Design of statistically based buffer control policies for compressed digital video

The authors present an investigation of statistically based approaches to the design of buffer-control algorithms for interfacing a compressed digital video source to a constant rate channel. Since compression techniques such as DPCM (differential pulse code modulation) or transform coding result in variable bit rate encoded video, rate buffering and adaptive encoding are often used to maintain high average image quality, while avoiding buffer overflow events. As a first step to a quantitative methodology for the design and evaluation of adaptive buffer-control algorithms, an example adaptive-spatio-temporal/DPCM source is statistically characterized in terms of long-term rate distributions for each potential encoding mode (which are useful for general mode selection), along with more detailed intermode and intramode rate statistics (which provide useful information for mode switching). Two specific adaptive mode control algorithms based on these statistical characterizations are proposed: (a) the expected rate policy line algorithm, driven by conditional intermode expectation statistics, and (b) the constant overflow probability algorithm driven by conditional intramode distributions. Simulation results comparing the proposed algorithms with a conventional buffer level based control heuristic are presented, demonstrating the potential for substantial performance improvement at the expense of only a modest increase in encoder complexity.