Speech noise reduction system based on combined filter
Ruoting Liu · 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS) · 2022
People's communication efficiency will be affected by noise, especially in non face-to-face voice communication. These noises can only be weakened as much as possible but can not be completely eliminated. The method proposed in this paper uses the system filter of combined band-pass filter and Wiener filter to weaken three noisy speech with different signal-to-noise ratio respectively -5dB, -10dB and -12dB. By comparing signal-to-noise ratio improvement of single band-pass filter that is 2.133289dB, 2.574053dB and 2.573509dB and system filter that is 11.312919dB, 15.730713dB and 17.015864dB, it can be concluded that the overall filtering effect is better than that of single filter. Therefore, the combined algorithm of band-pass filter and Wiener filter proposed in this paper overcomes the limitation of a single band-pass filter, improves the noise reduction performance, and has certain practical value.