Energy Optimization for Latency- and Quality-Constrained Video Applications

The trade-off between energy consumption and media quality is a fundamental design issue in mobile multimedia. This energy optimization technique based on frame skipping and buffering exploits the characteristics of video applications, particularly their tolerance to variation in latency and video quality to increase slack time and its use in dynamic voltage scaling. The technique adapts the frame rate to the degree of motion activity by skipping redundant frames. Thus, it intentionally creates slack time as long as video quality is not unacceptably degraded. We implemented this technique with only a slight modification of the original encoder algorithm and no modification to the decoder.