C
Carol Y. Espy-Wilson
Researcher at University of Maryland, College Park
Publications - 180
Citations - 4675
Carol Y. Espy-Wilson is an academic researcher from University of Maryland, College Park. The author has contributed to research in topics: Vocal tract & Formant. The author has an hindex of 32, co-authored 166 publications receiving 4161 citations. Previous affiliations of Carol Y. Espy-Wilson include University of Southern California & Massachusetts Institute of Technology.
Papers
More filters
Proceedings Article
Analysis of i-vector Length Normalization in Speaker Recognition Systems.
TL;DR: The proposed approach deals with the nonGaussian behavior of i-vectors by performing a simple length normalization, which allows the use of probabilistic models with Gaussian assumptions that yield equivalent performance to that of more complicated systems based on Heavy-Tailed assumptions.
Journal ArticleDOI
Articulatory tradeoffs reduce acoustic variability during American English /r/ production.
Frank H. Guenther,Carol Y. Espy-Wilson,Suzanne Boyce,Melanie L. Matthies,Majid Zandipour,Joseph S. Perkell +5 more
TL;DR: Analysis of acoustic and articulatory variabilities revealed that these tradeoffs act to reduce acoustic variability, thus allowing relatively large contextual variations in vocal tract shape for /r/ without seriously degrading the primary acoustic cue.
Journal ArticleDOI
Acoustic measures for linguistic features distinguishing the semivowels /w j r l/ in American English
TL;DR: In this article, acoustic properties related to the linguistic features which characterize the semivowels in American English were quantified and analyzed statistically, and the features of interest were sonorant, syllabic, consonantal, high, back, front, and retroflex.
Journal ArticleDOI
A magnetic resonance imaging-based articulatory and acoustic study of "retroflex" and "bunched" American English /r/.
TL;DR: The results suggest that the F4/F5 differences between the variants can be largely explained by differences in whether the long cavity behind the palatal constriction acts as a half- or a quarter-wavelength resonator.
Proceedings ArticleDOI
Multicondition training of Gaussian PLDA models in i-vector space for noise and reverberation robust speaker recognition
TL;DR: A multicondition training strategy for Gaussian Probabilistic Linear Discriminant Analysis (PLDA) modeling of i-vector representations of speech utterances using a collection of individual subsystems tuned to specific conditions is presented.