Go To Program 134 Blind Source Separation with Distributed Microphone Pairs Using Permutation Correction by Intra-Pair TDOA Clustering Takuma Ono 1 Shigeki Miyabe 1 Nobutaka Ono 1 Shigeki Sagayama 1 1 Graduate School of Information Science and Technology, The University of Tokyo, Tokyo, Japan In this paper, we present a novel framework of distributed microphone array for blind source separation (BSS), where stereo microphones or proximately-placed microphone pairs are distributed. Recently, the framework of distributed microphones has been investigated in order to separate wide-spread sources. Although frequency-domain independent component analysis (ICA) is a standard approach of BSS, in the distributed microphone arrays the distance between microphones could be large, which makes it difficult to correct the permutation problem of frequency-domain ICA. In order to correct it, we present the framework of distributing sub-arrays, which consist of closely spaced microphones for avoiding spatial aliasing, and using the time difference of arrivals (TDOAs) within sub-arrays which give definite clue for the permutation correction. We focus on specifically distributing paired microphones because many recording devices have stereo channels and it would be very suitable with the context of the distributed microphone array. Based on it, the permutation problem is corrected by clustering the intra-pair TDOAs. In addition, we also discuss using multiple stereo recording devices, where inter-pair channels are asynchronized. Finally, we compare the separation performance of the proposed method and several existing methods by experimental results in real environment. View pdf