Study On The Improved Two-Pass Algorithm For The Rate Control In HEVC

With the develop of video coding technology, the mainstream video gradually becomes high-definition and ultra-high definition video, in order to achieve higher compression performance, JCT-VC Joint Coordination Group released High Efficiency Video Coding(HEVC) standard, the compression efficiency of which is twice than H.264/AVC. In the practical application of video coding standards, ensuring the compressed video quality under the limited bandwidth, that is rate control(RC), is a key module of the video encoder, and is one of the hot research issues in the field of video compression technology.In this paper, the algorithms of rate control for HEVC are deeply researched with a focus on the improvement of rate control precision and video compression quality under the target bit rate. HM standard test model is the model which has been widely adopted for the research on HEVC video compression encoding technology. Through the study and analysis of the JCTVC-K0103 classic rate control algorithm in HM model, it is found that the complexity measurement baseline of the K0103 algorithm is not accurate. At the same time, the classic R-? model in the K0103 algorithm has unified parameters, which can not fit for every video sequences. These problems are likely to cause the overflow or underflow of the buffer, and lead to a decrease of video compression quality. To solve the above problems, this paper proposes the following improvements and implementations of the rate control algorithm for HEVC:(1) The realization of Two-Pass rate control algorithm based on HM model. In this algorithm, the number of bits obtained from the First-Pass pre-coding process was used as the benchmark for the complexity of the current coding unit, which improves the accuracy of the complexity measurement baseline. Experimental results show that, compared to the latest HM16.0 standard rate control algorithm, the PSNR of our Two-Pass model has greatly improved, especially for the situation under low target bit rate, but the rate control accuracy has declined.(2) The realization of an improved Two-Pass rate control fast algorithm. In order to cure the fall rate control accuracy in the Two-Pass basic algorithm implemented above, this paper proposes two improvements based on the Two-Pass basic model above. First, inorder to solve the problem that the R-? model in the Two-Pass basic algorithm has unified model parameters which is not fit for every video sequences, this paper proposes to train the model parameters online as an improved method; secondly, in order to avoid the overall rate control coding time multiplied, we propose a fast algorithm for the First-Pass, that is, do pre-coding only on the third layer and collect relevant statistical data. After a lot of experimental tests, it is found that compared to HM16.0, the PSNR of the improved Two-Pass fast algorithm increases by an average of 0.42 d B, the rate control accuracy improves by an average of 4.81%, furthermore, the total encoding time of the First-Pass reduces 47.35% compared to the Two-Pass basic model.In this paper, the algorithms of rate control for HEVC are deeply researched, faced the imbalance in the complexity measurement and the inaccurate model parameters in the current rate algorithms, two improved algorithms are proposed and implemented based on the HM platform in this paper. The experiment results show that the algorithms proposed effectively improve the video quality under target rate and enhance the accuracy of rate control, meanwhile, the improved Two-Pass fast algorithm greatly reduce the First-Pass encoding time. In summary, the two rate control algorithms proposed in this paper greatly improves HEVC encoding performance and coding efficiency under the given bandwidth conditions, which has great significance for the design of HEVC encoder, and has broad application prospects.