14th International Congress of Phonetic Sciences (ICPhS-14)

San Francisco, CA, USA
August 1-7, 1999

Time-Series Analysis of Conversational Prosody for the Identification of Rhythmic Units

Eugene H. Buder (1), Anders Eriksson (2)

(1) The University of Memphis, USA; (2) University of Stockholm, Sweden

The notion that conversational speech is rhythmically organized has been explored in a variety of disciplinary frameworks, but the acoustic phonetic basis for this notion has been problematic. Intensive analysis of a one-minute fragment of dialogue between two American men utilizing segmental and prosodic analysis reveals a hierarchy of units that appear to establish rhythmic coherence across utterances within and across speakers. The prosodic analysis utilizes spectral frequency, magnitude and phase data resulting from FFTs of normalized F0 and intensity records from overlapping 8-second frames. The segmental analysis utilizes spectrographic-based temporal data specifying word units and vowel onsets. The two types of data are interpreted together to associate specific sinusoidal models (phase, frequency, and magnitude) with specific speech events (e.g., sequences of utterances, chains of prosodic prominence or word stress intervals, or dyadic events such as question-answer sequences or simultaneous speech onsets).

Full Paper

Bibliographic reference.  Buder, Eugene H. / Eriksson, Anders (1999): "Time-series analysis of conversational prosody for the identification of rhythmic units", In ICPhS-14, 1071-1074.