Dynamic bit-rate estimation and control for constant-quality communication of video

In video communications, the complexity of Image contents, the available bit-rats for communication, end the quality of video are all closely related. This paper describes a new quality control scheme for variable-bit-rate (VBR) video encoder. The Bayesian estimation approach is used to establish a bit-rats prediction model. The model defines the required bandwidth for a future frame according to prior information and performs a posterior estimation after the bit-rate of the new frame is determined. When the estimated bit-rats exceeds a certain threshold, adjustment to the quantisation parameters has to be made such that the quality of the video is kept constant.