DEC voice synthesiser

Press, 1 May 1984, Page 20

DEC voice synthesiser

This $BOOO box from Digital Equipment Corporation is a sophisticated voice synthesiser which the company claims is a considerable advance over existing products.

The box is a self-con-tained unit that usually connects between a computer and a visual display unit and turns text displayed on the visual display unit into speech. Called DECtalk, it works with computers that use an R 5232 line and the ASCH character set.

DECtalk can speak in a number of different voices. The programmer can control the pitch, resonance and speed of delivery, plus many other factors. The voice can be male, female, young or old.

DECtalk does not usd stored digitalised speech or vocabulary - synthesised voice technology. It has built in rules that allow it to pronounce words, pronounce amounts displayed in figures and to differentiate between words that look the same in print but are pronounced differently according to the context — for instance St (saint) and St (street). DECtalk can also parse a sentence and place expression and emphasis according to punctuation. The programmer can also tell DECtalk about colloquial pronunciation. The existing program has American spelling and pronunciation (for example,

reesearch instead of research). These pronunciations can be easily re-programmed. DECtalk was developed by graduate students at the Massachussets Institute of Technology over a five-year period. Digital computers were used in the development and Digital refined the initial concept into a marketable product. Mr Harker, general manager of Digital Equipment Corporation (New Zealand), said he believed the most significant applications of DECtalk would be an aid to professional proof readers and to enable blind and visually impaired people full use of computer equipment.