Scalable video transportation using look ahead scheduling

This thesis introduces many video streaming algorithms and techniques to improve the performance of scalable video dissemination over a packet network. As the quality of visual content in a best-effort network is susceptible to average bandwidth availability, packet loss, packet delay and packet delay variation (jitter), the study proposes a scheduling algorithm and buffering technique. To combat the variation of delay (jitter) in a best-effort network, the study proposes Look Ahead Scheduling Algorithm (LASA) which provides unequal look-ahead by safeguarding the base layer of the scalable video. The results of the experiment show that LASA reduces the number of frames with violate deadlines and significantly improves the continuity of video stream without compromising the average Y-PSNR. The LASA is also extended to support muti-server scalable video streaming. The performance evaluation is investigated by using a network simulation tool. This thesis also presents the weighted multi-playback buffer management system for scalable video streaming in a congested network. The approach intrinsically provides weighted protection to the more important video layers within a video stream. The study compares the performance of a weighted scheme with a multi-playback buffer which gives equal protection to every video layer. The simulation results show that weighted multi-playback buffer management can improve the average Y-PSNR and the continuity of the scalable video stream.