Two-Pass Rate Control for Improved Quality of Experience in UHDTV Delivery

Abstract

Rate control plays an important role in any video 5 coding application and it was extensively studied in the context 6 of previous video coding standards. However, the current state-of-7 the-art high efficiency video coding (HEVC) standard introduces 8 many flexible tools making previous rate-distortion models used 9 in rate control insufficiently accurate. Recently, a few rate control 10 methods have been developed for HEVC that introduce many 11 useful features, such as a robust correspondence between the rate 12 and Lagrange multiplier λ. Nonetheless, previous rate control 13 algorithms for HEVC do not address typical content in television 14 applications that consists of frequent scene changes. Furthermore, 15 the new ultra high definition television (UHDTV) format, which 16 is expected to become widespread in the future, demands for 17 even higher compression efficiency. To overcome these issues, a 18 two-pass rate control method is proposed in this paper, targeting 19 the encoding of UHDTV content. In the first pass, a fast encoder 20 with limited set of coding tools is used during pre-encoding step 21 to obtain the data used for rate allocation and model parameter 22 initialization, which will then be used during the second pass. To 23 avoid multiple encoding steps when deriving this information, 24 a variable quantization parameter framework is proposed. 25 Experimental results show that the proposed rate control method 26 outperforms the well-known HEVC rate control method. When 27 compared with variable bit-rate encoding mode, the proposed 28 two-pass rate control method achieves on average 2.9% BD-rate 29 losses. That is significantly better than the state-of-the-art HEVC 30 rate control method, which achieves an average 8.8% BD-rate 31 loss. The proposed method also provides a more consistent quality 32 fluctuation with time, measured with standard deviation of frame 33 PSNR values, required for high Quality of Experience. Q1 Q2 34 Index Terms-HEVC, quality of experience, rate control, UHD 35 video, video streaming. 36 I. INTRODUCTION 37 U LTRA high definition television (UHDTV) is the new for-38 mat which is expected to deliver a greater impact, more 39 presence and immersion than the current high definition tele-40 vision (HDTV). UHDTV is not just about more pixels but it 41 has the potential to deliver wider color gamut, high dynamic 42

Rate Estimation Techniques for Encoder Parallelization

“…The QP value is determined based on the Sum of Absolute Transformed Difference (SATD) with four-dimensional first-pass statistics. In [6], based on the R-model, Zupancic et. al.…”

Implementation-Free Forensic Watermarking for Adaptive Streaming with A/B Watermarking

“…Moreover, they can also change (the granularity of) other coding deci-sions such as the block partitioning and motion vectors. In adaptive-streaming applications, this is typically done using a two-pass encoding [19]. In a first pass, the content is analyzed and the result is stored in a log file.…”

Complexity-driven rate-control for parallel HEVC coding

“…Temporal complexity was also used [12]. In the context of R-λ models, a two-pass RC scheme was proposed [13], highlighting the importance of having content-dependent initial parameters. Unfortunately, two-pass approaches may not be suitable for some applications in which low complexity is crucial.…”