Computer-assisted pronunciation tutoring (CAPT) methods have been well-established in research and education. Common system approaches include the phonetic quality assessment, highlight problematic sections in the speech signal and usually rely on automatic speech recognition (ASR) regarding the target language L2. The contribution deals with the audiovisual CAPT system AzAR. An extensive feedback mechanism and several speech databases for Slavonic learners of German were developed. Currently, AzAR is adapted for Chinese Learners of German and for learners of the Basque language. This extended abstract summarizes some experiences of the multilingual experiments and system tests - focusing on phonetic and prosodic assessment aspects.
Pronunciation tutoring (CAPT), Slavonic languages, Mandarin Chinese, German, Basque
© Association for Information and Communication Technologies www.ict-act.org Developed by tdelev