H.264 bit-rate control using the 3-D perceptual quantization modeling

Bit-rate control has a critical influence in video coding and multimedia streaming. This paper proposes a novel H.264 bit-rate control using a 3-D perceptual quantization modeling (PQrc), including two major encoding modules: the perceptual frame-level bit-allocation and the fast macroblock-level quantizer decision. The frame-level budget bit depends on the frame complexity (c) and buffer fullness, in which c is weighted by the predicted mean-absolute-difference (MAD) and the just-noticeable-difference (JND) PSNR. Considering the MB-level quantizer decision, the 3-D bits-complexity- quantization, which is denoted as B.C.Q., model is established, in which the B.C.Q. curve's tangent slope is a piece of unique information to find a proper quantizer. In comparison with the latest H.264 JM10.2, our experiment results show that the proposed PQrc can improve the SNR quality and keep the stable buffer fullness with less computational cost.