DOA-informed switching independent vector extraction and beamforming for speech enhancement in underdetermined situations

Abstract This paper proposes novel methods for extracting a single Speech signal of Interest (SOI) from a multichannel observed signal in underdetermined situations, i.e., when the observed signal contains more speech signals than microphones. It focuses on extracting the SOI using prior knowledge o...

Full description

Bibliographic Details

Published in:	EURASIP Journal on Audio, Speech, and Music Processing
Main Authors:	Tetsuya Ueda, Tomohiro Nakatani, Rintaro Ikeshita, Shoko Araki, Shoji Makino
Format:	Article
Language:	English
Published:	SpringerOpen 2024-10-01
Subjects:	Speech enhancement Underdetermined situations Switching mechanism Time-varying Gaussian distribution
Online Access:	https://doi.org/10.1186/s13636-024-00373-3