Smarter speech technology to enable smooth-talking gadgets

Voice-activated devices that can interact with people in a natural, intelligent way could soon be available, thanks to a major new project involving researchers at the Universities of Sheffield, Edinburgh, and Cambridge.

Researchers will use advanced computing techniques to create a combined system that can not only understand the spoken word well, but can also synthesize voices that sound natural. A University of Sheffield release reports that the technology will use machine learning techniques to enable it to become familiar with a user’s voice.

Scientists say the technology could have many benefits, including improved voice-activated computers and communications technology. It may also help to develop Web search engines for audio clips.

Technology that responds to speech commands could soon be able to recognize an individuals’ voice and become familiar with a person’s vocabulary, accent, and vocal expressions.

The technology could also be able to follow the flow of a conversation and to cope with changes in background noise.

The release notes that scientists at the three universities involved aim to develop the systems over the next five years in a £6.2 million project funded by the U.K. Engineering and Physical Sciences Research Council.

Dr. Thomas Hain, who is leading the research at the University of Sheffield’s Department of Computer Science, said: “Speech technology is clearly on its way into the mainstream, but the key to its success is human-like performance. We are working on the foundations of technology that will bring speech to new users and exciting new applications, from clinical assistive technology to robotics, from schools to board rooms.”

Speech technology also has the potential to create voice-controlled devices for the home, which could help older people stay independent and could also be used in speech therapy.

Professor Phil Green from the University of Sheffield leads this theme.

Professor Steve Renals, of the University of Edinburgh’s School of Informatics, who is leading the project, said: “We are working to develop technologies that can recognize and generate natural-sounding speech. This could open the door to computer speech technology becoming commonplace throughout our lives – at home, at work, and in our leisure time.”