ShlokaDhara: A Multi-Speaker Sanskrit Speech Corpus for Learning Meter-Aware Prosody in Text-to-Recitation

Abstract

Sanskrit verse recitation is governed by $\text{Chandas}$, a strict prosodic framework defined by rhythmic $\text{Laghu}$ (light) and $\text{Guru}$ (heavy) syllables. Because existing Sanskrit speech datasets prioritize ASR or conversational TTS, they lack the metrical fidelity required for authentic verse synthesis. We introduce $\textbf{$\textit{ShlokaDhara}$}$, a 43-hour, multi-speaker Sanskrit speech corpus capturing diverse meters, recitation styles, and melodic variations. We specifically focus on $\text{Text-to-Recite}$, a synthesis task explicitly requiring both prosodic accuracy and metrical correctness. Benchmarking state-of-the-art TTS architectures via an automated UTMOS (neural MOS) predictor demonstrates that XTTS-v2 achieves the highest perceptual quality. We further validate the XTTS-v2 generated samples using two paradigms: subjective assessments by 16 Sanskrit experts and novel empirical metrics, specifically, syllable-level $\text{Laghu-Guru}$ duration analysis to objectively quantify structural metrical awareness. Ultimately, this work establishes a robust dataset and evaluation benchmark for prosody-aware computational Sanskrit recitation. Generated Samples: https://shlokadhara.vercel.app/

arXiv commentUnder submission at a conference. v2: Typo corrected in author name, content unchanged