摘要 |
The reference speech data of foreign language textbooks for introductory, intermediate, and advanced courses are stored in a database (16). When one of the sentences in a lesson menu is selected on a CRT (26), the corresponding reference speech data are reproduced, and a voice synthesizer (34) replays model speech, after which the user repeats using a microphone (20). The user speech input from the microphone (20) is subjected to spectrum analysis in a speech recognizer (22) of the computer to obtain speech data, which are then compared with the reference speech data of the database (16) to evaluate the user's pronunciation. If the user's pronunciation is determined to be so good that the user can communicate sufficiently, a passing mark appears on the screen, and another sentence follows. If a passing mark does not appear, the same sentence may have to be repeated over and over. Since the users can evaluate their pronunciation of foreign language by themselves, pronunciation is effectively practiced by repetition. |