Method and device of denoising voice signal

Patent metadata

Field	Value
Publication number	US-11475907-B2
Application number	US-201716766236-A
Country	US
Kind code	B2
Filing date	Dec 20, 2017
Priority date	Nov 27, 2017
Publication date	Oct 18, 2022
Grant date	Oct 18, 2022

Abstract

The present disclosure provides a method and a device of denoising a voice signal. The method portion includes the following steps: filtering out an environmental noise signal in an original input signal according to an interference signal related to the environmental noise signal in the original input signal to obtain a first voice signal; obtaining a sample signal matching the first voice signal from a voice signal sample library; and filtering out other noise signal in the first voice signal according to the sample signal matching the first voice signal, to obtain an effective voice signal. The method provided by the present disclosure may effectively filter out the environmental noise signal and other noise signal in the voice signal.

First claim

What is claimed is:

1. A method of denoising a voice signal, comprising:
filtering out an environmental noise signal in an original input signal according to an interference signal related to the environmental noise signal in the original input signal to obtain a first voice signal;
obtaining a sample signal matching the first voice signal from a voice signal sample library; and
filtering out other noise signal in the first voice signal according to the sample signal matching the first voice signal, to obtain an effective voice signal;
wherein the obtaining a sample signal matching the first voice signal from a voice signal sample library, comprises:
performing voiceprint recognition on the first voice signal, and obtaining a spectrum feature of the first voice signal from the first voice signal after being voiceprint recognized; and
obtaining a sample signal matching the first voice signal from a voice signal sample library according to the spectrum feature of the first voice signal;
wherein the performing voiceprint recognition on the first voice signal, and obtaining a spectrum feature of the first voice signal from the first voice signal after being voiceprint recognized, comprises:
performing windowing on the first voice signal to obtain at least one frame of voice signal;
performing Fourier transform on the at least one frame of voice signal to obtain at least one frame of frequency domain signal; and
extracting a spectrum feature of the at least one frame of frequency domain signal to obtain the spectrum feature of the first voice signal;
wherein the extracting a spectrum feature of the at least one frame of frequency domain signal to obtain the spectrum feature of the first voice signal, comprises:
selecting one frame of frequency domain signal from the at least one frame of frequency domain signal as a first frequency domain signal;
mapping a signal amplitude at each frequency in the first frequency domain signal to a grayscale value in accordance with a preset amplitude-grayscale mapping relationship; and
taking the grayscale value corresponding to each frequency in the first frequency domain signal as the spectrum feature of the first voice signal.

2. The method according to claim 1 , wherein the obtaining a sample signal matching the first voice signal from a voice signal sample library according to the spectrum feature of the first voice signal, comprises:
calculating a similarity between the spectrum feature of the first voice signal and a spectrum feature of each sample signal stored in the voice signal sample library; and
taking a sample signal with the highest similarity to the spectrum feature of first voice signal as the sample signal matching the first voice signal.

3. The method of claim 1 , wherein the filtering out other noise signal in the first voice signal according to the sample signal matching the first voice signal, to obtain an effective voice signal, comprises:
calculating other noise value in each frame frequency domain signal by adopting a least mean square algorithm according to the sample signal matching the first voice signal;
subtracting other noise value in each frame frequency domain signal from each frame frequency domain signal to obtain an effective frequency domain signal of each frame;
performing inverse Fourier transform on the effective frequency domain signal of each frame to obtain an effective time domain signal of each frame; and
combining the effective time domain signal of each frame in sequence to obtain the effective voice signal.

4. The method according to claim 1 , wherein prior to the filtering out an environmental noise signal in an original input signal according to an interference signal related to the environmental noise signal in the original input signal, the method further comprising:
collecting the original input signal through a first microphone within a first specified distance from a sound source; and
collecting the interference signal through a second microphone outside the first specified distance and within a second specified distance from the sound source;
and wherein the second specified distance is greater than the first specified distance.

5. The method according to claim 1 , wherein the filtering out an environmental noise signal in an original input signal according to an interference signal related to the environmental noise signal in the original input signal to obtain a first voice signal, comprises:
filtering out environmental noise signal in the original input signal by adopting a least mean square algorithm according to the interference signal related to the environmental noise signal in the original input signal to obtain the first voice signal.

6. An electronic device, comprising: a processor and a memory connected to the processor;
the memory is used to store one or more computer instructions; and
the processor is used to execute the one or more computer instructions for:
filtering out an environmental noise signal in an original input signal according to an interference signal related to the environmental noise signal in the original input signal to obtain a first voice signal;
obtaining a sample signal matching the first voice signal from a voice signal sample library; and
filtering out other noise signal in the first voice signal according to the sample signal matching the first voice signal, to obtain an effective voice signal;
wherein the obtaining a sample signal matching the first voice signal from a voice signal sample library, comprises:
performing voiceprint recognition on the first voice signal, and obtaining a spectrum feature of the first voice signal from the first voice signal after being voiceprint recognized; and
obtaining a sample signal matching the first voice signal from a voice signal sample library according to the spectrum feature of the first voice signal;
wherein when performing voiceprint recognition on the first voice signal, and obtaining a spectrum feature of the first voice signal from the first voice signal after being voiceprint recognized, the processor is specifically used for:
performing windowing on the first voice signal to obtain at least one frame of voice signal;
performing Fourier transform on the at least one frame of voice signal to obtain at least one frame of frequency domain signal; and
extracting a spectrum feature of the at least one frame of frequency domain signal to obtain the spectrum feature of the first voice signal;
wherein when extracting a spectrum feature of the at least one frame of frequency domain signal to obtain the spectrum feature of the first voice signal, the processor is specifically used for:
selecting one frame of frequency domain signal from the at least one frame of frequency domain signal as a first frequency domain signal;
mapping a signal amplitude at each frequency in the first frequency domain signal to a grayscale value in accordance with a preset amplitude-grayscale mapping relationship; and
taking the grayscale value corresponding to each frequency in the first frequency domain signal as the spectrum feature of the first voice signal.

7. The electronic device according to claim 6 , wherein when obtaining a sample signal matching the first voice signal from a voice signal sample library according to the spectrum feature of the first voice signal, the processor is specifically used for:
calculating a similarity between the spectrum feature of the first voice signal and a spectrum feature of each sample signal stored in the voice signal sample library; and
taking a sample signal with the highest similarity to the spectrum feature of first voice signal as the sample signal matching the first voice signal.

8. The electronic device according to

Assignees

• Goertek Tech Co Ltd

Inventors

• Chen Weiliang

Classifications

• G10L2021/02087

  the noise being separate speech, e.g. cocktail party · CPC title

• G10L17/00

  Speaker identification or verification techniques · CPC title

• G10L21/0208 Primary

  Noise filtering · CPC title

• G10L21/0308

  characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques · CPC title

• G10L17/02

  Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction · CPC title