2007/01/10

Xu, Y. (2005). Speech Melody as articulatory implemented communicative functions. Speech Communication, 46, 220–251.

Presentation: Shelly
Summary: Chris

There are two fundamental facts related to speech prosody. First, speech conveys communicative meanings. Second, speech is produced under the constraint of human articulators. This paper investigates the correlation between the acoustic cues and communicative meanings. Xu argues that melodic functions are not encoded by unchanging surface acoustic forms. Instead, successful communicative functions are achieved by using a limited number of phonetic primitives including local pitch targets, pitch range, articulatory strength, and duration. The realizations of these primitives are determined by different encoding schemes that are tied into particular communicative functions. Xu then proposed the parallel encoding and target approximation (PENTA) model in which melodic components are defined by specific functions rather than forms. PENTA contains stringent articulatory assumptions but transmits multiple communicative functions in parallel. The communicative functions are ultimately converted to continuous surface acoustic patterns by a process of syllable-synchronized sequential target approximation.