Power reduction via macroblock prioritization for power aware H.264 video applications

As the importance of multimedia applications in hand-held devices increases, the computational strain and corresponding demand for energy in such devices continues to grow. Portable multimedia devices with inherently limited energy supplies face tight energy constraints and require optimization for energy conservation. Power-aware applications give their users flexibility to prioritize and trade between performance and battery-life. This paper introduces a power-aware technique for user selectable power reduction in exchange for controlled reductions in video quality for H.264 video streams. The technique uses an encoder-decoder pair. The encoder characterizes video streams and provides information to the decoder via Flexible Macroblock Ordering (FMO) by generating prioritized slice groups. The decoder selectively ignores low priority slice groups based on user selected preference effectively reducing the decoder workload. With a reduced computational requirement, processor voltage an...

• Type: Conference
• Year: 2008
• Where: CODES
• Authors: Michael A. Baker, Viswesh Parameswaran, Karam S. Chatha, Baoxin Li
• Keywords: CODES 2008, Decoder, Decoder Power Performance, Frequency Scaling, Software Engineering