Voice noise reduction method and device, storage medium and electronic equipment

Abstract

The invention relates to the technical field of audio data processing, in particular to a voice noise reduction method, a voice noise reduction device, a storage medium and electronic equipment. The method comprises the steps of obtaining a to-be-processed voice signal, and performing framing and windowing processing on the to-be-processed voice signal; acquiring a preset characteristic parameterof each frame of voice signal so as to calculate noise estimation of each frame of voice signal according to the characteristic parameter; obtaining a prior signal-to-noise ratio and a posterior signal-to-noise ratio of each frame of voice signal according to the noise estimation result; performing masking processing on noise in each frame of voice signal by using a statistical masking filter. Thestatistical masking filter is constructed based on the masking effect, the prior signal-to-noise ratio and the posterior signal-to-noise ratio. According to the method and the device, the statisticalmasking filter is used for masking the voice signal subjected to noise estimation, so that residual noise can be masked to the maximum extent. The extra music noise cannot be generated. Therefore, the filtering effect is effectively improved, and the listening effect is improved.

Description

technical field [0001] The present disclosure relates to the technical field of audio data processing, and in particular to a speech noise reduction method, a speech noise reduction device, a storage medium, and electronic equipment. Background technique [0002] With the rapid development of Internet technology and smart terminals, people can use smart terminals in various environments such as indoors and outdoors to perform activities using voice functions such as live broadcasting, games, and chatting. However, the background noise is diverse and complex in different environments, so the requirement for noise suppression when the user makes speech is getting higher and higher. [0003] Existing speech noise reduction methods generally use a noise reduction algorithm to remove background noise from sound data acquired by a microphone, thereby retaining the speech data. For scenarios such as live broadcasting, the commonly used noise reduction algorithm is based on Google'...

AI Technical Summary

Problems solved by technology

However, there are certain defects and deficiencies in its actual use. For example, the noise estimation in the real environment cannot be completely accurate, and the judgment-guided method is used to calculate the prior signal-to-noise ratio, which increases the noise while smoothing the spectrum. The inaccuracy of estimation, after filtering, some noise will remain, especially when the noise is large

In addition, although the use of the Wiener filter can minimize the amount of calculation, it will generate additional music noise, which will cause discomfort in the sense of hearing and affect the user experience.

Images