Sound files associated with the paper

A BAYESIAN APPROACH TO TIME-FREQUENCY BASED BLIND SOURCE SEPARATION
C. Févotte and S. J. Godsill


Submitted to 2005 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics



Stereo Mix

iid model (Section 3)


l_frame = 32 ms :      Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 64 ms :      Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 128 ms :    Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 256 ms :    Estimated source 1     Estimated source 2     Estimated source 3

frequency-dependent model (Section 4)


l_frame = 32 ms :      Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 64 ms :      Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 128 ms :    Estimated source 1     Estimated source 2     Estimated source 3
l_frame = 256 ms :    Estimated source 1     Estimated source 2     Estimated source 3



Original source 1      Original source 2      Original source 3