Towards achieving robust universal neural vocoding
Abstract
This paper explores the potential universality of neural vocoders. We train a WaveRNN-based vocoder on 74 speakers coming from 17 languages. This vocoder is shown to be capable of generating speech of consistently good quality (98% relative mean MUSHRA when compared to natural speech) regardless of whether the input spectrogram comes from a speaker or style seen during training or from an out-of-domain scenario when the recording conditions are studio-quality. When the recordings show significant changes in quality, or when moving towards non-speech vocalizations or singing, the vocoder still significantly outperforms speaker-dependent vocoders, but operates at a lower average relative MUSHRA of 75%. These results are shown to be consistent across languages, regardless of them being seen during training (e.g. English or Japanese) or unseen (e.g. Wolof, Swahili, Ahmaric).
Keywords
Cite
@article{arxiv.1811.06292,
title = {Towards achieving robust universal neural vocoding},
author = {Jaime Lorenzo-Trueba and Thomas Drugman and Javier Latorre and Thomas Merritt and Bartosz Putrycz and Roberto Barra-Chicote and Alexis Moinet and Vatsal Aggarwal},
journal= {arXiv preprint arXiv:1811.06292},
year = {2019}
}
Comments
4 pages, 1 extra for references. Accepted on Interspeech 2019