Fundamental frequency by rule for a text-to-speech system

An algorithm for the generation of fundamental frequency (Fo) in a speech synthesis-by-rule system is described. Input consists of syntactic information (parts-of-speech, and a syntactic structure tree) and lexical information (number of syllables, location of lexical stress, and simple phonemic information). The output Fo values are used to synthesize the voiced portions of the generated speech. The algorithm utilizes research findings of the relationship of Fo to such linguistic factors as stress, syntactic phrases, and sentence type. Word types and syntactic boundaries affect the degree and direction of Fo movements, respectively, and questions have high Fo.