Granted Patent B2

US 12,512,111 · App. 18/354,715 · Granted Dec 30, 2025

Inventors: Le Xiao (Shenzhen, CN); Chengqian Zhang (Shenzhen, CN); Fengyun Liao (Shenzhen, CN); Xin Qi (Shenzhen, CN)

Assignee: SHENZHEN SHOKZ CO., LTD.

G10L21/0264G10L21/02G10L21/0208G10L25/21G10L2021/02166

Abstract

Claims (92)

1 . A method for voice enhancement, comprising:

obtaining a first signal and a second signal of a target voice, the first signal being a signal of the target voice collected based on a first position, and the second signal being a signal of the target voice collected based on a second position;

determining a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal;

determining, based on the first signal and the second signal, a plurality of parameters related to a plurality of sound source directions, each parameter of the plurality of parameters corresponding to a probability that a sound is emitted from a sound source direction to form the first signal and the second signal;

determining, based on the plurality of parameters and the position of the target voice, a second coefficient, comprising:

determining a direction of a synthetic sound source based on the plurality of parameters, including:

determining that a sound source direction corresponding to a parameter with the greatest value among the plurality of parameters is the direction of the synthetic sound source; and

determining, based on the direction of the synthetic sound source and the position of the target voice, the second coefficient; and

obtaining a voice-enhanced first output voice signal corresponding to the target voice by processing, based on the first coefficient and the second coefficient, the first signal and/or the second signal.

2 . The method of claim 1 , wherein the determining a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal comprises:

obtaining a signal pointing to a first direction and a signal pointing to a second direction by performing, based on the position of the target voice, the first position, and the second position, a differential operation on the first signal and the second signal, the signal pointing to the first direction and the signal pointing to the second direction containing different proportions of a valid signal;

determining a third signal corresponding to the valid signal based on the signal pointing to the first direction and the signal pointing to the second direction; and

determining the first coefficient based on the third signal.

3 . The method of claim 2 , wherein the determining a third signal corresponding to the valid signal comprises:

determining a fourth signal by performing an adaptive differential operation on the signal pointing to the first direction and the signal pointing to the second direction; and

obtaining the third signal by enhancing a low-frequency component in the fourth signal.

4 . The method of claim 3 , further comprising:

updating one or more adaptive parameters of the adaptive differential operation based on the fourth signal, the signal pointing to the first direction, and the signal pointing to the second direction.

5 . The method of claim 1 , wherein the determining a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal, comprises:

obtaining a signal pointing to the first direction and a signal pointing to the second direction by performing, based on the position of the target voice, the first position, and the second position, a differential operation on the first signal and the second signal, the signal pointing to the first direction and the signal pointing to the second direction containing different proportions of a valid signal;

determining an estimated signal-to-noise ratio (SNR) of the target voice based on the signal pointing to the first direction and the signal pointing to the second direction; and

determining the first coefficient based on the estimated SNR.

6 . The method of claim 1 , wherein the determining, based on the first signal and the second signal, a plurality of parameters related to a plurality of sound source directions comprises:

determining one or more parameters related to each sound source direction by performing, based on the each sound source direction, the first position, and the second position, a differential operation on the first signal and the second signal.

7 . The method of claim 1 , wherein the determining, based on the direction of the synthetic sound source and the position of the target voice, the second coefficient comprises:

determining whether the position of the target voice is located in the direction of the synthetic sound source;

in response to determining that the position of the target voice is located in the direction of the synthetic sound source, setting the second coefficient as a first value; or

in response to determining that the position of the target voice is not located in the direction of the synthetic sound source, setting the second coefficient as a second value.

8 . The method of claim 1 , wherein the determining, based on the direction of the synthetic sound source and the position of the target voice, the second coefficient comprises:

determining, based on an angle between the position of the target voice and the direction of the synthetic sound source, the second coefficient using a regression function.

9 . The method of claim 1 , further comprising:

performing a smoothing operation on the second coefficient based on a smoothing factor.

10 . The method of claim 1 , further comprising performing at least one of the following operations on the first signal and the second signal:

framing the first signal and the second signal;

performing a windowed smoothing operation on the first signal and the second signal; or

converting the first signal and the second signal to a frequency domain.

11 . The method of claim 10 , further comprising:

determining a fourth coefficient based on a power level difference between the first signal and the second signal; and

obtaining a voice-enhanced fourth output voice signal corresponding to the target voice by processing, based on the first coefficient, the second coefficient, and the fourth coefficient, the first signal and/or the second signal.

12 . The method of claim 11 , wherein the determining a fourth coefficient based on a power level difference between the first signal and the second signal comprises:

obtaining a noise power spectral density based on a silent interval in the first signal and the second signal;

obtaining the power level difference based on a first power spectral density of the first signal, a second power spectral density of the second signal, and the noise power spectral density; and

determining the fourth coefficient based on the power level difference and the noise power spectral density.

