Research Committee | Current Projects
Forensic comparison of Spanish twins and non-twin brothers' voices
Eugenia San Segundo Fernández1 and Prof. Dr. Hermann J. Künzel2
1Consejo Superior de Investigaciones Científicas, Laboratorio de Fonética, Madrid, Spain
2Department of Phonetics, University of Marburg, Germany
This project investigates the effectiveness of voice-source features, such as jitter and shimmer, for forensic voice comparison. We will employ Glottex software (Gómez-Vilda et al., 2009) which includes 51 algorithms for the extraction of voice-source properties. Testing will be conducted using a database of voice recordings of 60 Spanish brothers, including twins. We consider that both twins and non-twin brothers may be useful to test the extent to which a parameter is robust for forensic comparison of the voices of very similar speakers. The validity and reliability of the likelihood-ratio results will be assessed (Morrison, Thiruvaran, & Epps, 2010).
References
Gómez-Vilda, P., Fernández-Baillo, R., Rodellar-Biarge, V., Nieto-Lluis, V., Álvarez-Marquina, A., Mazaira-Fernández, L.M., Martínez-Olalla, R. and Godino-Llorente, J.I. (2009). Glottal Source biometrical signature for voice pathology detection. Speech Communication, 51, 759-781.
Morrison, G.S., Thiruvaran, T., Epps, J. (2010). Estimating the precision of the likelihood-ratio output of a forensic-voice-comparison system. In H. Cernocký and L. Burget (Eds.), Proceedings of Odyssey 2010: The Language and Speaker Recognition Workshop, Brno, Czech Republic (pp. 63-70). International Speech Communication Association.