Presenter(s): Johan Classon (Conversy)
Event: MDN 2022 Online Conference
This talk goes through several different voice technologies now, makes a case for how they can be applied now, and predicts what the development might look like in the near future. The solution presented recognizes languages and can switch code and language models accordingly and in real time. Gender and age can be determined, which is useful for further metadata-driven services and analytics. Speakers can be identified and verified through a so-called voice-print, which can have applications in journalism and news coverage.