PDF《Now and Beyond》

4 in 1995, and we have been proposing new technologies and editing specification documents since then. Some of the proposed technologies incorporating MPEG-4 include motion-compensation error- controlling technology(1), high-speed-warping technology, and global-motion-compensation technology, and a technology incorporating H.264 is spatial direct motion prediction. As a good example of such a technology, motion-compensation error control is described in the following. This technology can prevent the reddening of pictures caused by accumulation of errors in motion compensation. The occurrence of reddening is detected by rounding-up error processing to whole numbers, and a method for determining these numbers by means of implementing control variables was proposed. An example of the results from this method is shown in Fig. 3. This method has been adopted into the MPEG-4 standard. technology is laid down by the VCEG (Video Coding Experts Group) under ITU-T (International Telecommunication Union-Telecommunication Standardization Sector) and the MPEG (Moving Picture Experts Group) under the authority of the ISO (International Organization for Standardization) and the IEC (International Electrotechnical Commission). Hitachi is not only involved in commercial application of these standards but also actively involved in developing other standards. As well as giving a simple overview of MPEG and VCEG video coding standards, this report presents the current activities of Hitachi regarding video coding standards and the future prospects for these standards. OVERVIEW OF VIDEO CODING STANDARDS The specifications standardized by MPEG and VCEG are shown in Fig. 1.The video coding standards laid down in MPEG consist of several categories: MPEG-1 for recording images and sound on CDs (compact discs);

MPEG-2 for normal TV-size images up to HD (high-definition) images used in DVD and satellite broadcasts;

and MPEG-4 used extensively for low-bit-rate applications such as sending images via mobile phones. Moreover, the video coding standards laid down in VCEG are comprised of H.261 and H.263, which are mainly used for communication devices such as video-conferencing systems. The latest coding standard―H.264/AVC (advanced video coding)―has been drawn up by the JVT (JointVideo Team) under a collaboration between MPEG and ITU-T;

it is referred to as H.264 by ITU-T and MPEG-4 AVC by MPEG. Video coding technology has been designated new standards every several years or so, and with each re- standardization, coding performance has been improved. The latest coding standard―H.264/AVC― attains two to three times higher compression performance than MPEG-2 (used for DVDs). From now onwards, H.265 will be further improved and implemented in the latest technologies, thereby making it easy to enjoy even higher picture quality. These technologies (each shown in Fig. 2) form the base of so-called hybrid video coding. Already coded images are taken as reference pictures, and their movement is estimated. The difference between these imagesandtheremainingoriginalimagesiscompressed by means of a DCT (discrete cosine transform). Up until now, the coding performance of video coding technology has been improved with each new Fig. 2―Principle of Hybrid Coding. According to a motion compensation calculation and DCT, by transmitting only the difference of the frame in question with an already coded frame, high compression performance is attained. Time direction of frames Movement is predicted from an already coded reference image, and the size of the movement is coded as a motion vector. Lacking parts (difference between images) have their data compressed by means of a DCT. By transmitting only the difference with the previous frame, high compression performance is attained. Reference image Coded image DCT: discrete cosine transform Hitachi Review Vol.