BROWSE
Detail
Conference Paper
Implementation of Accelerated 4K HEVC Encoder using SIMD Optimization and Data-Level Parallelism
Cited - time in 
Share 
Share
- Authors
- Dongsan Jun, Sung-Chang Lim, Jinho Lee, Hahyun Lee, Jongho Kim, Jungwon Kang, Jinwook Seok, Younhee Kim, Soonheung Jung, Hui-Yong Kim, Jin Soo Choi
- Issue Date
- 2016-01
- Citation
- International Workshop on Advanced Image Technology (IWAIT) 2016, pp.1-4
- Language
- English
- Type
- Conference Paper
- Abstract
- High Efficiency Video Coding (HEVC) is the latest video coding standard which had been developed by Joint Collaborative Team on Video Coding (JCT-VC) formed by the ITU-T Video Coding Experts Group (VCEG) and the ISO/IEC Moving Pictures Experts Group (MPEG). To achieve improved coding efficiency, HEVC adopted various advanced coding tools. Even though the coding efficiency is highly increased compared to previous video coding standard, the computational complexity of encoder is very heavy. Due to the dramatically increased complexity, it has been difficult to develop the commercial HEVC encoder such as ultra-high resolution (UHD) dedicated live and/or VoD service. Based on the analysis of the computational complexity of HEVC encoder, we implemented both single instruction multiple data (SIMD) operations and tile based datalevel parallel processing on multi-core platform. In addition, we investigated optimal encoding parameters for encoder simplicity and presented several fast encoding algorithms. Experimental results show that the proposed HEVC encoder significantly reduces the computational complexity while maintaining reasonable RD performance compared to HEVC reference model (HM) ver. 14.0.
- KSP Keywords
- Coding Tools, Coding efficiency, Computational complexity, Data-level parallelism, Fast encoding, HEVC encoder, ITU-T, Joint Collaborative Team on Video Coding(JCT-VC), Parallel Processing, Reference Model, Single Instruction Multiple Data(SIMD)