8 papers
Preferences of a Voice-First Nation: Large-Scale Pairwise Evaluation and Preference Analysis for TTS in Indian Languages
Srija Anand, Ashwin Sankar, Ishvinder Sethi +10
Crowdsourced pairwise evaluation has emerged as a scalable approach for assessing foundation models. However, applying it to Text to Speech(TTS) introduces high variance due to lin…
Mark My Words: A Robust Multilingual Model for Punctuation in Text and Speech Transcripts
Sidharth Pulipaka, Sparsh Jain, Ashwin Sankar +1
Punctuation plays a vital role in structuring meaning, yet current models often struggle to restore it accurately in transcripts of spontaneous speech, especially in the presence o…
RASMALAI: Resources for Adaptive Speech Modeling in Indian Languages with Accents and Intonations
Ashwin Sankar, Yoach Lacombe, Sherry Thomas +3
We introduce RASMALAI, a large-scale speech dataset with rich text descriptions, designed to advance controllable and expressive text-to-speech (TTS) synthesis for 23 Indian langua…
Rethinking MUSHRA: Addressing Modern Challenges in Text-to-Speech Evaluation
Praveen Srinivasa Varadhan, Amogh Gulati, Ashwin Sankar +8
Despite rapid advancements in TTS models, a consistent and robust human evaluation framework is still lacking. For example, MOS tests fail to differentiate between similar models,…
Towards Building Large Scale Datasets and State-of-the-Art Automatic Speech Translation Systems for 14 Indian Languages
Ashwin Sankar, Sparsh Jain, Nikhil Narasimhan +6
Speech translation for Indian languages remains a challenging task due to the scarcity of large-scale, publicly available datasets that capture the linguistic diversity and domain…
IndicVoices-R: Unlocking a Massive Multilingual Multi-speaker Speech Corpus for Scaling Indian TTS
Ashwin Sankar, Srija Anand, Praveen Srinivasa Varadhan +7
Recent advancements in text-to-speech (TTS) synthesis show that large-scale models trained with extensive web data produce highly natural-sounding output. However, such data is sca…
Rasa: Building Expressive Speech Synthesis Systems for Indian Languages in Low-resource Settings
Praveen Srinivasa Varadhan, Ashwin Sankar, Giri Raju +1
We release Rasa, the first multilingual expressive TTS dataset for any Indian language, which contains 10 hours of neutral speech and 1-3 hours of expressive speech for each of the…
Enhancing Out-of-Vocabulary Performance of Indian TTS Systems for Practical Applications through Low-Effort Data Strategies
Srija Anand, Praveen Srinivasa Varadhan, Ashwin Sankar +2
Publicly available TTS datasets for low-resource languages like Hindi and Tamil typically contain 10-20 hours of data, leading to poor vocabulary coverage. This limitation becomes…