The paper presents MUNTTS, a text-to-speech (TTS) system developed specifically for Mundari, a low-resource Indian language. The authors have collected and processed substantial data to build this speech synthesis system. The system was evaluated by native speakers and objective metrics to demonstrate its effectiveness. The development of such a system presents several challenges, such as the need for large amounts of audio recordings and text transcriptions, the scarcity of native speakers familiar with technology, and the high cost of obtaining quality audio recordings. Despite these challenges, the authors were able to build a TTS system for Mundari, contributing to the preservation and promotion of this underrepresented language in the digital age.
Publication date: 31 Jan 2024
Project Page: https://aka.ms/MUN-TTS
Paper: https://arxiv.org/pdf/2401.15579
