Research On Rate Control Algorithm Based On H.264/AVC

With the development of network and multimedia communication technology, the demand for video communication services is increasing. In order to obtain better quality of video transmission and broaden the application of video coding, digital video coding technology has been constantly advancing and perfecting. As the latest international video coding standard, H.264/AVC achieves great improvements in compression efficiency, images quality and adaptability of network. However, due to limited bandwidth of communication channel, it is necessary to control the bit rate output of video coding. Therefore, rate control plays a key role in the video encoder, which can ensure video images to code efficiently and transmit stably.Based on H.264/AVC video coding theory, this paper discussed the relevant technologies to realize rate control and the applications of rate-distortion theory in video encoder and rate control optimization, and emphatically analyzed the H.264/AVC rate-distortion optimization problems. Subsequently, the JVT-G012 proposal, recommended rate control algorithm by H.264/AVC, was further studied. The main work in this paper is as follows:Aimed at integer DCT coefficients'probability distribution characteristics of H.264/AVC encoder, an improved Cauchy-distribution-based rate-quantization (R-Q) model was proposed so as to describe more accurately the rate-distortion performance of video source. Through introducing mean squared error of energy (EMSE) as a measure of images distortion and analyzing statistics of extensive video sequences with different complexities, we developed a simply linear distortion-quantization (D-Q) model. According to the proposed models, the Lagrangian method was adopted to derive a bit allocation solution to the optimization of rate distortion.Combined with the improved Cauchy distribution model and the optimal rate-distortion bit allocation scheme, an optimized rate control algorithm for H.264/AVC macroblock layer was presented, which could improve rate distortion performance. Lagrangian multiplier method was adopted to calculate the optimized quantization parameter of macroblocks, and optimal bit allocation was realized by rate-distortion optimization strategy in all frames to make the output streams closer to target bits. Finally, target bits were jointly determined by the complexities of coded macroblocks, the distortion degree of images and buffer capacities. Experimental results demonstrate that the proposed algorithm can not only enhance the accuracy of rate control, decrease the average of bit error rate prediction, but also obtain better rate distortion optimization effect, and promote the PSNR as well.