Video quality balance algorithm between source and channel losses using H.264 packet loss visibility model

• Main Authors: Hsin-Chin Lee, 李信瑾
• Other Authors: Ting-Lan Lin
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/y8a97a

碩士 === 中原大學 === 電子工程研究所 === 102 === In the multimedia network, transmission is increasingly popular. Video transmissions are becoming more important. Topics of video transmission include bit rate control encoding (bandwidth) and encoding quality. In the transmission among the general Internet, packet losses will be another important topic. Usually in H.264 coding standard, we can use the mode decision in the standard to balance bit rate and a distortion; this emphasizes the "source distortion". However, the standard does not focus on how to optimize the possible "channel distortion", that is the relationship between the packet-loss distortion and bit rates. In this paper, we propose the use of Packet loss visibility model to measure the possible distortion of the channel, and use the packet loss visibility value in the standard mode decision, to perform the joint Rate distortion optimization algorithm. With this modification, it is possible to take into account the source distortion, bit rate and channel distortion, to optimize the balance among the three. Our experimental results show that, with the weight adjustment between the source distortion and channel distortion in the optimization, we can successfully adjust the final proportion of the overall video quality distortion between the source distortion and channel distortion, and can ultimately enhance the overall video quality.