English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this paper, we propose a text-to-speech (TTS)-driven data augmentationmethod for improving the quality of a non-autoregressive (AR) TTS system.Recently proposed non-AR models, such as FastSpeech 2, have successfullyachieved fast speech synthesis system. However, their quality is notsatisfactory, especially when the amount of training data is insufficient. Toaddress this problem, we propose an effective data augmentation method using awell-designed AR TTS system. In this method, large-scale synthetic corporaincluding text-waveform pairs with phoneme duration are generated by the AR TTSsystem and then used to train the target non-AR model. Perceptual listeningtest results showed that the proposed method significantly improved the qualityof the non-AR TTS system. In particular, we augmented five hours of a trainingdatabase to 179 hours of a synthetic one. Using these databases, our TTS systemconsisting of a FastSpeech 2 acoustic model with a Parallel WaveGAN vocoderachieved a mean opinion score of 3.74, which is 40% higher than that achievedby the conventional method.

TTS-by-TTS: TTS-driven Data Augmentation for Fast and High-Quality Speech Synthesis