Characterization of Rhythms of Mising Using Time and Frequency Domain Correlates

Abstract

This thesis presents a comprehensive analysis of the rhythmic characteristics of Mising, a low-resource language spoken in northeastern India, utilizing both time and frequency domain techniques. We consider two Mising dialects, viz. Pagro and Delu, along with Assamese, the official language of Assam, India, to study their rhythmic characteristics in the case of spontaneous speech. In the time-domain analysis, we investigate automated rhythm measure calculation using vowel onset point (VOP) and vowel end point (VEP) detection algorithms without the need for manual annotation or forced alignment models. We employ machine learning models to classify Assamese and Mising using automatically computed rhythm measures and speech rate. Results indicate that Mising varieties exhibit stress-timed characteristics, in contrast to Assamese’s mora-timed nature.