Granted Patent B2

US 11,363,147 · App. 16/141,446 · Granted Jun 14, 2022

Inventor: Mark Adams (Lehi, UT)

Assignee: Sorenson IP Holdings, LLC

H04M9/087G10L21/0232H04M9/082H04M9/10G10L2021/02082H04M2201/40

Abstract

Claims (67)

1. A method comprising:

obtaining a transmit speech signal generated by a first device during a communication session between the first device and a second device;

obtaining a receive-path signal that includes an echo speech signal of the transmit speech signal and a receive speech signal originating at the second device during the communication session;

identifying a first portion of the receive-path signal that occurs at a first time frame and that includes a first frequency component that corresponds to the echo speech signal and a second frequency component that corresponds to the receive speech signal in which the first frequency component is different from the second frequency component;

attenuating the first frequency component of the first portion while avoiding attenuating the second frequency component of the first portion based on the first frequency component corresponding to the echo speech signal and the second frequency component corresponding to the receive speech signal;

identifying a second portion of the receive-path signal that occurs at a second time frame different from the first time frame and that includes a third frequency component of the echo speech signal and a fourth frequency component of the receive speech signal in which the third frequency component is the same as the fourth frequency component; and

avoiding attenuating the second portion of the receive-path signal based on the fourth frequency component being the same as the third frequency component.

2. The method of claim 1 , further comprising amplifying the second frequency component of the first portion while avoiding amplifying the first frequency component based on the second frequency component corresponding to the receive speech signal and based on the first frequency component corresponding to the echo speech signal.

3. The method of claim 1 , further comprising:

transmitting the transmit speech signal to the second device by way of a communication interface of the first device,

wherein

the obtaining the transmit speech signal includes generating by the first device the transmit speech signal, and

the obtaining the receive-path signal includes receiving, at the communication interface of the first device, the receive-path signal.

4. The method of claim 1 , further comprising avoiding amplifying the second portion of the receive-path signal based on the fourth frequency component being the same as the third frequency component.

5. The method of claim 1 , wherein attenuating the first frequency component is in response to determining that the echo speech signal at the first portion corresponds to voiced speech.

6. The method of claim 1 , further comprising:

identifying another portion of the receive-path signal that includes the receive speech signal but that does not include the echo speech signal; and

amplifying the other portion of the receive-path signal in response to the other portion including the receive speech signal and not including the echo speech signal.

7. The method of claim 6 , wherein amplifying the other portion of the receive-path signal includes amplifying a specific frequency component of the other portion using a filter that is based on the specific frequency component in response to determining that the receive speech signal at the other portion corresponds to voiced speech that has the specific frequency component.

8. The method of claim 6 , wherein amplifying the other portion of the receive-path signal includes:

determining that the receive speech signal at the other portion corresponds to unvoiced speech;

identifying a difference between an unvoiced speech frequency spectrum of the unvoiced speech and an estimated noise floor frequency spectrum of the receive-path signal; and

amplifying, in response to determining that the receive speech signal at the other portion corresponds to the unvoiced speech, the other portion based on the difference between the unvoiced speech frequency spectrum and the estimated noise floor frequency spectrum such that frequency components of the other portion that correspond to the unvoiced speech frequency spectrum are amplified more than frequency components of the other portion that correspond to the estimated noise floor frequency spectrum.

9. The method of claim 1 , further comprising:

identifying another portion of the receive-path signal that includes the echo speech signal but that does not include the receive speech signal;

determining that the echo speech signal at the other portion corresponds to voiced speech; and

attenuating the other portion of the receive-path signal in response to determining that the echo speech signal corresponds to the voiced speech.

10. At least one non-transitory computer-readable media configured to store one or more instructions that, in response to being executed by at least one computing system, cause performance of the method of claim 1 .

