Concepedia

Common Voice: A Massively-Multilingual Speech Corpus

Rosana Ardila, Megan Branson, Kelly Davis, Michael Henretty, Michael Köhler, Josh Meyer, Reuben Morais, Lindsay Saunders, Francis M. Tyers, Gregor Weber

arXiv (Cornell University) · 2019 · 209 citations · 6 references

DOI Full text

Open access

TL;DR

Common Voice is a large, multilingual speech corpus of transcribed audio designed primarily for automatic speech recognition but also useful for other language tasks, currently covering 38 languages. The authors demonstrate the corpus by training and evaluating Mozilla DeepSpeech models on it. The corpus is built via crowdsourced collection and validation, and the authors use it to train DeepSpeech models. The corpus now contains 2,500 hours of audio from over 50,000 volunteers, making it the largest public speech recognition dataset, and transfer‑learning experiments on twelve languages achieved a 5.99% average character‑error‑rate improvement, providing the first published end‑to‑end ASR results for many languages.

Abstract

The Common Voice corpus is a massively-multilingual collection of transcribed speech intended for speech technology research and development. Common Voice is designed for Automatic Speech Recognition purposes but can be useful in other domains (e.g. language identification). To achieve scale and sustainability, the Common Voice project employs crowdsourcing for both data collection and data validation. The most recent release includes 29 languages, and as of November 2019 there are a total of 38 languages collecting data. Over 50,000 individuals have participated so far, resulting in 2,500 hours of collected audio. To our knowledge this is the largest audio corpus in the public domain for speech recognition, both in terms of number of hours and number of languages. As an example use case for Common Voice, we present speech recognition experiments using Mozilla's DeepSpeech Speech-to-Text toolkit. By applying transfer learning from a source English model, we find an average Character Error Rate improvement of 5.99 +/- 5.48 for twelve target languages (German, French, Italian, Turkish, Catalan, Slovenian, Welsh, Irish, Breton, Tatar, Chuvash, and Kabyle). For most of these languages, these are the first ever published results on end-to-end Automatic Speech Recognition.