Robust determination of periodic correlation of speech signals using empirical mode decomposition and higher-order spectra

This paper presents a new method of periodic/non-periodic (P/nP) classification of noisy speech signals. Empirical mode decomposition (EMD), a newly developed tool to analyze nonlinear and non-stationary signals is used to filter the additive noise with the speech signal. The normalized autocorrelation of the filtered speech signal is computed to enhance the periodicity of the analyzing speech signal if any. It is considered that the voiced speech (with fundamental periodicity) signal is periodically correlated and the unvoiced signal is not. A noise robust P/nP decision rule is formulated based on third-order autocumulants of the autocorrelation function of speech signal. The experimental results show that the use of EMD improves the classification performance and the overall efficiency is noticeable as compared to other existing algorithms.