| 摘要 |
<p>PURPOSE:To make it easy to discriminate between a voiceless plosive consonant and voiced consonant by extracting the features of a voice to be recognized by adding a redundancy part and by deciding a comparison between the features and information from a dictionary part stored with information corresponding to a redundant voice. CONSTITUTION:When an user pronounces the same voice twice toward microphone 1, an aural signal from microphone 1 is frequency-analyzed 4 and its features are extracted 5 and input to comparative decision part 6. In decision part 6, characters of the same sound are repeated twice and information is transferred one after another from dictionary part 3 stored with aural information corresponding to the characters and then compared to information from extraction part 5; every time information with the possibility of the input voice is found, the information is displayed at display positions D1-D4 of display part 8 through control part 7. The user, when finding a required character among characters displayed on display part 8, operates one of keys K1-K4 of operation part 2 corresponding to the display position to store the required character information in input data storage part 9. Since a voiceless plosive consonant has a voiceless pause, it can be discriminated from a voiced consonant easily.</p> |
