Global syllable set for building speech synthesis in Indian languages

Raghavendra, E. Veera; Desai, Sanjay; Yegnanarayana, B.; Black, Alan W.; Prahallad, Kishore

doi:10.1109/slt.2008.4777837

Cited by 13 publications

(6 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There have been previous attempts to train generic voice models from different perspectives-polyglot synthesis [17]- [28], code-mixing 2 [23], [26], [29]- [36], cross-lingual voice conversion [37], [38], and data augmentation [21], [22], [25], [26], [28], [39]- [44]. Polyglot synthesis aims to synthesise texts of multiple languages in the voice of a single speaker.…”

Section: Related Workmentioning

confidence: 99%

Exploring the Role of Language Families for Building Indic Speech Synthesisers

Prakash

Murthy

2023

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Building end-to-end speech synthesisers for Indian languages is challenging, given the lack of adequate clean training data and multiple grapheme representations across languages. This work explores the importance of training multilingual and multi-speaker text-to-speech (TTS) systems based on language families. The objective is to exploit the phonotactic properties of language families, where small amounts of accurately transcribed data across languages can be pooled together to train TTS systems. These systems can then be adapted to new languages belonging to the same family in extremely low-resource scenarios.TTS systems are trained separately for Indo-Aryan and Dravidian language families, and their performance is compared to that of a combined Indo-Aryan+Dravidian voice. We also investigate the amount of training data required for a language in a multilingual setting. Same-family and cross-family synthesis and adaptation to unseen languages are analysed. The analyses show that language family-wise training of Indic systems is the way forward for the Indian subcontinent, where a large number of languages are spoken.Index Terms-end-to-end speech synthesis, Indian languages, language families, low-resource• This work is one of the first attempts to study the importance of language families in the context of speech synthesis.• We compare language family-specific Indo-Aryan (IA) and Dravidian (Dr) models with a combined Indo-Aryan+Dravidian (IA+Dr) system. • We also assess the performance of models trained in datastressed situations. We reduce the training data used per language in the multilingual voice.

show abstract

Section: Related Workmentioning

confidence: 99%

Exploring the Role of Language Families for Building Indic Speech Synthesisers

Prakash

Murthy

2023

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

show abstract

“…These tools are not able to cope when subjected to syllables, due to the increase in the length of the sub-word unit (syllables are longer in length than phonemes). The total number of phonemes in Tamil language is only 41 [14] whereas the total number of syllables is definitely greater than the phonemes (number of syllables in any language will be in thousands) [32]. This increase in the total count of sub-word unit and the usage of a bi-gram prediction model hinders the performance of these data driven toolkits.…”

Section: Related Workmentioning

confidence: 99%

Formulation of Syllable Based Pronunciation Models for Tamil Text-to-Speech Synthesizer

Rajendran

Kumar

2020

MJCS

View full text Add to dashboard Cite

The primary aim of Human-Computer Interaction (HCI) is to deliver the power of computers and communication systems to people in an easily accessible and understandable form. HCI in a person’s native/first language is always invigorate. Developing a Tamil Text-To-Speech (TTS) system will facilitate a convenient medium of interaction for people who speak Tamil language. This paper emphasizes on the development of pronunciation models, a vital component of a Tamil TTS. Developing a pronunciation model for Tamil is more arduous when compared to other languages due to the non-triviality between the letter to sound correspondence. Veritably, two syllable-based pronunciation models developed by us are discussed in this paper. First, is a syllable-centric rule-based pronunciation model that generates a well-founded training data which is ingrained into the second, Conditional Random Field (CRF) enforced model. It is evident that both of these models are dominions with a high Mean Similarity Score of 0.97 and 0.94 respectively in comparison to the other existing rule driven and data driven models in the literature. These syllable-based pronunciation models will enrich the performance of a Tamil TTS.

show abstract

“…Most existing work in speech synthesis for Indian languages uses unit selection [13] with syllable-like units [14,15]. Re- cently, based on the observation that Indian languages share many commonalities in phonetics, a language independent phone set was proposed, and was used in building statistical parametric (HMM-based) speech synthesis systems [8].…”

Section: Statistical Speech Synthesismentioning

confidence: 99%

DNN-based Speech Synthesis for Indian Languages from ASCII text

Ronanki¹,

Gangireddy²,

Bollepalli³

et al. 2016

9th ISCA Workshop on Speech Synthesis Workshop (SSW 9)

View full text Add to dashboard Cite

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.

show abstract

Global syllable set for building speech synthesis in Indian languages

Cited by 13 publications

References 8 publications

Exploring the Role of Language Families for Building Indic Speech Synthesisers

Exploring the Role of Language Families for Building Indic Speech Synthesisers

Formulation of Syllable Based Pronunciation Models for Tamil Text-to-Speech Synthesizer

DNN-based Speech Synthesis for Indian Languages from ASCII text

Contact Info

Product

Resources

About