The design of a speech synthesis system for Afrikaans

Abstract

The design of a speech synthesis system for the generation of speech from text is presented. The system accepts any Afrikaans sentence and then derives segmental and suprasegmental informa­tion through various linguistic processes. The ­resultant information is used to generate control parameters for a formant synthesizer. Relevant features of the most well-known experimental­ speech synthesis systems are given, followed by a specification of the design objectives of the system. The system design is ex­plained in terms of data modelling and data flow analysis. Further explanation is provided for the implementation of the data structure, the data structure manager and a typical linguistic process.