Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Multi-Channel Speech Enhancement Method Based on Time-Frequency Domain Binary Mask

A binary mask and speech enhancement technology, which is applied in speech analysis, instruments, etc., can solve the problems of estimation distortion and speech enhancement effect decline, and achieve the effect of eliminating influence and high speech enhancement effect

Active Publication Date: 2022-07-29
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0010] In the above scheme, when the SNR of the received signal continues to be high or low, the algorithm’s estimation of beamforming related parameters is severely distorted, resulting in a decline in the effect of speech enhancement

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 Multi-Channel Speech Enhancement Method Based on Time-Frequency Domain Binary Mask
  • A Multi-Channel Speech Enhancement Method Based on Time-Frequency Domain Binary Mask
  • A Multi-Channel Speech Enhancement Method Based on Time-Frequency Domain Binary Mask

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0041] The basic idea of ​​the present invention is to create a new beamforming parameter estimation method by constructing a binary mask estimation based on the estimation value of the speech existence probability in the time-frequency domain, and use the binary mask estimation to classify the time-frequency components of the signal , to eliminate the influence of the noise part on beamforming as much as possible.

[0042] Example steps such as figure 1 shown:

[0043] Step 1. Generate convolutional neural network CNN input features according to the voice signal data, and estimate the probability of voice existence.

[0044] Suppose the received signal in the time-frequency domain is:

[0045] X i (f k )=a i (f k , θ)·S i (f k )+N i (f k )

[0046] where S i (f k ) is the ith frame frequency f k The sound source signal component of , a i (f k , θ)∈C M×1 represents the array pair f k Steering vector of frequency signal, N i (f k ) ∈ C M×1 zero mean additiv...

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 relates to a multi-channel speech enhancement method based on a time-frequency domain binary mask for an array received speech signal. Using the network model to output the probability estimation of speech existence, a binary mask is calculated, and the classification of the signal in the time-frequency domain and the estimation of the corresponding beamforming parameters are realized through the binary mask, so as to obtain a better speech enhancement effect. The realization process is as follows: first, the network model is used to estimate the probability of speech existence in the time-frequency domain for the received signal of the array, and then the estimation result and the received signal are used to calculate the threshold, so as to calculate the binary mask estimation and beamforming related parameter estimation to realize multi-channel. Voice enhancement. Compared with the existing speech enhancement algorithm of the array received signal, the present invention has higher output signal-to-noise ratio and PESQ score of subjective speech quality evaluation.

Description

technical field [0001] The invention belongs to beam forming technology, and particularly relates to a multi-channel language enhancement technology for time-frequency domain binary mask estimation. technical background [0002] With the research and development of pattern recognition and machine learning, some methods have been borrowed from the field of speech enhancement, and a series of speech enhancement algorithms that combine machine learning and multi-channel speech enhancement have emerged. Compared with traditional multi-channel speech enhancement algorithms, these algorithms use machine learning models to perform specific mask estimation on the received signal, and then more accurately estimate beamforming related parameters, which can avoid the spatial distribution of the microphone array and the target direction. A priori assumptions to obtain better speech enhancement performance. However, there are many types of machine learning models, and the features of sp...

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): G10L21/0216G10L21/0224G10L21/0232
CPCG10L21/0216G10L21/0224G10L21/0232G10L2021/02166
Inventor 江家麒
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com