Speech Prosody 2010

Chicago, IL, USA
May 10-14, 2010

A Preliminary Analysis of the Relationship of Speech Rate to Speech-Timing Metrics as applied to Large Corpora of Non-Laboratory Speech in English and Chinese Broadcast News

Matthew Benton

Department of Linguistics and TESOL, The University of Texas at Arlington, USA

A renewed interest, in recent years, has occurred in the area of speech rhythm (traditionally defined by categories of speech timing patterns based on perceptual or acoustic durations of stresses, syllables, or moras). Since accurate categorization seems to be a three dimensional problem (durations of vocalic intervals, intervocalic intervals, and speech rate/tempo), some studies have made provision for differences in speech tempo by providing metrics with rate normalizing parameters based on the intervocalic intervals or the vocalic intervals (VarcoC & nPVI-V respectively). This study applies these different metrics on larger corpora of many speakers and more naturally occurring speech.

Index Terms: speech rhythm, pairwise variability, speech tempo, syllable timing, stress timing

Full Paper

Bibliographic reference.  Benton, Matthew (2010): "A preliminary analysis of the relationship of speech rate to speech-timing metrics as applied to large corpora of non-laboratory speech in English and Chinese broadcast news", In SP-2010, paper 423.