• DocumentCode
    3080044
  • Title

    Analyzed rate distortion model in standard video codecs for rate control

  • Author

    Rezaei, Mehdi ; Gabbouj, Moncef ; Wenger, Stephan

  • Author_Institution
    Inst. of Signal Process., Tampere Univ. of Technol., Finland
  • fYear
    2005
  • fDate
    2-4 Nov. 2005
  • Firstpage
    550
  • Lastpage
    555
  • Abstract
    In this paper we propose a general and precise rate-distortion (RD) model for standard video codecs, including MPEG-4, H.263 and AVC H.264. The model can be used in almost all video rate control algorithms for different applications, from one-pass to multi-pass, and from constant rate to constant quality. It is designed taking into account previous theoretical results, and based on new assumptions which are confirmed by comprehensive practical experiments. We conceptually divide the RD model into two separate parts. The first part is codec specific and reflects the codec´s properties. The second part is video content specific and describes the interaction between the video content and the encoder. This division provides several advantages. The model performs more precisely but needs less updating efforts. Furthermore, in this paper we introduce a simple complexity measure for the video content. As an example, we parameterized the RD model by the complexity measure for the estimation of intra frame bits. This example can be used for complexity estimation in a start frame and in the scene cuts. The experimental results show that both the RD model and complexity measure perform well, with high precision as estimation tools.
  • Keywords
    rate distortion theory; video codecs; video coding; AVC H.264; H.263; MPEG-4; complexity estimation; intra frame bits estimation; rate distortion model; standard video codecs; video content; video rate control algorithms; Automatic voltage control; Code standards; Distortion measurement; Layout; MPEG 4 Standard; Rate-distortion; Signal processing algorithms; Video codecs; Video compression; Video signal processing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing Systems Design and Implementation, 2005. IEEE Workshop on
  • ISSN
    1520-6130
  • Print_ISBN
    0-7803-9333-3
  • Type

    conf

  • DOI
    10.1109/SIPS.2005.1579928
  • Filename
    1579928