Abstract: Speech signal processing and modeling in systems for continuous speech recognition and Text-to-Speech synthesis in Serbian language are described in this paper. Both systems are fully developed by the authors and do not use any third party software. Accuracy of the speech recognizer and intelligibility of the TTS system are in the range of the best solutions in the world, and all conditions are met for commercial use of these solutions.
Keywords: Signal processing, speech processing, speech recognition, text-to-speech synthesis.