Spatial Characteristic Based Rate Control Initialization Method for Conversational Video Communication of H.264/AVC Using Linear Prediction

In this paper, we propose a method of initializing the quantization parameter (QP) in H.264 rate control algorithm based on spatial characteristic of conversational video. Primarily, we choose the fourteen various conversational videos and calculate the complexity of the macroblock (MB) using variance of MB. The spatial characteristic of each video can be quantized by the proportion of the number of complex MB in the first frame. And then we find their optimal initial QPs which can guarantee that video sequences have consistent quality by minimizing the variation of QPs in a GOP, while ensuring the minimizing actual encoding bit rate closer to the target bit rate over various low bit rates. And then we propose the fitted models of optimal initial QPs against spatial characteristic over different low bit rates. The proposed model can be used to determine optimal initial QP to guarantee consistent quality and to meet target bit rate simultaneously. The objective and subjective simulation results show that the proposed method achieves better balance quality than JM10.2 and previous research result.