Yc. Cao et al., MULTICHANNEL SPEECH SEPARATION BY EIGENDECOMPOSITION AND ITS APPLICATION TO CO-TALKER INTERFERENCE REMOVAL, IEEE transactions on speech and audio processing, 5(3), 1997, pp. 209-219
This paper describes the concept of eigendecomposition for multichanne
l signal separation, an alternative method of enhancing the desired si
gnal corrupted by interference, The method uses two observations that
come from a pair of sensors, both of which contain the desired signal
and the undesired signal(s), The method assumes that the desired signa
l and the undesired signal(s) are uncorrelated, and the signal-to-nois
e ratios (SNR's) of each observation are different and not necessarily
greater than unity, The technique has been successfully used to separ
ate speech signals corrupted heavily by ambient noise, co talker inter
ference, and other sources such as background music, The method can al
so be applied to instances where more than two sensors are used by fir
st forming two beams and then processing the output of the two beamfor
mers.