Abstract: We propose a novel dual-path state-space model within an encoder-decoder architecture for multichannel speech enhancement that leverages joint temporal and spectral modeling to significantly ...