Article Preview
TopIntroduction
The increasing popularity of high definition TV, mobile phone, digital camera, surveillance, broadcasting and so on stimulates the rapid evolution of multimedia technologies for meeting the demand of high quality and low delay while operating the intensively complicated processes. Since videos involved plenty of useful information gradually dominates the consumer market, the corresponding coding technologies become critical for effective storage or transmission, which has been gaining a wide interest in research and industrial communities.
Due to its high coding efficiency, the video codec of H.264/AVC (Wiegand et al., 2003) has been widely employed by industry. However, with the ever-growing diversification of video applications, e.g. large-screen digital video, mobile entertainment, the conventional codec of H.264/AVC is facing severe challenges in coding efficiency (Video and Requirements SGs, 2008). So the latest codec of H.265/HEVC (Sullivan et al., 2012) has been developed for achieving a significant improvement in coding performance, where numerous advanced tools, such as coding-tree block structures, advanced motion vector prediction, unified directional intra prediction and so on, are adopted and contributed for high efficiency compression.
The video codec of H.265/HEVC employs the same “hybird” approach (intra/inter prediction, transform, quantization, entropy coding) used in compression standard H.264/AVC. Figure 1 depicts the block diagram of H.265/HEVC, where each candidate coding mode is firstly processed by a sequence of modules to achieve the distortion and bit-rate. Then the optimal option modeopt can be obtained by calculation and comparison of the Rate-Distortion (RD) cost (Sullivan et al., 1998):
,
(1) where the
modei denotes the
ith element in the candidate set
C, the parameter of
λ represents the Lagrange multiplier. The symbols of
D and
R indicate the coding distortion and corresponding bits.
Figure 1. The coding structure of H.265/HEVC
However the intensive computational complexity is simultaneously introduced into H.265/HEVC, which may cause a bottleneck for the actual application. It is known that videos relate to abundant data in storage or transmission, especially for the ultra high resolution services, which may probably make the task of low-delay coding difficult and eventually lead to users’ annoyance and dissatisfaction. Therefore, efficient and fast video coding scheme is indispensable for H.265/HEVC to conform the requirement of low-delay application.
In this paper, an effective rate estimation algorithm for H.265/HEVC encoder is developed and incorporated into the rate distortion optimization (RDO) for fast mode decision. The syntax elements of residuals and the principle of entropy coding are first analyzed, and then, a highly accurate and efficient model is constructed to skip the actual entropy coding of CABAC for the calculation of RD cost in mode decision. Experimental results show that the CABAC encoding time for residuals can be reduced by 68% in average using the proposed fast scheme, with negligible degradation of coding performance.
The remainder of this paper is organized as follows. In the next section, we further review related work. Then we give a brief illustration of the syntax elements regarding residuals. Afterwards we propose a fast bit-rate estimation algorithm for significantly accelerates the encoding process while well maintaining the video quality. For comparison and verification of the proposed method, finally we provide the experimental results, followed by the conclusion.