Technology giants are actively embracing artificial intelligence and speech recognition technology, but the construction of a voice recognition system source library needs a huge resource, not a small developer or individual developers can afford. The Mozilla Foundation is bragging about a public-friendly source and identification library called Common Voice, which will be open to third-party testing for developers who need to apply voice recognition (voice-to-speech) training in development projects The
The goal of the project is to collect more than 10,000 hours of audio sources, including various accents and is expected to be open later this year. And now everyone can contribute to the site to upload their own source, by reading the specified statement naturally, or to determine the sound of others to hear the accent and tone is accurate. There are related iOS applications for mobile device users to upload voice sources directly. Of course, Mozilla has done a very strict privacy protection for project contributors.