Mozilla's Common Voice Library now offers up to 1,400 hours of speech

Mozilla's Common Voice Library include 18 different languages, comprises clips from 42,000 contributors

Mozilla may not have its personal voice based assistant like Google but it is taking initiative ro help startups, app developers build their own voice based app, services.

Mozillas Common Voice Library now offers up to 1,400 hours of speech

Mozilla's Common Voice Library might help researchers, developers to use voice clips and create their voice-based assistants.

To make that possible, Mozilla has released its  Common Voice Library, an open source collection of human voices, in 18 different languages including Dutch, Hakha-Chin, Esperanto, Farsi, Basque, Spanish, French, German, Mandarin Chinese (Traditional), Welsh and Kabyle.

The open source collection of transcribed voice data from Mozilla comprises over 1,400 hours of voice samples from 42,000 contributors including linguists, professionals working in voice technologies.

Notably, Mozilla launched its common voice library in 2017. Back then the company said that its aim was to "build a speech corpus that's free, open source, and big enough to create meaningful products with.”

When Mozilla made the Common Voice Library publicly available eight months ago, the library had 500 hours of voice samples from 20,000 contributors, and with current 1,400 hours the company claims its one the largest multi-lingual dataset of its kind. The company is said to be planning to use the clips it collects to improve Text-to-speech, speech-to-text and DeepSDeepeech engines.

 According to a VentureBeat report, Mozilla is working to bring its open source collection of transcribed voice data in 70 other languages and it is said to be 'actively underway' via Common Voice webaite and mobile apps.

As per the report, Mozilla will roll out an improved Common Voice web tool that will have simplified prompts to help vary clip-to-clip, and include new controls for reviewing, re-recording, and skipping clips.

While Mozilla's open voice-recognition library is said to help those nascent developers and researchers to create their own voice-enabled assistants who doesn't have the resources of big companies like Apple, Google, the company's  effort of providing a huge library of human voices will help users as well to access information and use a massive collection of voice clips in dozens of languages that too for free.

Tech2 is now on WhatsApp. For all the buzz on the latest tech and science, sign up for our WhatsApp services. Just go to Tech2.com/Whatsapp and hit the Subscribe button.






also see

science