R Joshi, N Garera - International Conference on Speech and Computer, 2023 - Springer
… Code-mixed TTS is also relevant in these applications since the product … text is in a regional language. In this work, we describe our approaches for production quality code-mixed Hindi-…
S Zhao, TH Nguyen, H Wang, B Ma - arXiv preprint arXiv:2010.08136, 2020 - arxiv.org
… text-to-speech (TTS) synthesis models have dramatically improved intelligibility and naturalness of generated speech from text. … corpus and the converted speech in the other language, …
Y Cao, S Liu, X Wu, S Kang, P Liu, Z Wu… - … Acoustics, Speech …, 2020 - ieeexplore.ieee.org
… (CS) speech with consistent voice using only monolingual … (PPG) based CS speech synthesizer using only monolingual … Sitaram and AW Black, “Speechsynthesis of code-mixedtext,…
… text. This paper describes our two stage approach used for text normalization developed as part of our ongoing speech … Choudhury, “Pos tagging of English hindi code-mixed social …
S Guo, L Wang, S Li, J Zhang, C Gong, Y Wang… - … , ICONIP 2020, Bangkok …, 2020 - Springer
… Speechsynthesis aims to generate human-like speech with high … -end text-to-speech system allow us to produce a human-like synthesizedspeech with an integrated end-to-end text-to-…
D Sravani, R Mamidi - Proceedings of the 27th Conference on …, 2023 - aclanthology.org
… code-mixedtext which trained using a silver parallel corpus created by filtering a large synthetic code-mixed … of code-mixed sentences (Sec.3.1.1) and distributions in humangenerated …
… For the baseline model, we test Bnc, and Bncr code-mixed translation without training on CM text in a multilingual manner (MMT), ie, {Hi, Hir, Bn, Bnr}⇌En + Hirn→En. Interestingly, MMT …
M Chai, S Guo, C Gong, L Wang… - … Automatic Speech …, 2021 - ieeexplore.ieee.org
… In this study, to address the poor performance of speechsynthesis at CS points, we investigate two different textrepresentation methods: implicit and explicit. For implicit method, we use …
… the effective use of CS text-tospeech (TTS) … text from CS-Real as input and generate speech samples from these speaker-adapted models that amount to roughly 100 hours of speech …