Experimental evaluation of source separation with only one sensor

We report on the evaluation of a new method for audio source separation using only one sensor. The method can be viewed as a generalization of Wiener filtering to locally stationary signals, where the sources are modelled using power spectral density dictionaries which are estimated during a training step. The experiments were designed to measure how separation performance varied with amount of training data, model complexity and the representativity of the training data. The results show that model complexity and training data representativity are more important than the amount of training data.