Multilingual Tt
Multilingual text-to-speech (TTS) research aims to create systems capable of generating natural-sounding speech in multiple languages, often addressing challenges posed by limited data availability for less-resourced languages. Current efforts focus on developing robust model architectures that leverage techniques like self-supervised learning, disentangled representations, and translation-enhanced approaches to improve cross-lingual transfer and maintain speaker identity and accent. These advancements are significant because they promise to broaden access to speech technology across diverse linguistic communities and facilitate applications such as multilingual dubbing and cross-lingual communication aids.
Papers
July 19, 2024
May 30, 2023
March 1, 2023
January 24, 2023
October 27, 2022
June 24, 2022