An Experiment in Paratone Detection in a Prosodically Annotated EAP Spoken Corpus

Abstract

This article describes an experiment in paratone detection based on a spoken corpus of English for Academic Purposes (EAP) recently automatically re-annotated with prosodic information. The Momel and INTSINT annotations were carried out using SPPAS. The EIIDA corpus was chosen as it offered long uninterrupted stretches of speech of academic presentations. We describe the clustering method adopted for automatic detection, contrasting a supervised and an unsupervised method of paratone boundary detection. We showcase the relevance of the annotation scheme followed for this corpus and contribute to the investigation of the phonostyle of lecture delivery. We discuss the relevance of clustering methods applied to the labels of the pitch targets for the analysis of paratones.

Revisiting Paratone Prosodic Features with the EIIDA corpus

“…One could try to build the algorithm hierarchically, taking into account the most robust prosodic cues for paratones. The alternative approach would consist in using co-clustering techniques (see [32]) simultaneously dealing with relevant cues such as rhythmic cues (both as spacing of syllable peaks and phone intervals) and INTSINT pitch targets and F0 associated values.…”

“…The expected gradual baseline was not observed (claim 3), as can be seen on Figure 1. Although detection based on absolute levels of pitch has been fruitless, [32] reports success in capturing the detection of pitch resets as the prediction of recurrent INTSINT patterns, finding that 3gram clusters of INTSINT pitch targets correlate with paratone boundaries. Figures 2 and 3 show that typical INSTINT patterns can be observed in initial and final position of our manually annotated paratones.…”