Introduction - If you have any usage issues, please Google them yourself
The background noise is the most common factor degrading the quality and intelligibility of speech in recordings. Spectral subtraction can lower the noise level without affecting the speech signal quality. In this paper, we have studied and implemented to reduce the background noise of the voice signal, and to make improvements to reduce the background noise of the experiment. Not only with the spectral subtraction of the noisy signal of the transformation, we improve the traditional spectral subtraction based on the introduction of the parameters, by adjusting the parameters(α、β、γ) to achieve better enhancement.