(A) new rate control algorithm of MPEG video transcoder based on the bit-rate reduction characteristics in requantization

Abstract

Video transcoding techniques offer the possibility of matching coded video to transmission channels of lower capacity by reducing the bit-rate of compressed bitstreams. Applications such as video on demand (VOD) and multicasting of video to heterogeneous networks motivate the idea of transcoding. The performance of transcoding is evaluated by two factors; the accuracy on the target bit-rate and the complexity of hardware implementation. In this thesis, we propose a new rate control algorithm which has an accurate bit-rate control characteristics with a small computational complexity. For the rate control method, we empirically observe the relationship between the quantization step size and the generated bits in the requantization process of the MPEG coded video sequence. It is found that the regenerated bit-rate can be characterized as a piecewise linear function of requantization step size. This property provides an accurate quantization step size for new target bit-rate. For the complexity problem, we reduce the computational complexity of target bit allocation drastically and remove the need of additional computation for adaptive quantization. It is shown by simulation that the proposed method gives much better performance in accuracy and picture quality than conventional rate control method.