• DocumentCode
    2287744
  • Title

    DCELP: a low bit rate and low delay speech coding method

  • Author

    Sabbarwal, A. ; Jandhyala, Vikram ; Prasad, Surendra

  • Author_Institution
    Dept. of Electr. Eng., Ohio State Univ., Columbus, OH, USA
  • fYear
    1994
  • fDate
    13-16 Apr 1994
  • Firstpage
    476
  • Abstract
    Methods for obtaining high quality speech at very low bit rates have become very important in the digital communication environment. Good quality speech has been reported at bit rates as low as 2.4 kb/s. Apart from the bit rate, another important consideration is the delay due to the coding methods used. For applications like digital telephony, a one way delay of less than 2 ms is required for ease in echo cancellation. Low delay methods have been developed; however, these work at much higher bit rates than methods with high delay. The authors present a method that can be used to obtain good quality speech at a bit rate as low as 2.66 kb/s and a delay of less than 2 ms. The method, termed differential code excited linear prediction (DCELP) is a CELP based method differing from other CELP variations in that a low delay version of CELP is used to code the error signal rather than the actual speech signal. DCELP is a vector based system comprising two main blocks, a vector predictor and CELP acting as a vector quantizer
  • Keywords
    channel capacity; delays; digital communication systems; linear predictive coding; speech coding; vector quantisation; 2.66 kbit/s; DCELP; coding methods; differential code excited linear prediction; digital communication environment; digital telephony; echo cancellation; error signal; low bit rate coding method; low delay speech coding method; vector based system; vector predictor; vector quantizer; Bit rate; Delay; Digital communication; Image processing; Neural networks; Phase change materials; Pulse modulation; Speech coding; Speech processing; Vectors;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Speech, Image Processing and Neural Networks, 1994. Proceedings, ISSIPNN '94., 1994 International Symposium on
  • Print_ISBN
    0-7803-1865-X
  • Type

    conf

  • DOI
    10.1109/SIPNN.1994.344867
  • Filename
    344867