Unlock instant, AI-driven research and patent intelligence for your innovation.

Time-domain blind separating method for in-line natural voice convolution mixing signal

A technology of speech signal and mixed signal, applied in the field of signal processing

Inactive Publication Date: 2011-05-18
FUDAN UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method is applied in the actual speech processing, which can effectively separate the speech signals of different speakers, and solve the limitation of only a single speaker in the field of speech recognition. It is also applicable to speech communication and robot interaction. can achieve better results

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
  • Time-domain blind separating method for in-line natural voice convolution mixing signal
  • Time-domain blind separating method for in-line natural voice convolution mixing signal
  • Time-domain blind separating method for in-line natural voice convolution mixing signal

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0053] In the following, two sections of actual speech signals collected are used to separate the signals using the traditional mutual information method and the new online convolution blind separation time domain method proposed by the present invention, and the results are compared. In order to reflect the generality and robustness of the algorithm, the speakers in the processed speech signal have pauses and silences (unavoidable in practical situations).

[0054] Simultaneously in order to quantitatively compare the separation effect of the method proposed by the present invention and the existing method, we adopt the mixed convolution matrix of artificial construction to simulate the actual speech environment, and the z domain expression of its linear convolution mixed matrix is:

[0055] H ( z ) = 0.65 + 0.27 ...

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 related unsupervised learning algorithm to on-line treat speech signal for speech interactive front-end process comprises: with sufficient respect to time-domain relevance and non-stationary of natural speech signal, constructing a group of adaptive feed-forward filter networks to separate speech signal from multiple sources. This invention enhances speech signal, reduces algorithm complexityfor hardware implementation, and has wide application.

Description

technical field [0001] The invention belongs to the technical field of signal processing, and specifically relates to an unsupervised learning algorithm capable of online processing the problem of convolution blind separation of speech signals. The algorithm fully considers the time-domain correlation and non-stationary characteristics of natural speech signals, can separate speech signals from multiple sources in real time, and overcomes the system instability problem of existing algorithms when the number of sources changes. Therefore, it has broad application prospects in the field of voice signal processing such as teleconferencing, intelligent robots, and automotive communications. technical background [0002] The Blind Signal Separation Problem (BSS) refers to the process of signal propagation, when the parameters of the source signal and the propagation channel are unknown, the source is recovered only by the observed mixed signal and some statistical characteristics...

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): G10L15/28G10L15/22G10L15/00H03H17/02G10L19/26
Inventor 鲁晓丹张立明
Owner FUDAN UNIV