A quality measure-based rate control strategy for MPEG video encoders

Rate control is an essential part of most video encoders. In this paper, a rate control scheme based on block based-features which has a high correlation with subjective test experiments is presented. Four block-based features (the average of DFT differences, the standard deviation of DFT differences, the mean absolute deviation of wepstrum differences, and the variance of UVW differences) are extracted from the input and output video sequences and fed into a four-layer back-propagation neural network that has been trained by subjective testing. The bit assignment between frames and macroblocks is selected based on the nonconvex quality and bitcount functions and a subjective activity measure which is computed as the average of the nonconvex quality bitcount ratios over the mquant index ranges [1...31]. The proposed rate control scheme has a higher, and also a more consistent, quality measure than the MPEG Test Model 5 scheme. The objective quality, as measured by the signal-to-noise ratio is also improved, and the variations in the signal-to-noise ratio and bit-rate from frame to frame are reduced