A binaural speech enhancement method based on deep learning

A speech enhancement and binaural technology, applied in speech analysis, stereo systems, instruments, etc., can solve the problems of poor non-stationary noise suppression and no special processing of target speech space information, so as to improve robustness and suppress Effects of noise interference and binaural speech enhancement

Active Publication Date: 2021-03-23
INST OF ACOUSTICS CHINESE ACAD OF SCI
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In most traditional speech enhancements with two-channel output, most of them only consider removing interference, and there is no special processing for the spatial information of the target speech, and the suppression effect on non-stationary noise is not good.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A binaural speech enhancement method based on deep learning
  • A binaural speech enhancement method based on deep learning
  • A binaural speech enhancement method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

[0032] figure 1 It is a flowchart of a binaural speech enhancement method based on deep learning. Such as figure 1 shown, including:

[0033] Step S101: Framing, windowing, and Fourier transform are respectively performed on the left channel noisy speech signal and the right channel noisy speech signal to obtain the left channel noisy spee...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a binaural speech enhancement method based on deep learning, comprising the following steps: respectively processing a left / right channel noisy speech signal containing a target speech signal to be enhanced so as to obtain a left / right frequency domain signal, and combining amplitudes thereof to obtain single-channel complex features; using the frequency domain signal of the left / right channel and corresponding target frequency domain signal theoretical value to calculate corresponding target speech ideal complex masking respectively; combining to form a target speech single-channel complex masking theoretical value, combining the single-channel complex features to train a complex feedforward neural network so as to obtain a binaural speech enhancement model; usingthe target speech single-channel complex masking estimated value outputted by the model to respectively process the left / right channel noisy speech signal so as to obtain a left / right channel frequency domain signal; and finally obtaining a corresponding target speech time-domain signal. By the method, noise interference can be suppressed and spatial information of a target sound source can be maintained. By making full use of the generalization ability of deep neural networks, the enhancement of binaural speech is achieved.

Description

technical field [0001] The invention relates to the technical field of speech enhancement, in particular to a binaural speech enhancement method based on deep learning. Background technique [0002] At present, speech enhancement technology is mainly to remove background noise and directional noise interference in speech signals, improve speech quality and intelligibility, and achieve better performance in speech recognition and human ear understanding. In the enhancement technology with single-channel speech as the output, the background noise can be suppressed by using the different characteristics of speech and noise in the time-frequency domain of the single-channel input, and the spatial information of the target speech and interference signals in the multi-channel input can be better. Remove directional noise. In binaural hearing, the human ear can use the spatial information difference between the target and the interference signal in the dual-channel speech to impro...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G10L21/0232G10L25/30
CPCG10L21/0232G10L25/30H04S2420/01
Inventor 李军锋孙兴伟夏日升颜永红
Owner INST OF ACOUSTICS CHINESE ACAD OF SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap