Text-to-speech technologies for mobile telephony services

Abstract

Text-To-Speech (TTS) systems aim to transform arbitrary textual input into spoken output. At first glance, this may seem a relatively simple task of determining the phonetic sounds of the input and outputting a corresponding sequence of audible signals. However, it is in fact quite a difficult task to produce intelligible and natural results in the general case. This is due to linguistic and vocalization subtleties at various levels that human speakers take for granted when interpreting written text. In fact, the task requires a considerable grasp of both Natural Language Processing and Digital Signal Processing techniques. The potential application of such functionality is varied, including its use for language edu- cation, as an aid to handicapped persons, for implementing talking books and toys, for vocal monitoring and other man-machine communication facilities. The area is currently being explored in order to address its application for the Maltese language within the context of the mobile communications industry. This paper’s main aim is to provide a brief overview of current TTS approaches and techniques, and the way these may be implemented. For further insight, reference should be made to the selected bibliography.