• DocumentCode
    3004379
  • Title

    Proper scalings in recursive lattice digital filters for minimum round-off noises

  • Author

    Lei, Shaw-Min ; Yao, Kung

  • Author_Institution
    Dept. of Electr. Eng., California Univ., Los Angeles, CA, USA
  • fYear
    1988
  • fDate
    11-14 Apr 1988
  • Firstpage
    1810
  • Abstract
    A systolizable digital filter is considered which is capable of achieving a high-performance with respect to good finite-word-length effects as well as high-throughput and is also suitable for VLSI implementation. A simple method to compute the average power at each internal node and a method to compute the output roundoff noise in the proposed digital filter are presented. Based on this approach, for a given transfer function, the authors consider the design of a minimum roundoff noise filter with the proposed systolizable structure under the constraints that the signal powers of internal nodes are well scaled. Some results for this problem are obtained by using a local optimization technique. Specific numerical examples show this approach yields less roundoff noise as compared to the Rao-Kailath´s orthogonal digital filters
  • Keywords
    digital filters; filtering and prediction theory; noise; optimisation; transfer functions; Rao-Kailath´s orthogonal digital filters; VLSI implementation; constraints; finite-word-length; high-throughput; local optimization; minimum roundoff noise filter; proper scalings; recursive lattice digital filters; systolizable digital filter; transfer function; Digital filters; Finite impulse response filter; IIR filters; Lattices; Signal design; Signal processing; Signal to noise ratio; Throughput; Transfer functions; Very large scale integration;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics, Speech, and Signal Processing, 1988. ICASSP-88., 1988 International Conference on
  • Conference_Location
    New York, NY
  • ISSN
    1520-6149
  • Type

    conf

  • DOI
    10.1109/ICASSP.1988.196973
  • Filename
    196973