| 摘要 |
PURPOSE: To recognize the contents of a voice without deteriorating the articulation of a voiceless sound by providing a voiced sound/voiceless sound decision means which decides a voice included in a signal, a multiplication data generating means which varies the output level at the time of the voiceless-sound, a delay means which delays the signal by a necessary time, and a multiplying means which varies the output level. CONSTITUTION: This device is equipped with the voiced sound/voiceless sound decision means 12 which decides which of the voiced sound and voiceless sound the signal is, the multiplication data generating means 13 for level variation which varies the output level at the time of the voiceless sound, the delay means 14 which delays the signal by the time needed for the decision means 12 and generating means 13, and the multiplying means 15 which varies the output level. Consequently, a decrease in the articulation of the voiceless sound due to the deletion of information by the sound-speed reproducing device is eliminated by the increase in the level at the time of the voiceless sound. |
