[Univ of Cambridge]alt[Dept of Engineering]

MIL Speech Seminars 2004-2005

The MIL Speech Seminar series schedule for Easter Term 2005 is as follows:

17th May 2005 Paul Taylor (MIL Visitor) Unit Selection Speech Synthesis This talk will give an overview of the state of the art in generation natural sounding speech. Unit selection is the current dominant technique for speech synthesis and works by a process of search through a large database of recorded speech. The "training" phase for a unit selection system involves carefully recording and analysing speech from a single talker. This is then linguistically annotated, and at run time a search is used to match units of annotated speech with like units found from text analysis of the input sentence. The talk with give an overview of this technique, but will then concentrate on the current issues of research. The talk will also discuss issues of automatic training and evaluation, and end with a discussion of current trends and future progress.
5th July 2005 Heiga Zen (Nagoya Institute of Technology, Japan) Reformulating the HMM as a trajectory model by imposing explicit relationship between static and dynamic features A trajectory model, derived from the HMM by imposing explicit relationship between static and dynamic features, is developed and evaluated. The derived model, named "trajectory-HMM", can alleviate some limitations of the standard HMM, which are i) piece-wise constant statistics within a state and ii) conditional independence assumption of state output probabilities, without increasing the number of model parameters. In this talk, a Viterbi-type training algorithm is also derived. This model was evaluated both in speech recognition and synthesis experiments. In speaker-dependent continuous speech recognition experiment, the trajectory-HMM achieved error reduction over the standard HMM. The experimental results of subjective listening tests shows that introduction of the trajectory-HMM can improve the quality of synthetic speech generated from HMM-based speech synthesis system which we have proposed.