13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Acoustic and Perceptual Similarity in Coarticulatorily Nasalized Vowels

Rebecca Scarborough (1), Georgia Zellou (1,2)

(1) Linguistics Department, University of Colorado, Boulder, CO, USA
(2) Linguistics Department, University of Pennsylvania, Philadelphia, PA, USA

This study investigates the acoustic and perceptual consequences of nasal coarticulation in English. Nasalized (coarticulated) vowels were found to be closer in the F1-F2 acoustic vowel space than corresponding oral (non-coarticulated) vowels, indicating that contrast is reduced in the nasal vowel space, relative to the oral vowel space. With respect to perception, listeners are, perhaps unsurprisingly, more accurate in identifying oral vowels than nasalized vowels. Interestingly, however, while listeners take longer to identify nasalized vowels than oral vowels when they hear those vowels in isolation, this difference in processing time disappears when the nasalized and oral vowels are heard in lexical contexts. We take these findings to indicate that listeners compensate for nasal coarticulation, albeit sometimes incompletely, attributing the acoustic effects to their consonantal source, and that this compensation takes place instantaneously.

Index Terms: nasality, coarticulation, perception, similarity

Full Paper

Bibliographic reference.  Scarborough, Rebecca / Zellou, Georgia (2012): "Acoustic and perceptual similarity in coarticulatorily nasalized vowels", In INTERSPEECH-2012, 2694-2697.