11. A method comprising:

generating, at a first device, a transmit speech signal during a communication session between the first device and a second device;

transmitting, via a voice network, the transmit speech signal to the second device by way of a communication interface of the first device;

receiving, at the communication interface of the first device, a receive-path signal that includes an echo speech signal of the transmit speech signal and a receive speech signal originating at the second device during the communication session and transmitted over the voice network, the echo speech signal being an echo of the transmit speech signal that results from transmission of the transmit speech signal by way of the communication interface via the voice network;

identifying a portion of the receive-path signal that includes, at a particular time, a first frequency component that corresponds to the echo speech signal and a second frequency component that corresponds to the receive speech signal in which the first frequency component is different from the second frequency component;

attenuating the first frequency component of the portion while avoiding attenuating the second frequency component of the portion based on the first frequency component corresponding to the echo speech signal and the second frequency component corresponding to the receive speech signal, the attenuating of the first frequency component being in response to determining that the echo speech signal corresponds to voiced speech; and

amplifying the second frequency component of the portion while avoiding amplifying the first frequency component based on the second frequency component corresponding to the receive speech signal and based on the first frequency component corresponding to the echo speech signal.

12. The method of claim 11 , further comprising:

identifying another portion of the receive-path signal that includes, at another particular time, a third frequency component of the echo speech signal and a fourth frequency component of the receive speech signal in which the third frequency component is the same as the fourth frequency component; and

avoiding attenuating the other portion of the receive-path signal based on the fourth frequency component being the same as the third frequency component.

13. The method of claim 11 , further comprising:

identifying another portion of the receive-path signal that includes the receive speech signal but that does not include the echo speech signal; and

amplifying the other portion of the receive-path signal in response to the other portion including the receive speech signal and not including the echo speech signal.

14. The method of claim 13 , wherein amplifying the other portion of the receive-path signal includes amplifying a specific frequency component of the other portion using a filter that is based on the specific frequency component in response to determining that the receive speech signal at the other portion corresponds to voiced speech that has the specific frequency component.

15. The method of claim 13 , wherein amplifying the other portion of the receive-path signal includes:

determining that the receive speech signal at the other portion corresponds to unvoiced speech;

identifying a difference between an unvoiced speech frequency spectrum of the unvoiced speech and an estimated noise floor frequency spectrum of the receive-path signal; and

amplifying, in response to determining that the receive speech signal at the other portion corresponds to the unvoiced speech, the other portion based on the difference between the unvoiced speech frequency spectrum and the estimated noise floor frequency spectrum such that frequency components of the other portion that correspond to the unvoiced speech frequency spectrum are amplified more than frequency components of the other portion that correspond to the estimated noise floor frequency spectrum.

16. The method of claim 11 , further comprising:

identifying another portion of the receive-path signal that includes the echo speech signal but that does not include the receive speech signal;

determining that the echo speech signal at the other portion corresponds to voiced speech; and

attenuating the other portion of the receive-path signal in response to determining that the echo speech signal corresponds to the voiced speech.

17. A system comprising:

one or more processors; and

one or more computer-readable media configured to store instructions that in response to being executed by the one or more processors cause the system to perform operations, the operations comprising:

obtaining a transmit speech signal generated by a first device during a communication session between the first device and a second device;

obtaining a receive-path signal that includes an echo speech signal of the transmit speech signal and a receive speech signal originating at the second device during the communication session;

identifying a first portion of the receive-path signal that occurs at a first time frame and that includes a first frequency component that corresponds to the echo speech signal and a second frequency component that corresponds to the receive speech signal in which the first frequency component is different from the second frequency component;

attenuating the first frequency component of the first portion while avoiding attenuating the second frequency component of the first portion based on the first frequency component corresponding to the echo speech signal and the second frequency component corresponding to the receive speech signal;

identifying a second portion of the receive-path signal that occurs at a second time frame different from the first time frame and that includes a third frequency component of the echo speech signal and a fourth frequency component of the receive speech signal in which the third frequency component is the same as the fourth frequency component; and

avoiding attenuating the second portion of the receive-path signal based on the fourth frequency component being the same as the third frequency component.

18. The system of claim 17 , wherein attenuating the first frequency component is in response to determining that the echo speech signal at the first portion corresponds to voiced speech.

19. The system of claim 17 , wherein the operations further comprise:

identifying another portion of the receive-path signal that includes the receive speech signal but that does not include the echo speech signal; and

amplifying the other portion of the receive-path signal in response to the other portion including the receive speech signal and not including the echo speech signal.

20. The system of claim 17 , wherein the operations further comprise:

identifying another portion of the receive-path signal that includes the echo speech signal but that does not include the receive speech signal;

determining that the echo speech signal at the other portion corresponds to voiced speech; and

attenuating the other portion of the receive-path signal in response to determining that the echo speech signal corresponds to the voiced speech.

Assignments (11)

Continuity (1)