Mozilla has updated the dials of voice data common voice , including examples of the pronunciation of more than 200 thousand people. Data published as a public treasure ( cc0 ). The proposed sets can be used in machine learning systems to build models of recognition and synthesis of speech.
Compared to the last update, the volume of speech material in the collection increased from 23.8 to 25.8 thousand hours of speech. More than 88 thousand people who had 3161 hours of speech took part in the preparation of materials in English (there were 84 thousand participants and 3098 hours). A set for the Belarusian language covers 7903 participants and 1419 hours of speech material (there were 6965 participants and 1217 hours), the Russian language – 2815 participants and 229 hours (there were 2731 participants and 215 hours), Uzbek – 2092 participants and 262 hours (there were 2025 participants and 258 hours), the Ukrainian language – 780 participants and 87 hours (there were 759 participants and 87 hours).
The Common Voice project is aimed at organizing joint work to accumulate the base of voice templates, taking into account the whole variety of voices and manners of speech. Users are invited to voice phrases displayed on the screen or evaluate the quality of data added by other users. The accumulated database with records of various pronunciations of typical phrases of human speech without restrictions can be used in machine learning systems and in research projects.