The present doctoral research project is being developed in the following sequential phases: 1. Creation of a motion capture data and audiovisual expressive speech database for Brazilian Portuguese; 2. Study of sample-based synthesis techniques suitable for the reproduction of expressive speech; 3. Development of an expressive talking head for Brazilian Portuguese; 4. Proposal of a strategy to generalize the synthesis methodology to a compact database of samples captured under partially controlled conditions. The project is entering into its third phase, moving towards the proposal of an expressive speech synthesis methodology with a pilot implementation being developed for Brazilian Portuguese.
展开▼
