SPEECH TECHNOLOGIES FOR PRONUNCIATION AND PROSODY TRAINING OF MACEDONIAN LANGUAGE

Abstract

In this paper, we introduce a concept of Computer assisted pronunciation training (CAPT) e-learning platform on Macedonian language where the issues for speech technologies (Automatic Speech recognition and synthesis) on Macedonian language are also described. Review of the speech technologies used in CAPT is given and detailed description of the proposed framework with the development steps and a number of technological challenges is presented. The presented CAPT concept will take advantages of speech technologies to provide audio-visual feedback to the language learners and provides language training on both subsegmental (pronunciation) and suprasegmental level (prosody). The specifications for the required speech databases (native speech database, non-native speech database and source-language accent database) and the speech corpora development are described. Also, general guidelines are given for preparation of the speech material and its lexical structure for prosody tests and exercises, based on the contrastive analysis research studies for Macedonian as L2.