Speech enhancement method and device

Abstract

The invention discloses a speech enhancement method and device. The method comprises steps as follows: the characteristic quantity of noise in a silence section of a speech signal is acquired; a noise class matched with the noise in the silence section is determined from multiple preset noise classes according to the characteristic quantity of the noise in the silence section, and the multiple noise classes are acquired after clustering of multiple noise samples according to characteristic information of the multiple noise samples; a noise model corresponding to the noise class matched with the noise in the silence section is determined according to the noise class matched with the noise in the silence section as well as the mapping relation between the noise class and the noise model; speech enhancement is performed on the speech signal according to the noise model corresponding to the noise class matched with the noise in the silence section. With the adoption of the speech enhancement method and device, speech enhancement is performed on the speech signal according to the noise model corresponding to the noise class, and the speech enhancement effect can be improved.

Description

technical field [0001] Embodiments of the present invention relate to the field of speech processing, and more specifically, to a method and device for speech enhancement. Background technique [0002] With the rapid development of communication technology and network technology, voice communication has far surpassed the traditional fixed telephone as the main form, and has been widely used in many fields such as mobile communication, TV / teleconferencing, vehicle hands-free communication, IP telephone, etc. widely used. In the application of voice communication, it is a challenging problem to ensure the clarity and high quality of the voice signal and to eliminate various noises in the voice signal caused by these new voice communication methods. [0003] At present, the biggest difficulty in speech enhancement in speech communication lies in the unknown (or the uncertainty of statistical characteristics) of the noise environment. The speech enhancement method based on hid...

AI Technical Summary

Problems solved by technology

[0005] However, the number of noise samples and noise models included in the noise sample and noise model mapping relationship list is limited. When the noise in the input speech signal does not have a relatively close noise sample in the mapping relationship list, according to the noise sample and noise model The enhancement effect of the noise model determined by the mapping relationship list is poor, so the universality of this method needs to be strengthened

In addition, this method requires separate training and model building for each noise, which requires a large storage space, which limits its practical usability

Images