13 . The method of claim 1 , further comprising:

determining at least one target sub-band signal in the first output voice signal; and

obtaining a second output voice signal by processing, based on a single-mic filtering algorithm, the at least one target sub-band signal.

14 . The method of claim 13 , wherein the determining at least one target sub-band signal in the first output voice signal comprises:

obtaining a plurality of sub-band signals based on the first output voice signal;

determining an SNR of each of the sub-band signals; and

determining, based on the SNR of each of the sub-band signals, the target sub-band signal.

15 . The method of claim 1 , further comprising:

determining a third coefficient by processing, based on a single-mic filtering algorithm, the first signal and/or the second signal; and

obtaining a third output voice signal by processing, based on the third coefficient, the first output voice signal.

16 . The method of claim 1 , wherein

the determining, based on the first signal and the second signal, a plurality of parameters related to a plurality of sound source directions comprises:

constructing a plurality of directional microphones whose poles point to the plurality of sound source directions, respectively, wherein

a cardioid shape corresponding to each of the plurality of sound source directions is a simulated cardioid shape;

a pole of each simulated cardioid shape points to a corresponding sound source direction;

and

determining, based on the first signal and the second signal, a likelihood function corresponding to the each sound source direction in the plurality of sound source directions, wherein a cardioid shape corresponding to the likelihood function is a synthetic cardioid shape;

and

the determining a direction of a synthetic sound source based on the plurality of parameters further comprises:

determining a simulated cardioid shape that is closest to a pole of an actual cardioid shape; and

determining a sound source direction corresponding to the simulated cardioid shape as the direction of the synthetic sound source.

17 . A system for voice enhancement, comprising:

at least one storage medium including a group of instructions; and

at least one processor in communication with the at least one storage medium, wherein when executing the group of instructions, the at least one processor is directed to cause the system to:

obtain a first signal and a second signal of a target voice, the first signal being a signal of the target voice collected based on a first position, and the second signal being a signal of the target voice collected based on a second position;

determine a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal;

determine, based on the first signal and the second signal, a plurality of parameters related to a plurality of sound source directions, each parameter of the plurality of parameters corresponding to a probability that a sound is emitted from a sound source direction to form the first signal and the second signal;

determine, based on the plurality of parameters and the position of the target voice, a second coefficient, wherein to determine, based on the plurality of parameters and the position of the target voice, the second coefficient, the system is further caused to:

determine a direction of a synthetic sound source based on the plurality of parameters, wherein to determine the direction of the synthetic sound source based on the plurality of parameters, the system is further caused to:

determine that a sound source direction corresponding to a parameter with the greatest value among the plurality of parameters is the direction of the synthetic sound source;

and

determine, based on the direction of the synthetic sound source and the position of the target voice, the second coefficient;

and

obtain a voice-enhanced first output voice signal corresponding to the target voice by processing, based on the first coefficient and the second coefficient, the first signal and/or the second signal.

18 . The system of claim 17 , wherein to determine a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal, the at least one processor is directed to cause the system to:

obtain a signal pointing to the first direction and a signal pointing to the second direction by performing, based on the position of the target voice, the first position, and the second position, a differential operation on the first signal and the second signal, the signal pointing to the first direction and the signal pointing to the second direction containing different proportions of valid signals;

determine an estimated signal-to-noise ratio (SNR) of the target voice based on the signal pointing to the first direction and the signal pointing to the second direction, and

determine the first coefficient based on the estimated SNR.

19 . A non-transitory computer-readable storage medium storing computer instructions that, when executed by at least one processor, direct the at least one processor to perform a method comprising:

obtaining a first signal and a second signal of a target voice, the first signal being a signal of the target voice collected based on a first position, and the second signal being a signal of the target voice collected based on a second position;

determining a first coefficient by processing, based on a position of the target voice, the first position, and the second position, the first signal and the second signal;

determining, based on the first signal and the second signal, a plurality of parameters related to a plurality of sound source directions, each parameter of the plurality of parameters corresponding to a probability that a sound is emitted from a sound source direction to form the first signal and the second signal;

determining, based on the plurality of parameters and the position of the target voice, a second coefficient, comprising:

determining a direction of a synthetic sound source based on the plurality of parameters, including:

determining that a sound source direction corresponding to a parameter with the greatest value among the plurality of parameters is the direction of the synthetic sound source;

and

determining, based on the direction of the synthetic sound source and the position of the target voice, the second coefficient;

and

obtaining a voice-enhanced first output voice signal corresponding to the target voice by processing, based on the first coefficient and the second coefficient, the first signal and/or the second signal.

Assignments (2)

Continuity (2)

References Cited (27)