Speech Database/Tool System and Preliminary Accent Recognition Study
Arthur Phidd, Padmashree Thimmappa, Richard Sauther, Shankar Vijayakumar, Charles C. Tappert · 2005
The voice/tool database System is a web-based system connected to a relational backend database that stores voice files (.wav format) along with user demographic information. The database is used to test the ability of an examiner, usually a briefly trained non-linguist, to identify the nationality of the speaker of the recorded voice. Here we provide background on voice recognition biometrics, describe the design and functionality of the pronunciation information system, and present some preliminary results on the nationality classification of voice samples by examiners. We also provide the ability to play the voice files in a speech processing tool for further research in the field.