National Natural Science Foundation of China (No.U19A2052, No.61571102);Key Program of Department of Science and Technology of Sichuan Provicne (No.2018JY0035);Supported by the Special Basic Cooperative Research Programs of Yunnan Provincial Undergraduate Universities Association (No.2018FH001-056)
Rate-distortion optimization (RDO) can effectively improve the compression performance for encoders
which is an important topic in video coding. In this paper
the Lagrange multiplier based RDO method is reviewed comprehensively from several aspects including independent RDO
dependent RDO
bit allocation for rate control
etc. First
an example is employed to explain how the constrained optimization problem is converted into an unconstrained optimization problem based on the Lagrange multiplier method
and the relationship between Lagrange multiplier and quantization parameter is analyzed according to relevant references. Second
the rate-distortion dependencies in the process of video coding are analyzed
and the dependent RDO methods presented in recent years are classified and discussed. Third
the effect of the rate-distortion dependencies on bit allocation for rate control is discussed. Finally
the features and performance of different RDO methods are tabulated for analysis and comparison
and the existing challenges and the research works deserving further exploration are presented.