This is an example of the MacinTalk speech synthesis system. The MacinTalk system consists of three components: the MacinTalk speech driver, the Reader English to phonetics package, and the Exceptions Editor program. The Reader package accepts input in the form of unrestricted English text, and using its internal rules, converts the input to phonetic codes suitable as input to MacinTalk. Since English spelling often bears little relationship to pronounciation, Reader will mis-pronounce some words. The Exceptions Editor was created to allow the user to explicitly control how Reader will pronounce particular words. Using the Exceptions Editor, words and their phonetic transcriptions are entered into an exceptions file. Reader will scan this file first, and, if found, use the stored phonetic transcription for the word in question. The final component, MacinTalk, accepts input from Reader or directly from the user, and produces high quality speech in real time. This entire discourse is being converted from English text to speech in real time, as you currently hear it. The MacinTalk system gives the user explicit control of pitch, speaking rate, and pitch contour. In the future, MacinTalk may also be able to speak with a female voice and in languages other than English.##
