Mozilla Common Voice
Platform
Crowdsourced project creating open, public multilingual voice datasets.
About Mozilla Common Voice
Mozilla Common Voice: Open, Crowdsourced Voice Dataset Project
Mozilla Common Voice is a project dedicated to making voice recognition technology open and accessible to everyone by building a publicly available, multilingual dataset of voice recordings. The initiative invites contributors from around the world to donate their voices by reading prescribed sentences aloud and to validate the accuracy of other contributors’ recordings. This collaborative approach helps create a diverse and representative dataset for speech technology development.
Key Features
- Voice Donation: Contributors record voice clips by reading from a bank of donated sentences, helping to expand the dataset with diverse accents and languages.
- Validation: Users can listen to submitted voice clips and validate whether the sentences were read correctly, improving the quality and accuracy of the dataset.
- Open Source Dataset: All collected voice data is made publicly available for anyone to use, supporting open research and development in speech technology.
- Multi-language Support: The platform supports contributions in multiple languages, with a dashboard designed to be friendly for global users.
- Progress Tracking: Contributors can review their progress and stay motivated with goal-oriented dashboards and statistics.
Participation and Governance
- Community Involvement: Contributors can participate not only by donating or validating voice clips but also by engaging in decision-making processes through surveys and the Representatives Council.
- Decision-Making Structure: The Representatives Council and advisory committees, including language experts and technical advisors, guide project development and resolve conflicts using a prioritization matrix that weighs public interest and cost-benefit.
How It Works
- Record your voice by reading sentences provided on the platform.
- Validate other users’ recordings for accuracy and clarity.
- Help build a diverse, open dataset for speech technology research and development.
Getting Started
- Website: https://commonvoice.mozilla.org/en
- Contribute: Visit the website to donate your voice or validate recordings.
- Learn More: Explore the dashboard and multilingual support features directly on the homepage.
Mozilla Common Voice empowers global communities to contribute to and benefit from open, inclusive voice technology by crowdsourcing a public dataset of diverse voice recordings.