A quality/energy tradeoff approach for IDCT computation in MPEG-2 video decoding

It would be desirable, in terms of energy conservation, to use a low complexity approximate algorithm to do all IDCT computation in an MPEG-2 decoder. However, there is a significant quality penalty associated with this approach that may not always be acceptable. A practical algorithmic method is presented here for achieving quality/energy tradeoff for a hardware 2-D IDCT in an MPEG-2 decoder. By allowing multiple algorithms to run on the same IDCT hardware, multiple quality/energy tradeoff modes can be supported, and the quality that results when an approximate algorithm is used can be increased. For example, by simply using an approximate algorithm for all B frame IDCT processing, while using a conventional algorithm for I and P frames, it is shown that a 23% energy reduction can be achieved compared to using the conventional algorithm exclusively. This significant energy reduction can be achieved with a quality tradeoff of less than 0.5 dB average PSNR