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

Frequency domain voice blind separation method for multi-frequency-band switching call media node (CMN) nonlinear function

A nonlinear function and sub-band technology, applied in speech analysis, instruments, etc., can solve the problems of CMN algorithm performance fluctuation and performance limitation, and achieve the effect of improving speech separation performance, improving overall performance and stable performance.

Inactive Publication Date: 2012-07-04
DALIAN UNIV OF TECH
View PDF4 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a frequency-domain voice blind separation method for switching CMN nonlinear functions in frequency bands, to match the Gaussian and symmetric changes of different voices, or the Gaussian and symmetric of the same voice in the whole frequency band Change to solve the performance fluctuation and performance limitation of the single nonlinear function CMN algorithm

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
  • Frequency domain voice blind separation method for multi-frequency-band switching call media node (CMN) nonlinear function
  • Frequency domain voice blind separation method for multi-frequency-band switching call media node (CMN) nonlinear function

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] A specific embodiment of the present invention will be described in detail below in conjunction with the technical scheme and accompanying drawings.

[0016] Assume that there are voice mixed signals recorded by two microphones, denoted as x1, x2, which contain voices s1, s2 of two speakers speaking at the same time, that is, x1, x2 are mixed signals of s1, s2, s1, s2 are in x1, x2 interfere with each other. In order to obtain two pure speeches s1, s2, it is necessary to separate the mixed signals x1, x2. The specific separation steps are as shown in the accompanying drawings.

[0017] In the first step, x1, x2 are subjected to windowing and framing processing and STFT transformation to obtain frequency-domain speech mixed signals x1(f, t), x2(f, t).

[0018] The second step is the low frequency and intermediate frequency segmentation. Any frequency in the range of 100Hz to 300Hz, such as 200Hz, can be selected as the cut-off point, and the frequency domain voice mixe...

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 frequency domain voice blind separation method for multi-frequency-band switching call media node (CMN) nonlinear function, which belongs to the technical field of speech enhancement and is characterized in that frequency domain speech is divided into the two frequency bands of low frequency and middle frequency based on kurtosis distribution characters. Three types of multi-frequency-band schemes are applied for switching a nonlinear function of a plurality of CMN algorithms, at least one scheme is led to be most matched with Gaussian performance and symmetry of frequency domain speech. Compared with single nonlinear function CMN algorithms, the frequency domain voice blind separation method for multi-frequency-band switching CMN nonlinear function is capable ofbeing adapted to voice changes in terms of Gaussian performance and symmetry, and remarkably improves voice separation performance. When an ordinary amplitude correlation method is adopted for voice sequence regulation of all frequency points, the separation signal to noise ratio of two paths of voice can be increased by 11dB at most, and the frequency domain voice blind separation method is stable is performance, easy in software and hardware achievement, and capable of being widely used in key technologies of computer perception and decision-making, unmanned driving and the like so as to achieve the speech enhancement function, and further improves entire performance of voice signal processing tasks such as voice recognition and content understanding.

Description

technical field [0001] The invention relates to a speech enhancement method, in particular to a frequency domain speech blind separation method. Background technique [0002] Speech recognition and content understanding are important functions in national key technologies such as computer perception and decision-making, and driverless driving. However, since speech in natural environments is often disturbed by environmental noise, multi-party conversations, etc., its signal-to-noise ratio and intelligibility are greatly reduced, and in severe cases, speech recognition and content understanding may fail. Therefore, eliminating various speech interferences (ie, speech enhancement) is the most important part of speech signal processing. Because the characteristics of speech and noise are complex and changeable, people have been exploring stable and effective speech enhancement methods, but they still face great challenges. [0003] Traditional speech enhancement methods inclu...

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
IPC IPC(8): G10L21/02G10L21/0272
Inventor 林秋华
Owner DALIAN UNIV OF TECH
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