Peter Bell, Joris Driesen, Steve Renals
Centre for Speech Technology Research, University of Edinburgh, UK