Network congestions can occur easily during video streaming sessions. This is caused by either variations of the available network bandwidth, or traffic bursts of the video stream itself. Hence, an encoder need employ rate control as a way to regulate varying bit rate characteristics of the coded bit-stream in order to produce high quality decoded frame at a given target bit rate. For live video streaming, the target bit rate is the actual available bandwidth. For off-line encoding, the target bit rate is bounded. The main objective of rate control is to balance the trade-off between the bit-rate of the video stream and its quality loss by determining the optimal choice of encoding parameter settings. Traditional rate control algorithm is limited to single one quantification parameter, although there are many other options available, such as GOP structure, chrominance sampling format, picture resolution and frame rate etc.

This thesis will investigate the relationship between network layered QoS parameters (throughput, packet loss rate) and multiple encoding parameters related to user perceptive quality. More challenge task is to search for optimal encoding configurations targeting the highest overall rate-distortion performance.

Prerequisites: English, excellent programming skills, C or C++, Fundamentals of multimedia, Computer networks