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

Method and device for extracting voice signal of desired sound source

A voice signal and sound source technology, applied in the direction of voice analysis, instruments, etc., can solve the problems of high power consumption, huge amount of calculation, and affecting the voice recognition rate, etc., and achieve the goal of increasing hardware cost, small calculation amount, and improving voice recognition rate Effect

Active Publication Date: 2019-12-24
ACTIONS ZHUHAI TECH CO
View PDF11 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] With the above solution, due to the need to perform ICA processing, the amount of calculation is very large, resulting in huge power consumption, therefore, an advanced voice processing engine is required to perform matching processing
[0010] However, the hardware cost of an advanced speech processing engine is very high, and it is not universal. If an ordinary speech processing engine is used instead, it may not be able to support such a complex processing process, resulting in the inability to correctly identify the speech signal of the desired sound source. Affects the speech recognition rate, thereby reducing the quality of service

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
  • Method and device for extracting voice signal of desired sound source
  • Method and device for extracting voice signal of desired sound source
  • Method and device for extracting voice signal of desired sound source

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0066] In the actual use environment, the speech processing device extracts features from the input speech signal for recognition, but there are various disturbances in the environment, such as reverberation, noise, and signal distortion. These interferences make the characteristics of the input speech signal very different from those of the speech recognition model, thereby reducing the recognition rate.

[0067] In the embodiment of the present invention, this difference is minimized under the principle of blind estimation and distortion-free filtering to improve speech recognition rate without increasing hardware cost.

[0068] The preferred embodiments of the present invention will be described in further detail below in conjunction with the accompanying drawings.

[0069] refer to figure 1 As shown, in the embodiment of the present invention, the speech processing device mainly includes the following functional modules:

[0070] Acoustic Echo Cancellation (AEC) is mainl...

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 the audio processing technology, in particular to a method and device for extracting a voice signal of a desired sound source. The method and device for extracting the voice signal of the desired sound source is used for ensuring the voice recognition rate without increasing hardware costs. The method for extracting the voice signal of the desired sound source comprises the following steps: obtaining the existence probability of the desired sound source and position information of the desired sound source based on relevant characteristics of corresponding voice signalsreceived through at least two microphones; then obtaining a preset target separation coefficient; and extracting the voice signal of the desired sound source from at least two voice signals of the corresponding the voice signals by using the target separation coefficient. Thus, since a stable corresponding relationship is preset between the position information and a target separation system, a stable direction can be formed based on the position information; therefore, the corresponding target separation coefficient is obtained quickly, and the voice signal of the desired sound source is extracted quickly and accurately from the reverberation environment; and therefore, the voice recognition rate in interference environments is greatly improved without increasing the hardware costs.

Description

technical field [0001] The invention relates to audio processing technology, in particular to a method and device for extracting a desired sound source voice signal. Background technique [0002] In the prior art, in the process of collecting voice signals, in order to improve data accuracy, dual microphones are usually used to extract voice signals from a desired sound source. [0003] However, there are usually other sources of interference around the desired sound source; for example, assuming that in a meeting scene, while the speaker serving as the desired sound source is speaking, other people participating in the meeting will also participate in comments. At this time, both microphones will simultaneously collect the speech signal of the desired sound source and the speech signals of other sources. Then, how to identify the speech signal of the desired sound source from the received signals of the two microphones has become an urgent problem to be solved. [0004] Cu...

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 Applications(China)
IPC IPC(8): G10L21/0272G10L21/0308G10L21/0208G10L21/0216G10L21/0232
CPCG10L21/0208G10L21/0216G10L21/0232G10L21/0272G10L21/0308G10L2021/02087G10L2021/02165G10L2021/02166
Inventor 余立志
Owner ACTIONS ZHUHAI TECH CO
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