Gelp: GAN-excited linear prediction for speech synthesis from mel-spectrogram

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Researchers

Research units

  • Research Organization of Information and Systems National Institute of Informatics
  • University of Edinburgh

Abstract

Recent advances in neural network -based text-to-speech have reached human level naturalness in synthetic speech. The present sequence-to-sequence models can directly map text to mel-spectrogram acoustic features, which are convenient for modeling, but present additional challenges for vocoding (i.e., waveform generation from the acoustic features). High-quality synthesis can be achieved with neural vocoders, such as WaveNet, but such autoregressive models suffer from slow sequential inference. Meanwhile, their existing parallel inference counterparts are difficult to train and require increasingly large model sizes. In this paper, we propose an alternative training strategy for a parallel neural vocoder utilizing generative adversarial networks, and integrate a linear predictive synthesis filter into the model. Results show that the proposed model achieves significant improvement in inference speed, while outperforming a WaveNet in copy-synthesis quality.

Details

Original languageEnglish
Title of host publicationProceedings of Interspeech
Publication statusPublished - 1 Jan 2019
MoE publication typeA4 Article in a conference publication
EventInterspeech - Graz, Austria
Duration: 15 Sep 201919 Sep 2019
https://www.interspeech2019.org/

Publication series

NameInterspeech - Annual Conference of the International Speech Communication Association
ISSN (Electronic)2308-457X

Conference

ConferenceInterspeech
CountryAustria
CityGraz
Period15/09/201919/09/2019
Internet address

    Research areas

  • GAN, Neural vocoder, Source-filter model, WaveNet

Download statistics

No data available

ID: 38773946