An analytic parahermitian matrix admits in almost all cases an eigenvalue decomposition
(EVD) with analytic eigenvalues and eigenvectors. We have previously defined a discrete …

Multi-channel signals captured by spatially separated sensors often contain a high level of
data redundancy. A compact signal representation enables more efficient storage and …

This article is devoted to the polynomial eigenvalue decomposition (PEVD) and its
applications in broadband multichannel signal processing, motivated by the optimum …

Voice activity detection (VAD) algorithms are essential for many speech processing
applications, such as speaker diarization, automatic speech recognition, speech …

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 …

Measuring the quality of noisy speech signals has been an increasingly important problem
in the field of speech processing as more and more speech-communication and human …

Array processing is widely used in many speech applications involving multiple
microphones. These applications include automatic speech recognition, robot audition …

A multichannel speech enhancement system usually consists of spatial filters such as
adaptive beamformers followed by postfilters, which suppress remaining noise. Accurate …

The power method in conjunction with deflation provides an economical approach to
compute an eigenvalue decomposition (EVD) of a low-rank Hermitian matrix, which typically …

As the number of connected devices equipped with multiple microphones increases,
scientific interest in distributed microphone array processing grows. Current beamforming …