• DocumentCode
    3582710
  • Title

    Exploiting MB level parallelism in H.264/AVC encoder for multi-core platform

  • Author

    Asif, Muhammad ; Majeed, Saqib ; Taj, Imtiaz A. ; Bin Ahmed, Maaz ; Ziauddin, S.M.

  • Author_Institution
    Mohammad Ali Jinnah Univ., Islamabad, Pakistan
  • fYear
    2014
  • Firstpage
    125
  • Lastpage
    130
  • Abstract
    H.264/AVC video coding standard has achieved a significant improvement in coding efficiency over previous standards, such as, H.261, H.263 and MPEG-4, at the cost of computational complexity. This paper takes advantage of the macroblock (MB) level parallelism in H.264/AVC encoder and based upon this represents a scheme to implement the encoder for multi-core platform. According to this technique an encoding process of an MB is divided into several independent phases and each phase is executed on a separate processing core. Moreover to efficiently utilize the multi-core processors and to improve the encoder performance, the encoder is pipelined at MB level. The experimental results show that encoding rate of pipelined encoder as compare to sequential encoder is improved from 6 fps to 72 fps and from 2 fps to 32 fps for FD1 (720 × 480) and HD 720p (1280 × 720) resolutions, respectively. The implemented encoder can be used for multimedia applications like video conferencing, smart phones, high quality mobile digital video recorders, tablet computers, mobile digital TVs etc.
  • Keywords
    codecs; computational complexity; telecommunication standards; video coding; FD1; H.261; H.263; H.264/AVC encoder; HD 720p; MPEG-4; computational complexity; high quality mobile digital video recorders; macroblock level parallelism; mobile digital TV; multi-core platform; multi-core processors; multimedia; sequential encoder; smart phones; tablet computers; video coding standard; video conferencing; Encoding; Graphics processing units; Hardware; Multicore processing; Parallel processing; Streaming media; Video coding; H.264/AVC; JZ4770; Multi-core; Pipelining;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer Systems and Applications (AICCSA), 2014 IEEE/ACS 11th International Conference on
  • Type

    conf

  • DOI
    10.1109/AICCSA.2014.7073188
  • Filename
    7073188