Sensory Partners on Technical Innovation Award Winning Solution

LONDON, ENGLAND - The VISTA (Virtual human Interface for a Set-Top box Agent) project won the Royal Television Society Technical Innovation Award 2003 for advanced engineering techniques. With the goal of improving visually impaired viewers' access to digital TV, the recently completed 18-month project developed a prototype Electronic Program Guide that viewers query by voice. The system then performs a search of standard EPG data and answers the viewers' queries in synthetic speech. The prototype system enables visually impaired viewers to select the programs they want to watch from the hundreds of digital channels. Once the viewer has decided what to watch, the VISTA system automatically changes to the appropriate digital TV channel, or, if the program is on at a later time, stores a reminder. "The VISTA partners have worked hard to make their vision of the VISTA system a reality," said Dr. Jonathan Freeman, VISTA project manager for the ITC. "I am delighted that the valuable work of the VISTA project towards improving the accessibility of digital TV to visually impaired viewers has been recognized by this prestigious award." Led by the Independent Television Commission, VISTA project partners (British Sky Broadcasting Limited, City University London, Sensory Inc., Televirtual Limited, The University of East Anglia and The Victoria University of Manchester) have integrated a range of technologies into its software, including a speaker independent speech recognition system, virtual human (avatar) technology, a speech synthesis system and EPG search software. Key development challenges were the optimization of the speech recognition capability of the prototype, effective and reliable integration of the technologies, and structuring the user-system interaction to enable viewers to search for the programs they want quickly and easily. The system was developed on the basis of extensive initial research on the television viewing habits, behaviors and needs of visually impaired and elderly viewers and the final integrated prototype was optimized on the basis of the results of user evaluations of early versions of the prototype.