![]() |
ESCA Workshop on Automatic Speaker Recognition, Identification, and VerificationMartigny, Switzerland |
![]() |
Log-power spectra were calculated in the centre of continuant sounds for speaker identification. Results indicate that there is a difference in speaker identification power among these sounds. Under the conditions of the experimental and computational set-up described in this paper, vowels and nasals seem to contain more information on speaker identity than fricatives.
Bibliographic reference. Greisbach, Reinhold / Esser, Otto / Weinstock, Constanze / Klinker, Theo (1994): "A comparison of the speaker identification power of continuant sounds", In ASRIV-1994, 177-180.