Combining evidence from residual phase and MFCC features for speaker recognition | IEEE Journals & Magazine | IEEE Xplore