Development overview of TTS-MK speech synthesizer for Macedonian language, and its application

Workshop

Authors

Slavcho Chungurski , Sime Arsenovski and Dejan Gjorgjevikj

Abstract

This paper shows the current results of development of TTS-MK – a speech synthesizer for Macedonian language. The basic principles for projecting and building of speech synthesizer for Macedonian language, based on con-catenation of speech segments, are shown. Because concatenative speech synthesizers are based on a collection of speech segments, the creation of the speech corpus for Macedonian language is described. Every language has its respective and specific speech norms and characteristics that should be observed during the speech synthesis. The Macedonian language is phonetic and every voice has its particular grapheme, hence the normative pronunciation does not contain great difficulty. However, there are some deviations of this approach, mainly in the written and spoken part of the language. These specific deviations should be taken into consideration according to the norms of the normative pronunciation. The inconsistencies in the normative pronunciation often are due to the influence of the speech in parts of the country can be expressed as the following forms: discharging, substitution, inserting and replacing of the voice. There are some issues about the pronunciation of the vocals positioned next to each other in different situations and possibilities. Furthermore, the pronunciation of the vowel R is observed, and the situation of the consonants sonority. The presentation also focuses on the accent in the Macedonian language. The accent is dynamic and positioned on the third syllable. Taking all this into consideration, the rules and regulations for the accent positioning in the Macedonian language can be easily derived. However, there are certain deviations, especially in the case of adopted foreign words and when the accent is considered for group of words. There are two versions of the system based on different segments corpora. Both of them are presented and the improvements of the second version, as well as its application are briefly explained.

Keywords

Text-To-Speech, Macedonian Language, TTS, TTS-MK, Orthoepy, Speech API – SAPI