| Most visual communication applications use a fixed rate transmission channel, which means the encoder's output bit rate must be regulated to meet the transmission bandwidth. It is necessary to build the rate control scheme according the application requirement within the given target bit rate to transmit the coded data effectively and the guaranteed steady received quality. Although H.264/AVC (MPEG-4 Part 10) has achieved the best coding efficiency among all coding standard, it can not handle the rate control mechanism well. It uses quantization parameter (QP) both in rate control and Rate-Distortion optimization (RDO), resulting in the chicken-egg dilemma. Thus it is difficult to realize an efficient rate control scheme.In this thesis, a new rate control Algorithm based on novel coding complexity metric and the relative improvement scheme is discussed. We find the chicken-egg dilemma can be avoided in macro-block (MB) layer coding. Meanwhile, it is used the mean absolutely difference (MAD) to denote the coding complexity and updated with the linear recession technology, which can not denote the coding complexity perfectly, especially in the situation of scene change. To modify this drawback, a new coding complexity metric called average pixel information bit (APIB) is proposed. This parameter can be obtained before quantization and perfectly reflect the image coding complexity of each frame and MB. The bit allocation scheme according the global complexity of coded frame is presented. In MB layer, we allocate bit for each MB according to the spatial and temporal complexity, then calculate QP and update the Rate-Distortion function. Furthermore, this paper proposes a new detected method and approach to handle with the scene change. The simulate results show that compared to H.264/AVC at the same condition, the proposed method gains more accurate control precision and higher visual quality; and controls the fluctuation on bit and PSNR more precisely at the same time. |
PDF Full Text Request