Common Voice project makes available for the speech community recordings in a great variety of languages from many speakers. This material can be used to improve the quality of synthetic voices in many different languages. In this master thesis we propose to use the available recording in Basque, Catalan, Spanish and English to improve the current AhoTTS multilingual synthesizer (aholab.ehu.eus/tts). Apart from data curation and development of voices, the student will perform subjective listening tests to assess the improvements.
Goal: Improve the quality of the current AhoTTS synthetic voices.
Requirements: Basic knowledge of scripting language and python programming will be appreciated.
Advisor: Eva Navas Cordón, Inma Hernáez Rioja