Advanced

Multitaper Estimation of Frequency-Warped Cepstra with Application to Speaker Verification

Sandberg, Johan LU ; Sandsten, Maria LU ; Kinnunen, Tomi; Saeidi, Rahim; Flandrin, Patrick and Borgnat, Pierre (2010) In IEEE Signal Processing Letters 17(4). p.343-346
Abstract
Usually the mel-frequency cepstral coefficients are estimated either from a periodogram or from a windowed periodogram. We state a general estimator which also includes multitaper estimators. We propose approximations of the variance and bias of the estimate of each coefficient. By using Monte Carlo computations, we demonstrate that the approximations are accurate. Using the proposed formulas, the peak matched multitaper estimator is shown to have low mean square error (squared bias + variance) on speech-like processes. It is also shown to perform slightly better in the NIST 2006 speaker verification task as compared to the Hamming window conventionally used in this context.
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Speech analysis, Multitapers, Speaker verification, Cepstral analysis, Multiple windows, MFCC
in
IEEE Signal Processing Letters
volume
17
issue
4
pages
343 - 346
publisher
IEEE--Institute of Electrical and Electronics Engineers Inc.
external identifiers
  • WOS:000274733400001
  • Scopus:77249096360
ISSN
1070-9908
DOI
10.1109/LSP.2010.2040228
language
English
LU publication?
yes
id
1c307540-6f78-49bb-8656-6a7e656cefee (old id 1515143)
date added to LUP
2009-12-14 11:06:11
date last changed
2016-10-13 04:40:07
@misc{1c307540-6f78-49bb-8656-6a7e656cefee,
  abstract     = {Usually the mel-frequency cepstral coefficients are estimated either from a periodogram or from a windowed periodogram. We state a general estimator which also includes multitaper estimators. We propose approximations of the variance and bias of the estimate of each coefficient. By using Monte Carlo computations, we demonstrate that the approximations are accurate. Using the proposed formulas, the peak matched multitaper estimator is shown to have low mean square error (squared bias + variance) on speech-like processes. It is also shown to perform slightly better in the NIST 2006 speaker verification task as compared to the Hamming window conventionally used in this context.},
  author       = {Sandberg, Johan and Sandsten, Maria and Kinnunen, Tomi and Saeidi, Rahim and Flandrin, Patrick and Borgnat, Pierre},
  issn         = {1070-9908},
  keyword      = {Speech analysis,Multitapers,Speaker verification,Cepstral analysis,Multiple windows,MFCC},
  language     = {eng},
  number       = {4},
  pages        = {343--346},
  publisher    = {ARRAY(0x93fce40)},
  series       = {IEEE Signal Processing Letters},
  title        = {Multitaper Estimation of Frequency-Warped Cepstra with Application to Speaker Verification},
  url          = {http://dx.doi.org/10.1109/LSP.2010.2040228},
  volume       = {17},
  year         = {2010},
}