Advances on Video Coding Algorithms for Streaming Applications

Video streaming refers to a method for delivery of video services in which compressed video data are transmitted as a continuous stream to the receiver and played as they arrive. Therefore, a transmission channel with enough bandwidth for the continuous transmission of the bit stream should be assured. In video streaming, the quality of service depends on the video encoding process and the allocated bandwidth to the bit stream. The encoding process is very important, because it has a great impact on the rate-distortion performance and also on the utilization of different resources such as processing power, transmission bandwidth, and end-to-end delay of streaming service. In real-time video streaming, compromising the rate-distortion performance, computational complexity, bandwidth efficiency, and the quality of service makes the encoding process a challenge. The challenge can be more serious when the overall performance over a number of streaming services is considered. The specifications of most video coding standards define only the bit-stream syntax and the decoding process. The encoding process is not standardized to allow flexible implementations. This thesis proposes algorithms for the encoding process with emphasis on rate control for streaming applications. In this thesis, novel fuzzy logic controllers have been developed and used in several video encoding and streaming scenarios. Video rate control algorithms for encoding variable bit rate video are proposed where the fuzzy controllers are deployed. The proposed algorithms in the new approach are very different from the conventional rate control algorithms in the way they operate. However, some theoretical and practical results of conventional rate control are used in the new approach. Furthermore, video encoding techniques optimized for streaming over DVB-H (Digital Video Coding for Handheld) are proposed in this thesis. The contributions of thesis are presented in three parts. Each part includes summary of a number of published papers and a selected subset of published papers is attached to the thesis. In the first part of the thesis, video rate control algorithms and tools for video streaming applications are proposed. The proposed rate control algorithms can be applied to other variable bit rate video applications such as recording on storage media. In the second part, video encoding and rate control algorithms are proposed and optimized for video streaming over DVB-H channels. The third part presents video encoding algorithms for a digital video broadcast system in which a number of video sources are encoded and broadcasted simultaneously. The proposed rate control algorithms in the thesis outperform sophisticate conventional rate control algorithms in term of rate-distortion performance and computational complexities. Moreover, they maximize utilization of the allocated resource to increase the overall quality of service. Furthermore, the proposed algorithms are very flexible so that they can be easily tuned for different operating points and applications.