MixedG2P-T5: G2P-free Speech Synthesis for Mixed-script texts using Speech Self-Supervised Learning and Language Model
Abstract
This study presents a novel approach to voice synthesis that can substitute the traditional grapheme-to-phoneme (G2P) conversion by using a deep learning-based model that generates discrete tokens directly from speech. Utilizing a pre-trained voice SSL model, we train a T5 encoder to produce pseudo-language labels from mixed-script texts (e.g., containing Kanji and Kana). This method eliminates the need for manual phonetic transcription, reducing costs and enhancing scalability, especially for large non-transcribed audio datasets. Our model matches the performance of conventional G2P-based text-to-speech systems and is capable of synthesizing speech that retains natural linguistic and paralinguistic features, such as accents and intonations.
Cite
@article{arxiv.2509.01391,
title = {MixedG2P-T5: G2P-free Speech Synthesis for Mixed-script texts using Speech Self-Supervised Learning and Language Model},
author = {Joonyong Park and Daisuke Saito and Nobuaki Minematsu},
journal= {arXiv preprint arXiv:2509.01391},
year = {2025}
}
Comments
In Proceedings of the 17th Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC 2025)