Supercharge Your Innovation With Domain-Expert AI Agents!

Channel Taylor series-based noise estimation method

A technique of noise estimation and Taylor series, which is applied in the field of noise estimation and can solve the problems that affect the application and the large amount of calculation.

Active Publication Date: 2018-05-18
南京土星视界科技有限公司
View PDF10 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, VTS noise estimation is performed in the cepstrum domain, which involves relatively complex matrix operations and a large amount of calculation, which will affect its application on mobile terminals and other devices

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
  • Channel Taylor series-based noise estimation method
  • Channel Taylor series-based noise estimation method
  • Channel Taylor series-based noise estimation method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0013] Below in conjunction with specific embodiment, further illustrate the present invention, should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention, after having read the present invention, those skilled in the art will understand various equivalent forms of the present invention All modifications fall within the scope defined by the appended claims of the present application.

[0014] Such as figure 1 As shown, the noise estimation method based on channel Taylor series mainly includes feature extraction in training phase, model training, feature extraction in testing phase and CTS noise estimation module. The specific implementation of each main module in the drawings will be described in detail below one by one.

[0015] 1. Feature extraction

[0016] First, fast Fourier transform (FFT: FastFourierTransform) is performed on the pure training speech or noisy test speec...

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 channel Taylor series-based noise estimation method. According to the method, Mel frequency logarithm coefficients (MFLC) are extracted from pure training speech; the pure MFLCs of all speech units are adopted to perform training so as to generate a Gaussian mixture model; and a nonlinear relationship between noise-containing test speech and pure training speech is approximated by using the CTS on each Mel channel, and the parameters of the pure Gaussian mixture model are transformed, so that the noise parameters of the noise-containing test speech can be obtained. According to the method of the invention, matrix calculation of traditional noise estimation is simplified into scalar calculation, and therefore, the operation complexity of a system can be significantly reduced under the premise of guaranteeing the accuracy of parameter estimation.

Description

technical field [0001] The invention belongs to the field of speech processing, and in particular relates to a noise estimation method for extracting characteristic parameters of a speech signal in a logarithmic spectral domain, and estimating the mean value and variance of environmental noise on each Mel channel through parameter adaptation of a Gaussian mixture model. Background technique [0002] In speech processing fields such as speech communication and speech recognition, the impact of environmental noise on speech is often unavoidable, which will lead to a decline in speech quality and affect the performance of the speech processing system. [0003] Enhancing or compensating the noisy speech features extracted in the test environment is one of the effective means to improve the speech quality. Estimation of noise parameters is the key technology of speech enhancement and feature compensation. In a stationary noise environment, the mean and variance of the noise can ...

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): G10L21/0216G10L15/06G10L15/20G10L25/24
CPCG10L15/063G10L15/20G10L21/0216G10L25/24
Inventor 吕勇
Owner 南京土星视界科技有限公司
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com