprietary "self-tuning" algorithm developed by Fujitsu Laboratories that automatically applies lighter compression method to high-action zones where compression artifacts are most noticeable, such as human faces or slow-moving objects, and stronger compression method in other zones. This enables image data size to be reduced to one-half to one-third that of the MPEG-2 format at an equivalent level of image quality.