Skip to main content


	title        = {Voice similarity — a comparison between judgements by human listeners and automatic voice comparison},
	abstract     = {Comparison between the way human listeners judge voice similarity and how state-of-the art GMM-UBM systems for voice recognition compare voices is a little explored area of research. In this study groups of informants judged the similarity between voice samples taken from a set of fairly similar male voices that had previously been used in a voice line-up experiment. The result from the listening tests was then compared to the scores from a UBM-GMM automatic voice
comparison system, built on the Mistral LIA_RAL open source platform. The results show a correlation between scores obtained from the automatic system and the judgements by the listeners. Listeners are, however, more sensitive to language dependent parameters or idiosyncratic phonetic features such as speaking tempo, while the system only bases its likelihood ratios on spectral similarities, i.e. timbre.},
	booktitle    = {Proceedings from FONETIK 2010, Working Papers},
	author       = {Lindh, Jonas and Eriksson, Anders},
	year         = {2010},
	volume       = {54},
	pages        = {63--69},