In this paper, we study the disentanglement of speaker and language representations in non-autoregressive cross-lingual TTS models from various aspects. We propose a phoneme length regulator that solves the length mismatch problem between IPA input sequence and monolingual alignment results. Using the phoneme length regulator, we present a FastPitch-based cross-lingual model with IPA symbols as input representations. Our experiments show that language-independent input representations (e.g. IPA symbols), an increasing number of training speakers, and explicit modeling of speech variance information all encourage non-autoregressive cross-lingual TTS model to disentangle speaker and language representations. The subjective evaluation shows that our proposed model can achieve decent naturalness and speaker similarity in cross-language voice cloning.
@article{arxiv.2110.07192,
title = {Exploring Timbre Disentanglement in Non-Autoregressive Cross-Lingual Text-to-Speech},
author = {Haoyue Zhan and Xinyuan Yu and Haitong Zhang and Yang Zhang and Yue Lin},
journal= {arXiv preprint arXiv:2110.07192},
year = {2022}
}