Reproduce “AutoProsody: A Prosodic Feature Extraction Tool for Indian Languages” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Preethi Thinakaran, Malarvizhi Muthuramalingam, Sooriya S, Anushiya Rachel Gladston, P. Vijayalakshmi, Hema A Murthy, T. Nagarajan
AutoProsody provides automatic, syllable-centered prosodic annotations across five Indian languages. Its pipeline aligns phones, syllables, and words, then labels pitch contours, relative intensity, and pauses; it accepts audio with optional transcript text. In read-speech evaluation, Tamil phoneme-segmentation accuracy was 95.2% with language-dependent models and 86.0% with language-independent models, alongside 0.98 pitch correlation and 0.96 break-index F1, for this setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.