SPEECH SYNTHESIZER BASED ON TIME DOMAIN SYLLABLE CONCATENATION

Ljubomir Josifovski, Dragan Mihajlov, Dejan Gorgevik

Abstract: In [2] we have presented a subsystem for text-to-speech (TTS) conversion for macedonian language as a part of a system for support of humans with damaged eyesight [1]. In this paper we present the speech synthesizer which is part of the TTS conversion subsystem. It's based on time-domain syllable concatenation. A novel module for duration and fundamental frequency (F₀) modification is introduced and discussed. We believe that the architecture presented is well suited to the nature of macedonian language, and fits well with prerequisites for real-time operation on a standard, of-the-shelf hardware. The prototype containing inventory of 1275 syllables and implementing modules for duration and F₀ modification was built and tested. The preliminary tests concerning the eligibility of the synthesized speech are most encouraging.