Title What Else is New Than the Hamming Window? Robust MFCCs for Speaker Recognition via Multitapering
Authors Tomi Kinnunen, Rahim Saeidi, Johan Sandberg, Maria Sandsten
Full-text Available as PDF, Restricted Access
Alternative Location http://cs.joensuu.fi/pages/...
Publication InterSpecch 2010
Year 2010
Pages 2734 - 2737
Document type Conference paper
Conference name Interspeech 2010
Conference Date September 2010
Conference Location Makuhari, Japan
Status Published
Quality controlled Yes
Language eng
Abstract English Usually the mel-frequency cepstral coefficients (MFCCs) are derived via Hamming windowed DFT spectrum. In this paper, we advocate to use a so-called multitaper method instead. Multitaper methods form a spectrum estimate using multiple window functions and frequency-domain averaging. Multitapers provide a robust spectrum estimate but have not received much attention in speech processing. Our speaker recognition experiment on NIST 2002 yields equal error rates (EERs) of 9.66 % (clean data) and 16.41 % (-10 dB SNR) for the conventional Hamming method and 8.13 % (clean data) and 14.63 % (-10 dB SNR) using multitapers. Multitapering is a simple and robust alternative to the Hamming window method.
Keywords speaker verification, multiple window method,

Questions: webmaster
Last update: 2013-04-11

Centre for Mathematical Sciences, Box 118, SE-22100, Lund. Telefon: +46 46-222 00 00 (vx)