Abstract
In our study, we have built a speed-displaying system called “KanNon”[1] which allows speech communication for people who are hard of hearing. When a hard of hearing person communicates with others, they currently employ sign-language and lip reading. The KanNon system is a new language display system, which uses sound spectrogram to display speech directly and to better convey the speaker's emotions. The Chinese characters that form the KanNon system's name literally means “seeing sound,” The system estimates the spectrum of the sound inputted from microphone with a parametric model and displays speech information, such as pitch and voice strength, in real time.
