Algorithm for frequency-domain blind source separation based on the pseudoanechoic model for 2 by 2 mixtures, according to the description in [1]. It implements also the automatic bin selection and correlated wiener postfilter as described in [2]. The algorithm needs two .wav files with the two mixtures as measured by the microphones, and also the window length, step size and microphone spacing. For the provided sample files, use wlength=256, step=128 and spacing=0.05.
Contact: Leandro Di Persia