| 摘要 |
PURPOSE: A method of removing an echo using a vocoder variable information rate is provided to operate an adaptive echo remover without changing a filter coefficient when a present frame is decided to a nonvoice section, thereby increasing stability of the adaptive echo remover. CONSTITUTION: A variable rate vocoder compresses a voice signal at several rates, and inputs the signal to a simultaneous call detector and an adaptive filter updater(30). The simultaneous call detector decides whether a compressed present frame is below a 1/8 rate, and if so, the simultaneous call detector recognizes a state as a simultaneous speech state. If the present frame is over the 1/8 rate, the simultaneous call detector does not recognize the state as the simultaneous speech state and updates an adaptive filter coefficient to operate an adaptive echo remover(32, 34). If the state is the simultaneous call state, the simultaneous call detector does not update the adaptive filter coefficient, and operates the adaptive echo remover by applying a present coefficient(33). |
