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

Unsupervised single microphone voice noise reduction method and system

A voice noise reduction and microphone technology, applied in the direction of frequency response correction, etc., can solve the problems of limited application

Active Publication Date: 2018-09-25
INST OF ACOUSTICS CHINESE ACAD OF SCI
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this type of algorithm also has some limitations, requiring specific speakers and specific noise types of training data, but in many scenarios it is difficult to obtain matching training data in advance, resulting in limited application

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
  • Unsupervised single microphone voice noise reduction method and system
  • Unsupervised single microphone voice noise reduction method and system
  • Unsupervised single microphone voice noise reduction method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0082] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0083] Such as figure 1 Shown, a kind of unsupervised single microphone speech noise reduction method, described method comprises:

[0084] Step 1) extract the spectrum of the speech training data covering all phonemes collected, then carry out k-means clustering to the amplitude spectrum, and obtain the corresponding speech dictionary of each category; then combine all the speech dictionaries of different categories into a complete Phonetic dictionary; specifically includes:

[0085] Step 101) collect a large amount of pure voices as voice training data;

[0086] Speech training data can be obtained from many open source speech databases, and the collected speech training data should cover all phonemes;

[0087] Step 102) preprocessing the speech training data collected above, and then extracting the frequency spectrum of the 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 an unsupervised single microphone voice noise reduction method. The method comprises the steps that 1, frequency spectrum extraction is conducted on acquired voice training data covering all phonemes, k-means clustering is conducted on amplitude spectra to obtain voice dictionaries corresponding to all classes, and all the different classes of voice dictionaries are combined into a complete voice dictionary WS; 2, short-time Fourier transform is conducted on a noisy voice reaching at the current moment to obtain an amplitude spectrum xt of a current frame, the amplitudespectrum is combined with the processed amplitude spectra of previous L frames to serve as an output voice spectrum X=[x<t-L>,..., x<t-1>, xt], a noise matrix WN estimated by the last frame and the voice dictionary WS are combined into a total dictionary matrix W=[WS WN], and non-negative matrix factorization is conducted on the output voice spectrum X by adopting an iterative algorithm to obtaina noise matrix and a voice noise weight vector corresponding to the current frame; and 3, a current-frame voice signal after noise reduction is reconstructed by means of the estimated noise matrix and noise weight vector.

Description

technical field [0001] The present invention relates to the field of speech signal processing, and more specifically, the present invention relates to an unsupervised single microphone speech noise reduction method and system. Background technique [0002] In many application scenarios (such as voice communication, automatic speech recognition, hearing aids), voice signals are inevitably disturbed by surrounding noise, such as road noise, wind noise, circuit noise, etc. The signal undergoes noise reduction processing. Moreover, many hearing devices (or instruments) usually have only one microphone to pick up the voice signal, and the algorithm needs to remove the noise signal from a noisy signal, which further increases the difficulty of solving the problem. [0003] The traditional single-microphone speech noise reduction algorithm mainly includes two parts: the noise estimation part and the gain calculation part. This type of algorithm generally assumes that the noise is...

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): H04R3/04
Inventor 李军锋李煦颜永红
Owner INST OF ACOUSTICS CHINESE ACAD OF SCI
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