The Mozilla Foundation, makers of the Firefox browser, have launched a new project called Common Voice, which the organization hopes to become the first open-source voice recognition engine on the market.
Mozilla launched Common Voice in mid-June, and the project is currently in a training phase. The organization is asking users to help train the engine by reading small pieces of text or by verifying the accuracy of previous voice recordings.
The aim is to collect at least 10,000 hours of voice recordings that Mozilla engineers feel would be enough to train their voice recognition system. Mozilla said it plans to release the Common Voice database into open-source later in 2017.
The organization says it embarked on this project because of a lack of an open-source voice recognition system on the market.
Currently, all voice recognition engines are locked up behind proprietary code at various companies, such as Amazon (Alexa), Apple (Siri), and Microsoft (Cortana), just to name a few.
"We think that stifles innovation," Mozilla says, "so we’re launching Project Common Voice, a project to help make voice recognition open to everyone."
Common Voice is to provide an accessible voice recognition technology to ordinary developers who can't afford to invest billions in Cortana-like technologies.
Users who want to contribute their voice to the Common Voice database can do so on the project's website. They don't need fancy microphones or sound-proof rooms to read out supplied texts. Mozilla engineers say they want to collect data from real-world environments, so it's OK if there's background noise or the user has an accent. The texts users are asked to read are in English.