Humanoid separation of speech sources in reverberant environments | IEEE Conference Publication | IEEE Xplore