DNN-based Speech Synthesis for Indian Languages from ASCII text

TL;DR

This paper explores deep neural network-based methods to synthesize speech from noisy ASCII transliterations of Indian languages, demonstrating competitive quality across Hindi, Tamil, and Telugu, and releasing datasets publicly.

Contribution

It evaluates three approaches for converting ASCII transliterations to speech using DNNs, addressing the challenge of noisy, non-standard text input.

Findings

Models produce speech comparable to native script-based synthesis.

Supervised G2P approach outperforms naive methods.

Datasets are publicly released for further research.

Abstract

Text-to-Speech synthesis in Indian languages has a seen lot of progress over the decade partly due to the annual Blizzard challenges. These systems assume the text to be written in Devanagari or Dravidian scripts which are nearly phonemic orthography scripts. However, the most common form of computer interaction among Indians is ASCII written transliterated text. Such text is generally noisy with many variations in spelling for the same word. In this paper we evaluate three approaches to synthesize speech from such noisy ASCII text: a naive Uni-Grapheme approach, a Multi-Grapheme approach, and a supervised Grapheme-to-Phoneme (G2P) approach. These methods first convert the ASCII text to a phonetic script, and then learn a Deep Neural Network to synthesize speech from that. We train and test our models on Blizzard Challenge datasets that were transliterated to ASCII using crowdsourcing. Our experiments on Hindi, Tamil and Telugu demonstrate that our models generate speech of competetive quality from ASCII text compared to the speech synthesized from the native scripts. All the accompanying transliterated datasets are released for public access.