Speech synthesis

 

System feedback is mostly given by speech prompts  played back to the user (production of audio signals equivalent to human speech). Speech output is the only possibility one has in telephone-based applications. (For a detailed account on speech synthesis see Chapter 12.)

The system's response generation (selection of the appropriate message) is achieved by the dialogue module and will not be described here. The system's responses through speech outputs are of many types which can be itemised as follows:

• The speech messages are recorded and stored, and optionally compressed, then played back to the user (like a computer-based digital recorder).
• Elementary speech messages are recorded, stored, and concatenated to produce the system's messages (canned speech ).
• The speech messages are generated through a text-to-speech synthesiser (TTS). 

 
• Speech recording, storage, and playback
• Canned speech
• Text-to-speech synthesis
• Multi-linguality