Higher - Order Spectral Analysis of Isolated Word Speech Signal.

Authors

• Abdul Barry Raaof, Departement of Computer &Information Technology /University of Mosul

Document Type

Original Study

Abstract

Second-order cumulant and fourth-order cumulant analysis of isolated word speech signal are introduced and compared. Investigation of the non-Guassianity and nonlinearity of such signal is carried out, and the bispectral estimation is proposed for extracting more information beyond the power spectra. The work includes estimating the frequencies of the fundamentals and the harmonics and confirming the frequency coupling. A package is established using Matlab language for examining and analyzing any isolated word. This analysis finds applications in automatic speech recognition (ASR), voice control systems isolated word speech recognition, text-to-speech and speech-to-text proceesing. Spectral and statistical results are obtained.

How to Cite This Article

Raaof, Abdul Barry (2006) "Higher - Order Spectral Analysis of Isolated Word Speech Signal.," Iraqi Journal of Computers, Communications, Control and Systems Engineering: Vol. 6: Iss. 1, Article 7.
Available at: https://ijccce.researchcommons.org/journal/vol6/iss1/7