This paper proposes a new method which reduces white gaussian noise using a variable step-size adaptive algorithm. The proposed method adjusts the step-size utilizing the fourth-order cumulant of the prediction error signal in a linear predictor. The performance of the proposed method is superior to that of fixed step-size adaptive algorithm in terms of tracking speed and steady-state error. Furthermore, a fourth-order cumulant of input signal is used to determine whether each of the audio segments is a speech or a non-speech segment and it is useful to improve the noise reduction in non-speech segments. As a result of the simulation, the effectiveness of the proposed method is confirmed.
展开▼
