Rate Control For Video Coding Using A Low-Memory-Cost Look-Up Table

Video coding and transmission techniques have become important in multimedia communication over the network. In the last few years, many video compression international standards are proposed, such as MPEG and H.26x standards. They are broadly used in digital broadcast TV standards, digital video disk (DVD), etc. After compression, the size of video data is largely decreased and the network bandwidth utilization is further reduced.The special techniques used in video coding result in a variable-bit-rate bitstream after compression. Unfortunately, in video communications, these compressed data are most likely to be delivered using constant-bit-rate (CBR) channels so as to avoid the network fluctuation. Thus, after video encoding, the compressed data are put into an encoder buffer, then the output data from the buffer are delivered using CBR channels;once these data are received by the video decoder, they are put into a decoder buffer, then the decoder fetchs these data from the buffer with a variable bit rate for decoding. As the buffer overflow or underflow may occur, rate control at the encoder is necessary to regulate the bitstream, so that the buffer exceptions can be avoided.Rate control plays an important role in the video coding process. It is introduced to regulate and control the output bit rates of video sources in the network to achieve the best trade-off between quality and bandwidth utilization. In the last few years, many rate control algorithms are proposed. Classic algorithms include the rate control for H.263 low-delay video communication, the scalable rate control for MPEG-4 video, etc. These rate control algorithms often employ rate-distortion (R-D) models, which involve complex floating-point computations. Thus, the computing complexity is high. Furthermore, the R-D models are obtained under the assumption of source statistics, so they are not always correct in the actual applications. In order to solve these problems, the paper proposes a new rate look-up table based rate control algorithm. The memory cost of the algorithm is low, and only some simple computations are involved, while the bit estimation and QP determination of the algorithm are still very well. Besides, in the proposed algorithm, an unequal frame bit allocation method is proposed for the GOP pattern in video coding. This method can achieve smoother video quality variations, especially when scene changes or high motions occur. After optimizations both in space and time, the proposed algorithm can be implemented in the video codec of mobile devices. Experimental results show that, the proposedalgorithm achieves better video quality, reduces the quality variation, and also retains the bit rate proximity and buffer regulation.