Build up TTS Prototypes

• set up architecture
• set up interfaces
• set up evaluation API
• specify language resources
• localise modules to UK ?English

Evaluation of TTS Prototypes and Modules

• set up evaluation criteria
• set up evaluation procedure

research topics

Text processing

Research is focused on techniques allowing robust grapheme-to-phoneme conversion and POS-tagging in an multilingual environment. A special research direction is the pronunciation of native and non native names.

Prosody Generation

Research concerns the generation of symbolic and acoustic prosody. Specific topics are the investigation of rhythm and emotional speech.

Acoustic Synthesis

The framework is based on non uniform unit selection. Research topics are

• finding appropriate sub units
• automatic segmentation
• concatenation and manipulation of speech units

The use of transducers (formalism of finite state machines) is explored

General research topics

Voice conversion

Dissemination

Organisation of Workshops