Emotional speech synthesis and its application to pervasive E-learning

With the significant development of information science and computer technology in recent years, Human-computer communication becomes the key technology for computer system to be personate and intelligent. Speech Synthesis has bright and potential prospect in the field of human-computer communication. Intelligibility, variability and naturalness are the features of speech that have been mostly used to measure the performance of synthetic speech. Thus, we have improved our speech synthesis system with a prosodic control module. Since there are more and more intercourse between Chinese and European, especially the Olympic Games will be held in Beijing, English becomes one of the most useful languages for the international communication. Most English learners need to find a way to improve their spoken language, and the learning places should not be restricted to traditional classrooms. Pervasive computing is the next generation computing paradigm, providing the right information to the right person at the right time no matter where he is. In this paper, a Pervasive Computing English E-learning System is presented applying HPC model, the LAAD method^[1], speech recognition and synthesis technologies. Especially, an improved stepwise FD (frequency-domain) &TD PSOLA emotional speech synthesis method is described in detail.