A comparison of convolutive blind source separation algorithms applied to speech signals

Blind source separation aims at estimating a number of source signals using from several mixtures of those source signals. In the case of acoustic applications the sources are people speaking the mixtures are microphone recordings. Many different algorithms have been proposed to solve this problem. There is however a need for a more involved comparison of the performance of the different algorithms. This dissertation examines the performance of several blind source separation algorithms.