18 February 2019

Research IT innovation fund for Aoju Chen

Prof. dr. Aoju Chen. Foto: Ivar Pel
Prof. Aoju Chen. Photo: Ivar Pel

Prof. Aoju Chen (Linguistic theory and language acquisition) has been granted a Research IT innovation fund for developing Automatic Analysis of Speech Prosody (AASP), a tool that automatically analyses the prosody of spoken Dutch. 

Prosody is the melody in our speech that binds words into sentences and communicates meaning in and beyond words (e.g. Coffee!  vs. Coffee?). AASP performs two levels of analysis: structural and holistic.

structural and holistic analysis

Structurally, AASP generates analysis of prosodic structural units in sentences, including pitch movements associated with the stressed syllable of a word (pitch accents), and pitch movements at the boundaries of a prosodic phrase, which may be (part of) a sentence. Holistically, AASP generates functional principle components, i.e. continuous mathematical functions that depict changes in the shape of a prosodic contour over a stretch of speech.

application AASP in research 

AASP can be used to analyse large speech samples to address both questions directly concerning prosody (e.g. prosodic development) and questions arising from interdisciplinary projects that require prosodic analysis as input for further analysis (e.g. the role of prosody in social robotics). Analysing the same data both structurally and holistically can shed light from different theorectical perspectives. We can also use AASP to generate low-level acoustic analyses to address acoustically-oriented questions.