Combining evidence from source, suprasegmental and spectral features for a fixed-text speaker verification system | IEEE Journals & Magazine | IEEE Xplore