Voice processing method, apparatus, electronic device, and storage medium

Patent metadata

Field	Value
Publication number	US-11587574-B2
Application number	US-202117404175-A
Country	US
Kind code	B2
Filing date	Aug 17, 2021
Priority date	Feb 22, 2019
Publication date	Feb 21, 2023
Grant date	Feb 21, 2023

Abstract

Provided in the present disclosure are a voice processing method, an apparatus, an electronic device, and a storage medium, the method comprising: detecting the working state of a current call system, and when the working state is a two-end speaking state or a remote-end speaking state, performing compression processing on a subsequent remote-end voice signal, acquiring a near-end voice signal by means of a microphone, performing echo processing on the basis of the near-end voice signal and the compression-processed remote-end voice signal to obtain an echo-processed near-end voice signal and a remaining echo signal, performing non-linear suppression processing on the near-end voice signal and the remaining echo signal, and performing gain control on the suppression-processed near-end voice signal.

First claim

What is claimed is:

1. A method for processing a voice, comprising:
detecting a working state of a communication system currently;
calculating a detection probability of a dual-end talk state;
determining a second reference attenuation threshold based on a first reference attenuation threshold and the detection probability, wherein the first reference attenuation threshold is a default;
attenuation a remote-end voice signal by the second reference attenuation threshold based on the working state being a dual-end talk state or a remote-end talk state, wherein the remote-end voice signal is subsequently received, and the subsequent remote-end voice signal, whose value is larger than the second reference attenuation threshold, is attenuated by adopting the second reference attenuation threshold;
obtaining a first near-end voice signal by a microphone;
obtaining a second near-end voice signal and a residual echo signal by performing echo processing on the first near-end voice signal and the attenuated remote-end voice signal, wherein the second near-end voice signal is subjected to echo processing;
performing non-linear suppression processing on the second near-end voice signal and the residual echo signal; and
performing gain control on a third near-end voice signal, wherein the third near-end voice signal is subjected to suppression processing, and a gain value of the gain control is different from the first reference attenuation threshold and the second reference attenuation threshold.

2. The method for processing the voice according to claim 1 , wherein prior to attenuating the remote-end voice signal, the method further comprises:
calculating an energy value of the remote-end voice signal; and
determining whether the energy value is greater than the first reference attenuation threshold; and
attenuating the remote-end voice signal based on the first reference attenuation threshold when the energy value is greater than the first reference attenuation threshold.

3. The method for processing the voice according to claim 1 , wherein when the working state is the remote-end talk state, said performing gain control on the third near-end voice signal comprises:
in response to original gain value being larger than 1, gradually decreasing the original gain value to approach 1 at a certain rate;
in response to the original gain value being smaller than 1, maintaining the original gain value not changed.

4. The method for processing the voice according to claim 1 , further comprising:
performing equalization filtering processing on a low-frequency signal in the remote-end voice signal, prior to attenuating the remote-end voice signal;
said attenuating the remote-end voice signal comprises:
attenuating the remote-end voice signal subjected to equalization filtering processing.

5. Electronic equipment, comprising:
a processor; and
a memory configured to store instructions that are executed by the processor;
wherein the processor is configured to:
detect a working state of a communication system currently;
calculate a detection probability of a dual-end talk state;
determine a second reference attenuation threshold based on a first reference attenuation threshold and the detection probability, wherein the first reference attenuation threshold is a default;
attenuate a remote-end voice signal by the second reference attenuation threshold based on the working state being a dual-end talk state or a remote-end talk state, wherein the remote-end voice signal is subsequently received, and the subsequent remote-end voice signal, whose value is larger than the second reference attenuation threshold, is attenuated by adopting the second reference attenuation threshold;
obtain a first near-end voice signal by a microphone;
obtain a second near-end voice signal and a residual echo signal by performing echo processing on the first near-end voice signal and the attenuated remote-end voice signal, wherein the second near-end voice signal is subjected to echo processing;
perform non-linear suppression processing on the second near-end voice signal and the residual echo signal; and
perform gain control on a third near-end voice signal, wherein the third near-end voice signal is subjected to suppression processing, and a gain value of the gain control is different from the first reference attenuation threshold and the second reference attenuation threshold.

6. The electronic equipment according to claim 5 , wherein the processor is further configured to:
calculate an energy value of the remote-end voice signal; and
determine whether the energy value is greater than the first reference attenuation threshold; and
attenuate the remote-end voice signal based on the first reference attenuation threshold when the energy value is greater than the first reference attenuation threshold.

7. The electronic equipment according to claim 5 , wherein when the working state is the remote-end talk state, the processor is further configured to:
in response to original gain value being larger than 1, gradually decrease the original gain value to approach 1 at a certain rate;
in response to the original gain value being smaller than 1, maintain the original gain value not changed.

8. The electronic equipment according to claim 5 , wherein the processor is further configured to:
perform equalization filtering processing on a low-frequency signal in the remote-end voice signal, prior to attenuating the remote-end voice signal;
attenuate the remote-end voice signal subjected to equalization filtering processing.

9. A non-transitory computer readable storage medium, wherein when instructions in the storage medium are executed by a processor of electronic equipment, the electronic equipment is caused to perform the method for processing the voice according to claim 1 .

10. The method for processing the voice according to claim 1 , wherein when the working state is the dual-end talk state, said performing gain control on the third near-end voice signal comprises:
maintaining original gain value unchanged until the dual-end talk state is exited.

11. The electronic equipment according to claim 5 , wherein the processor is further configured to:
maintain original gain value unchanged until the dual-end talk state is exited.

Assignees

• Beijing Dajia Internet Information Tech Co Ltd

Inventors

• Zhang Chen
• Dong Pei

Classifications

• H04M1/20

  Arrangements for preventing acoustic feed-back (H04M1/62 takes precedence) · CPC title

• H04M9/08

  Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic · CPC title

• G10L21/0216 Primary

  characterised by the method used for estimating noise · CPC title

• H04M9/082 Primary

  using echo cancellers (echo cancellers per se H04B3/23) · CPC title

• G10L2021/02082

  the noise being echo, reverberation of the speech · CPC title