An analytic parahermitian matrix admits an eigenvalue decomposition (EVD) with analytic
eigenvalues and eigenvectors except in the case of multiplexed data. In this paper, we …

A voice activity detection (VAD) algorithm identifies whether or not time frames contain
speech. It is essential for many military and commercial speech processing applications …

Speech enhancement is important for applications such as telecommunications, hearing
aids, automatic speech recognition and voice-controlled systems. Enhancement algorithms …

The paper provides a comprehensive overview of speech enhancement techniques and
their applications. It discusses challenges in non-stationary noise, reverberation, and …

The polynomial power method repeatedly multiplies a polynomial vector by a para-
Hermitian matrix containing spectrally majorised eigenvalue to estimate the dominant …

Extracting analytic eigenvectors from parahermitian matrices relies on phase smoothing in
the discrete Fourier transform (DFT) domain as its most expensive algorithmic component …

The second order sequential best rotation (SBR2) algorithm is a popular algorithm to
decompose a parahermitian matrix into approximate polynomial eigenvalues and …

The enhancement of noisy speech is important for applications involving human-to-human
interactions, such as telecommunications and hearing aids, as well as human-to-machine …

Speech enhancement algorithms using polynomial matrix eigenvalue decomposition
(PEVD) have been shown to be effective for noisy and reverberant speech. However, these …

Calculating the eigenvalues and eigenvectors of a polynomial matrix has proved to be an
important problem in signal processing, in recent years. There exists various iterative …