Temporal stability of long term measures of fundamental frequency
Author(s) -
Pablo Arantes,
Anders Eriksson
Publication year - 2014
Publication title -
speech prosody
Language(s) - English
Resource type - Conference proceedings
SCImago Journal Rank - 0.274
H-Index - 18
ISSN - 2333-2042
DOI - 10.21437/speechprosody.2014-220
Subject(s) - term (time) , stability (learning theory) , value (mathematics) , base (topology) , point (geometry) , mean value , mathematics , variance (accounting) , statistics , variation (astronomy) , computer science , mathematical analysis , physics , geometry , quantum mechanics , machine learning , accounting , astrophysics , business
Speech perception depends on the ability to rapidly accommodate considerable variability in speech rate. We present results from two eye-tracking experiments indicating that listeners use context speech rate to generate, maintain, and update probabilistic hypotheses about the timing and number of constituents in upcoming speech. Participants heard utterances containing polysyllabic nouns preceded by indefinite articles and followed by [s]-initial words (e.g. ...saw a raccoon slowly...). We altered the speech rate of the indefinite article and of the [s] with respect to surrounding context, manipulating the likelihood that the item would be perceived as singular (a raccoon) vs. plural (raccoons). Shorter indefinite articles elicited higher proportions of fixations to plural target pictures than longer articles both before and after the processing of [s], demonstrating that listeners made rapid use of prosodic cues to the presence or absence of the article. Importantly, fixations were also influenced by the duration of [s] relative to context speech rate. These findings suggest that listeners maintain and update provisional speech-rate hypotheses across multiple morphophonemic units. We interpret these results with respect to probabilistic approaches to spoken language understanding.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom