SRC_Num_TDOA: Multiple speech sources’ number and their TDOA Estimation from a stereo recorded mixture

We consider the problem of sources counting and Time Delays of Arrival (TDOAs) estimation from a two-channel reverberant mixture in the underdetermined case. Most existing estimating TDOA implementations are based on a prior knowledge about the number of involved sources in the mixture. Whereas, we develop a new function that only takes the observed mixtures as input. In this paper, we describe the basic of the proposed function, its development and its evaluation on real recorded mixtures with reference to widely used methods. Through simulations and real environment experiments, we indicate that our method outperforms other state-of-the-art TDOA methods in terms of accuracy, while being significantly more efficient in terms of computational complexity