论文部分内容阅读
This study addresses the problem of under-determined speech source separation from multichannel microphone signals,i.e.the convolutive mixtures of multiple sources.The time-domain signals are first transformed to the short-time Fourier transform(STFT)doma