Strength of forensic speaker identification evidence : multispeaker formant and cepstrum-based segmental discrimination with a Bayesian likelihood ratio as threshold

A forensic-phonetic speaker identification experiment is described which tests to what extent same-speaker pairs from a 60 speaker Japanese data base can be discriminated from different-speaker pairs using a Bayesian likelihood ratio (LR) as discriminant function. Non-contemporaneous telephone recordings are used, with comparison based on mean values from three segments only: a nasal, a voiceless fricative, and a vowel. It is shown that discrimination using the LR-based distance is better than with a conventional distance, and that the cepstrum outperforms the formants. A LR for the test of 50 is obtained for formant-based discrimination, compared to c. 900 for the cepstrum, and the tests are thus shown to be capable of yielding a probative strength of support for the prosecution hypothesis that is conventionally quantified as ‘moderate’ for formants but ‘moderately strong’ for the cepstrum. Comparisons are made with results from similar experiments

[1]  Philip Rose.  Forensic Speaker Identification , 2002 .

[2]  Yuko Kinoshita.  Use of likelihood ratio and Bayesian approach in forensic speaker identification , 2002 .

[3]  A. Broeders,et al.  Some observations on the use of probability scales in forensic identification , 1999 .

[4]  Javier Ortega-Garcia,et al.  On the application of the Bayesian approach in real forensic conditions with GMM-based systems , 2001, Odyssey.

[5]  Alvin F. Martin,et al.  The NIST 1999 Speaker Recognition Evaluation - An Overview , 2000, Digit. Signal Process..

[6]  Hirotaka Nakasone,et al.  Forensic automatic speaker recognition , 2001, Odyssey.

[7]  Douglas A. Reynolds,et al.  The SuperSID project: exploiting high-level information for high-accuracy speaker recognition , 2003, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03)..

[8]  David Lindley,et al.  A problem in forensic science , 1977 .

[9]  Didier Meuwly,et al.  Forensic speaker recognition based on a Bayesian framework and Gaussian mixture modelling (GMM) , 2001, Odyssey.

[10]  I W Evett,et al.  An illustration of the advantages of efficient statistical methods for RFLP analysis in forensic science. , 1993, American journal of human genetics.

[11]  Franco Taroni,et al.  Statistics and the Evaluation of Evidence for Forensic Scientists , 2004 .

[12]  Timothy J. Vance,et al.  An introduction to Japanese phonology , 1987 .

[13]  N. B. Skachkov,et al.  On the application of , 2002 .

[14]  Hermann J. Künzel,et al.  Beware of the ‘telephone effect’: the influence of telephone transmission on the measurement of formant frequencies , 2001 .