Speech output from a computer-controlled water-supply network

A system is described which provides for the output of information from a real engineering database in spoken form. Data extracted from the database is converted by the system, using its own predefined knowledge of the information domain and a knowledge of simple English, into a sequence of words and an associated pitch contour. The spoken output is then generated by the concatenation and resynthesis of previously analysed stored isolated words using a hardware digital speech synthesiser.