Frontier speech & language research, from India.
Built on sovereign speech technology since 2008 — delivering impact through ASR, TTS, translation, and assistive voice systems for Indian languages.
Explore by category
Automatic Speech Recognition
Multi-dialect and dysarthric-speech recognition systems for Indian languages.
Funded & supported by
Research
What we work on
Speech Recognition
Automatic Speech Recognition, multi-dialect ASR, and silent speech recognition aim to develop accurate, inclusive speech technologies capable of understanding diverse Indian languages, dialects, accents, and communication patterns.
Speech Synthesis & Prosody
Text-to-speech synthesis, prosody modelling, and voice conversion tuned for natural, expressive speech.
Language & Translation
Machine translation systems built for the linguistic structure of Indian Languages.
Assistive Speech technology
Speech-enabled assistive devices for cerebral palsy, autism spectrum disorder, and speech-input aids.
Signal Processing & Audio
Speech enhancement, music signal processing, and array microphone-based speech analysis focus on improving audio quality, separating speech and music signals, reducing noise, and enabling accurate spatial audio analysis.
Our thesis
“To significantly reduce communication barriers, between human beings and between humans and machines, that are predominantly caused by a language barrier, a disability, or illiteracy.”
- 01Develop a multidisciplinary environment to promote fundamental and applied research in speech and audio.
- 02Build state-of-the-art, socially-beneficial applications that enhance human–computer and human–human interaction.
- 03Promote collaborative research with industry to bridge the gap between research and available technology.
- 04Provide a forum to interact with other researchers, and offer training and consultancy to those who need it.
Standalone Speech-to-Speech Translator for English, Hindi and Tamil Languages
Assistive Speech Technologies — NLTM BHASHINI
Prosody Modelling for TTS — NLTM BHASHINI
Publications
Two years of published research.
20 papers since the start of 2025 alone — spanning sign language recognition, dysarthric speech, and whisper-to-speech conversion.
Leveraging Synthetic Speech for Dysarthric Speech Recognition
TTS-driven data augmentation that improves recognition accuracy for dysarthric speakers, published in Computer Speech & Language, Vol. 100.
People
Take a look at the people behind the lab.
Faculty, research scholars, and students working across speech and language technology.