TY - JOUR
T1 - Exploring the discriminatory potential of F0 distribution parameters in traditional forensic speaker recognition
AU - Kinoshita, Yuko
AU - Ishihara, Shunichi
AU - Rose, Philip
PY - 2009
Y1 - 2009
N2 - Despite its many prima facie attractive properties for forensic speaker recognition, F0 is regarded as having limited forensic value due to its large within-speaker variability. However, its forensic use to date has been limited mostly to its long-term mean and standard deviation. This paper examines the discriminatory potential, within a Likeli-hood Ratio-based approach, of additional parametric features from the distribution of long-term Fo: its skew, kurtosis, modal F0 and modal density. Motivated by the observation that the shape of the long-term F0 distribution shows less within-speaker occasion-to-occasion difference, we report a forensic discrimination experiment with non-contemporaneous speech samples from 201 male Japanese speakers. Using a multivariate Likelihood Ratio as discriminant distance with the six LTF0 distribution parameters, an equal error rate of 10.7% is obtained from 201 target and 80400 non-target trials. We also investigate how the EER degrades as a function of amount of voiced speech.
AB - Despite its many prima facie attractive properties for forensic speaker recognition, F0 is regarded as having limited forensic value due to its large within-speaker variability. However, its forensic use to date has been limited mostly to its long-term mean and standard deviation. This paper examines the discriminatory potential, within a Likeli-hood Ratio-based approach, of additional parametric features from the distribution of long-term Fo: its skew, kurtosis, modal F0 and modal density. Motivated by the observation that the shape of the long-term F0 distribution shows less within-speaker occasion-to-occasion difference, we report a forensic discrimination experiment with non-contemporaneous speech samples from 201 male Japanese speakers. Using a multivariate Likelihood Ratio as discriminant distance with the six LTF0 distribution parameters, an equal error rate of 10.7% is obtained from 201 target and 80400 non-target trials. We also investigate how the EER degrades as a function of amount of voiced speech.
KW - Duration
KW - Fo
KW - Forensic speaker recognition
KW - Japanese
KW - Likelihood ratio
KW - Multivariate likelihood ratio
KW - Non-contemporaneous recordings
KW - Spontaneous speech
KW - Univariate likelihood ratio
UR - http://www.scopus.com/inward/record.url?scp=70349732908&partnerID=8YFLogxK
U2 - 10.1558/ijsll.v16i1.91
DO - 10.1558/ijsll.v16i1.91
M3 - Article
SN - 1748-8885
VL - 16
SP - 91
EP - 111
JO - International Journal of Speech, Language and the Law
JF - International Journal of Speech, Language and the Law
IS - 1
ER -