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.

We describe the creation of the Artie Bias Corpus, an English dataset of expert-validated <audio, transcript> pairs with demographic tags for age, gender, accent. We also release open software which may be used with the Artie Bias Corpus to detect demographic bias in Automatic Speech Recognition systems, and can be extended to other speech technologies. The Artie Bias Corpus is a curated subset of the Mozilla Common Voice corpus, which we release under a Creative Commons CC0 license – the most open and permissive license for data. This article contains information on the criteria used to select and annotate the Artie Bias Corpus in addition to experiments in which we detect and attempt to mitigate bias in end-to-end speech recognition models. We we observe a significant accent bias in our baseline DeepSpeech model, with more accurate transcriptions of US English compared to Indian English. We do not, however, find evidence for a significant gender bias. We then show significant improvements on individual demographic groups from fine-tuning.