Speech enhancement method and device based on spatial features and electronic equipment

Abstract

The invention discloses a speech enhancement method and device based on spatial features and electronic equipment. The method comprises the steps of carrying out Fourier transform on dual-channel noisy speech to obtain a dual-channel complex spectrum; obtaining a first single-channel complex spectrum and a second single-channel complex spectrum of the dual-channel complex spectrum based on beam forming; calculating a logarithmic power spectrum of the first single-channel complex spectrum; calculating a direction energy ratio based on the energy of the first single-channel complex spectrum and the energy of the second single-channel complex spectrum, and taking a logarithm to obtain a logarithmic direction energy ratio; inputting the logarithmic power spectrum and the logarithmic direction energy ratio as features into a pre-trained speech enhancement neural network to obtain a masking value; and adding the masking value to the first single-channel complex spectrum, and performing inverse Fourier transform on the first single-channel complex spectrum after masking processing to obtain enhanced speech. According to the scheme provided by the embodiment of the invention, the speech distortion can be well avoided while the noise is effectively reduced.

Description

technical field [0001] The invention relates to the technical field of speech enhancement, in particular to a speech enhancement method, device and electronic equipment based on spatial features. Background technique [0002] Speech enhancement has always played an important role in the field of speech signal processing. The traditional speech enhancement method mainly estimates the spectral information of the noise, and then subtracts the noise from the original speech spectrum. However, sudden noise and random noise will make the spectrum The estimation of information becomes difficult. At the same time, the traditional method also needs to make independent assumptions on the signal and Gaussian assumptions on the feature distribution in advance. These assumptions are equivalent to setting boundaries for speech enhancement, resulting in limited noise reduction effects. [0003] Based on this, the neural network based on deep learning is widely used in the field of speech e...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problems of insufficient precision and low quantization efficiency of existing quantization methods, embodiments of the present invention provide a quantization method, device and electronic equipment for neural networks

Images