IndicVoices: Towards building an Inclusive Multilingual Speech Dataset for Indian Languages

TL;DR

This paper introduces INDICVOICES, a large, diverse multilingual speech dataset from India, along with an open-source blueprint for data collection, enabling the development of inclusive speech recognition models for 22 Indian languages.

Contribution

It provides a comprehensive, scalable data collection framework and the first multilingual ASR model supporting all 22 Indian languages.

Findings

Created 7348 hours of speech data from diverse speakers.

Developed INDICVOICES, a multilingual speech dataset for Indian languages.

Built IndicASR, the first ASR supporting all 22 Indian languages.

Abstract

We present INDICVOICES, a dataset of natural and spontaneous speech containing a total of 7348 hours of read (9%), extempore (74%) and conversational (17%) audio from 16237 speakers covering 145 Indian districts and 22 languages. Of these 7348 hours, 1639 hours have already been transcribed, with a median of 73 hours per language. Through this paper, we share our journey of capturing the cultural, linguistic and demographic diversity of India to create a one-of-its-kind inclusive and representative dataset. More specifically, we share an open-source blueprint for data collection at scale comprising of standardised protocols, centralised tools, a repository of engaging questions, prompts and conversation scenarios spanning multiple domains and topics of interest, quality control mechanisms, comprehensive transcription guidelines and transcription tools. We hope that this open source blueprint will serve as a comprehensive starter kit for data collection efforts in other multilingual regions of the world. Using INDICVOICES, we build IndicASR, the first ASR model to support all the 22 languages listed in the 8th schedule of the Constitution of India. All the data, tools, guidelines, models and other materials developed as a part of this work will be made publicly available