English

Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis

Computation and Language 2018-03-28 v1 Machine Learning Sound Audio and Speech Processing

Abstract

In this work, we propose "global style tokens" (GSTs), a bank of embeddings that are jointly trained within Tacotron, a state-of-the-art end-to-end speech synthesis system. The embeddings are trained with no explicit labels, yet learn to model a large range of acoustic expressiveness. GSTs lead to a rich set of significant results. The soft interpretable "labels" they generate can be used to control synthesis in novel ways, such as varying speed and speaking style - independently of the text content. They can also be used for style transfer, replicating the speaking style of a single audio clip across an entire long-form text corpus. When trained on noisy, unlabeled found data, GSTs learn to factorize noise and speaker identity, providing a path towards highly scalable but robust speech synthesis.

Keywords

Cite

@article{arxiv.1803.09017,
  title  = {Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis},
  author = {Yuxuan Wang and Daisy Stanton and Yu Zhang and RJ Skerry-Ryan and Eric Battenberg and Joel Shor and Ying Xiao and Fei Ren and Ye Jia and Rif A. Saurous},
  journal= {arXiv preprint arXiv:1803.09017},
  year   = {2018}
}