Maximum Likelihood Time Delay Estimation with Phase Domain Analysis in the Generalized Cross Correlation Framework

We propose a new method for efficiently estimating the maximum likelihood frequency weighting in the generalized cross correlation framework for time delay estimation. The estimation is based on the analysis of the cross spectrum between a pair of microphones. We model how phase distribution is affected by both noise and reverberation, and relax the common assumption that noise and reverberation are uncorrelated with the source. Thus, our method does not require knowledge of the noise spectrum or a detailed model of the reverberation. Experimental results show that the proposed method is superior to the PHAT method.