Technical Report of IEICE Japan, Vol. SP2004-164, pp. 1-6, 2005 (in Japanese)
The correspondences between the differences among the phones in human speaker identification and their acoustic properties
K. Amino, T. Sugawara and T. Arai
Abstract: The higher success of human speaker identification depends on the speech contents given to the listeners. In previous studies, the voiced sonorants, especially the nasals, were reported to be available for the identification of the speaker. In this study, perception tests were carried out in order to investigate the effectiveness of some Japanese phones excerpted from the sentences. Again, it was shown that the percent correct was the highest when the stimuli containing the nasal sounds were used. We have then attempted to acoustically explain these differences in stimuli, and we found out that there are correspondences between these differences among the phones and the cepstral distances of the stimuli.
Keywords: nasals, speaker identification, individuality, cepstral distance