Analyzing speech movement variability from audio recordings

Van Brenk, Frits and Lowit, Anja (2010) Analyzing speech movement variability from audio recordings. In: Conference on Cognitive and Physical Models of Speech Production, Speech Perception and Production-Perception Interaction, 2010-09-27 - 2010-10-01.

[img]
Preview
Text (van-Brenk-Lowit-CPMSP-2010-Analyzing-speech-movement-variability-from-audio)
van_Brenk_Lowit_CPMSP_2010_Analyzing_speech_movement_variability_from_audio.pdf
Final Published Version

Download (45kB)| Preview
    [img]
    Preview
    Text (van-Brenk-Lowit-CPMSP-2010-Analyzing-speech-movement-variability)
    van_Brenk_Lowit_CPMSP_2010_Analyzing_speech_movement_variability.pdf
    Accepted Author Manuscript

    Download (261kB)| Preview

      Abstract

      Producing fluent and intelligible speech involves quick, precise and coordinated movements of articulators. Developmental and acquired speech motor disorders often result in an instable gestural organization of speech movements, and thus more variable speech performance. In addition, research suggest that a number of environmental factors can influence articulatory movements, including speech rate modifications, linguistic and cognitive complexity of the speech task and concurrent motor tasks. The possible interaction of internal and external factors that influence speech motor control might give more insight in the nature of different speech disorders. The spatiotemporal index (STI) (Smith et al., 1995) and Functional Data Analysis (FDA) (Ramsay et al., 1996) are techniques that have been shown to be successful in investigating variability of kinematic movements obtained by lip displacement tracking, but may also be applied to other dimensions of speech, including amplitude envelopes and pitch and formant tracks. This pilot study presents some first results on STI and FDA variability measures applied to audio recordings of young, healthy adults. The study aims to investigate whether variability measures obtained from acoustic data reflect those from kinematic data.