Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

A Fast Source Signal Reconstruction Method for Blind Sound Source Separation with Two Microphones

A dual-microphone, source signal technology, applied in speech analysis, instruments, etc., can solve the problems of fast calculation speed, small calculation amount, high noise and so on

Inactive Publication Date: 2017-06-13
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The present invention aims at the disadvantages of the background technology that there is a large amount of computation and large noise when there are many source signals, and designs a fast source signal reconstruction method for double-microphone blind sound source separation, so as to achieve a small amount of computation, fast computation speed, and restored source signal information. The purpose of high noise ratio

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 Fast Source Signal Reconstruction Method for Blind Sound Source Separation with Two Microphones
  • A Fast Source Signal Reconstruction Method for Blind Sound Source Separation with Two Microphones
  • A Fast Source Signal Reconstruction Method for Blind Sound Source Separation with Two Microphones

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0067] The invention is verified by separating and processing two measurement signals containing four source signals on MATLAB R2008a. The 4 source signals come from 4 different languages, and there are two male voices and two female voices. The measured signal length is 5.03s, and the sampling rate is 10000hz. The specific implementation steps are as follows:

[0068] Step 1: Read the measurement signal and transform it to the time-frequency domain.

[0069] Framing and windowing are performed on the two measurement signals, and then the Fourier transform of nfft points is performed to obtain the time-frequency domain representation of the measurement signals.

[0070] Among them: frame length frame_len=512, frame shift nmov=256, fft transformation points nfft=512, window function selects Hanning window, the expression is:

[0071] Step 2: Extract single-source points.

[0072] Group the same frequency points of all frame signals into groups of 4 (finally make a group wi...

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 fast source signal reconstruction method for double-microphone blind sound source separation, which belongs to the field of speech signal processing, and is especially used for fast source signal reconstruction for double-microphone sound source separation where the number of sound sources and the surrounding environment are unknown. method. This invention adds a small additional phase to each frequency component of the reconstructed signal without affecting the time-domain envelope of the reconstructed source signal, and simulates it by approximately calculating the contribution of each source signal to the time-frequency point. The proportion of the time-frequency points is used to restore the source signal, which avoids the usual process of solving underdetermined equations, simplifies the calculation steps, and achieves the purpose of speed. Therefore, compared with the existing algorithm, the method has the advantages of small calculation amount and high signal-to-noise ratio when the number of sources increases.

Description

technical field [0001] The invention belongs to the field of speech signal processing, in particular to a fast source signal reconstruction method for double-microphone sound source separation where the number of sound sources and the surrounding environment are unknown. Background technique [0002] Blind sound source separation is a hot spot in the current signal processing field. It was originally to solve the "cocktail party" problem. Now it has been used in communication systems, speech enhancement, remote sensing, medical imaging, earthquake detection, geophysics, econometrics, data mining and other fields have a wide range of applications. [0003] The so-called blind sound source separation is the process of recovering the original signal only based on the observed mixed signal when the transmission channel and signal source are unknown. According to the relationship between the number of measured signals and the number of sound sources, the problem of blind sound s...

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): G10L21/0272
Inventor 贺知明刘茜茜张山陈燚
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
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
Eureka Blog
Learn More
PatSnap group products