Higher-order statistics-based non-linear speech analysis

J.J. Soraghan, A. Hussain, A. Alkulabi, T.S. Durrani

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)


A fast and robust three-level binary higher order statistics (HOS) based algorithm for simultaneous voiced/unvoiced detection and pitch estimation of speech signals in coloured noise environments with low SNR is presented. The use of the three-level binary speech signals dramatically reduces the computational effort required in evaluating the higher order cumulants. The superior performance of the new algorithm over the conventional autocorrelation method using real speech signals is demonstrated. The algorithm can easily be implemented in digital hardware using simple combinatorial logic.
Original languageEnglish
Pages (from-to)11-18
Number of pages7
JournalControl and Intelligent Systems
Issue number1
Publication statusPublished - 2002


  • higher order statistic
  • autocorrelation
  • statistical analysis
  • non-linear effect
  • speech analysis
  • vocal signal
  • speech processing


Dive into the research topics of 'Higher-order statistics-based non-linear speech analysis'. Together they form a unique fingerprint.

Cite this