Mozilla Updates Its Voice-Recognition Library Common Voice With 1,400 Hours Of Speech Across 19 Languages

Artificial Intelligence News

Mozilla-Updates-Its_Voice-Recognition-Library-Common-Voice-With-1400-Hours-Of-Speech-Across-19-Languages Mozilla Updates Its Voice-Recognition Library Common Voice With 1,400 Hours Of Speech Across 19 LanguagesOpen-Source Web Browser Platform, Mozilla Has Released The Latest Version Of Its Common Voice, An Open Voice-Recognition Library That Now Includes More Than 1,400 Hours Of Recorded Voice Clips From 42,000 Contributors Across 19 Languages Of English, French, Dutch, German, Hakha-Chin, Esperanto, Farsi, Basque, Spanish, Welsh, Mandarin Chinese, And Kabyle. Its Common Voice Projects Designed To Make It Easier For Developers Who Don’t Have The Resources To Develop Voice-Enabled Products That A Large Company Does.

Mozilla’s Common Voice Will Offer Researchers Access To A Large Collection Of Voices Without Any Charge. The Company Itself Plans To Utilize The Clips It Gathers To Advance Its Speech-To-Text, Text-To-Speech, And Deepspeech Engines, Where Anyone Who Could Use A Huge Collection Of Voice Clips In Dozens Of Languages Can Download The Set From The Common Voice Website. According To Mozilla, The Company, In The Coming Months, Will Test Various Approaches To Improve The Quantity And Quality Of Data That They Collected Through Community Efforts And New Partnerships. They Also Plan To Utilize Some Of The Recordings To Build Voice-Enable Productions. However, The Company Contended That Its Ultimate Goal Is To Offer More And Superior Speech Data To Those Who Look To Develop And Utilize Voice Technology.

Mozilla Intends To Provide A More Varied And Innovative Voice Technology Ecosystem. Its Common Voice Website Is One Of Our Main Vehicles For Creating Voice Data Sets That Are Constructive For Voice-Interaction Technology. Mozilla’s Common Voice That Can Be Integrated Into Deepspeech, A Suite Of Open-Source Speech-To-Text, Text-To-Speech Engines, And Trained Models, Not Only Of Voice Snippets, But Also Of Voluntarily Contributed Metadata That Constructive For Training Speech Engines Like Speakers’ Ages, Sex, And Accents.