Noise cancellation

Patent metadata

Field	Value
Publication number	US-11120795-B2
Application number	US-201916445226-A
Country	US
Kind code	B2
Filing date	Jun 19, 2019
Priority date	Aug 24, 2018
Publication date	Sep 14, 2021
Grant date	Sep 14, 2021

Abstract

A method for noise cancellation, the method may include detecting a voice trigger or receiving an indication regarding an occurrence of a voice trigger; searching, based at least on a timing of occurrence of the voice trigger, for at least one of a noise-period and a voice-period; when finding a noise period then updating one or more noise parameters based on one or more features of sensed audio signals received during the noise period; when finding a voice period then updating one or more voice parameters based on one or more features of sensed audio signals received during the voice period; estimating, based on the one or more noise parameters and the one or more voice parameters, a probability that voice is present at one or more subsequent periods; and cancelling noise at the one or more subsequent periods, based on the voice presence probability.

First claim

What is claimed is:

1. A method for noise cancellation, the method comprises:
detecting a voice trigger or receiving an indication regarding an occurrence of a voice trigger;
searching, based at least on a timing of occurrence of the voice trigger, for at least one of a noise period and a voice period;
when finding a noise period then updating one or more noise parameters based on one or more features of sensed audio signals received during the noise period;
when finding a voice period then updating one or more voice parameters based on one or more features of sensed audio signals received during the voice period;
estimating, based on the one or more noise parameters and the one or more voice parameters, a probability that voice is present at one or more subsequent periods;
cancelling noise at the one or more subsequent periods, based on the voice presence probability; and
forgetting the one or more noise parameters and the one or more voice parameters following a last interaction of a user with the device.

2. The method according to claim 1 , wherein the detecting of the voice trigger is executed by a user device that is voice activated, and wherein the voice parameters are parameters of a voice of a user of the user device.

3. The method according to claim 1 , comprising buffering features of sensed audio signals received during a buffering period to provide buffered features instead of storing the sensed audio signals received during the buffering period; wherein the buffering period comprises at least one out of the noise period and the voice period.

4. The method according to claim 1 , comprising finding a noise period that immediately preceded a beginning of the voice trigger.

5. The method according to claim 1 , comprising finding a voice period that immediately precedes an end of the voice trigger.

6. The method according to claim 1 , comprising determining that a period that immediately preceded a beginning of the voice trigger is not a noise period when a time difference between the voice trigger and a last voice trigger that preceded the voice trigger is below a timing threshold.

7. The method according to claim 1 , comprising searching for a residual echo period; and when finding a residual echo period then updating one or more residual echo parameters based on one or more features of sensed audio signals received during the residual echo period.

8. The method according to claim 7 , wherein the estimating of the probability that voice is present at the one or more subsequent periods is also based on the one or more residual echo parameters.

9. A non-transitory computer readable medium that stores instructions for:
detecting a voice trigger or receiving an indication regarding an occurrence of a voice trigger;
searching, based at least on a timing of occurrence of the voice trigger, for at least one of a noise-period and a voice-period;
when finding a noise period then updating one or more noise parameters based on one or more features of sensed audio signals received during the noise period;
when finding a voice period then updating one or more voice parameters based on one or more features of sensed audio signals received during the voice period;
estimating, based on the one or more noise parameters and the one or more voice parameters, a probability that voice is present at one or more subsequent periods; and
cancelling noise at the one or more subsequent periods, based on the voice presence probability; and
forgetting the one or more noise parameters and the one or more voice parameters following a last interaction of a user with the device.

10. The non-transitory computer readable medium according to claim 9 , wherein the detecting of the voice trigger is executed by a user device that is voice activated, and wherein the voice parameters are parameters of a voice of a user of the user device.

11. The non-transitory computer readable medium according to claim 9 , that stores instructions for buffering features of sensed audio signals received during a buffering period to provide buffered features instead of storing the sensed audio signals received during the buffering period; wherein the buffering period comprises at least one out of the noise period and the voice period.

12. The non-transitory computer readable medium according to claim 9 , that stores instructions for finding a noise period that immediately preceded a beginning of the voice trigger.

13. The non-transitory computer readable medium according to claim 9 , that stores instructions for finding a voice period that immediately precedes an end of the voice trigger.

14. The non-transitory computer readable medium according to claim 9 , that stores instructions for determining that a period that immediately preceded a beginning of the voice trigger is not a noise period when a time difference between the voice trigger and a last voice trigger that preceded the voice trigger is below a timing threshold.

15. The non-transitory computer readable medium according to claim 9 , that stores instructions for searching for a residual echo period; and when finding a residual echo period then updating one or more residual echo parameters based on one or more features of sensed audio signals received during the residual echo period.

16. The non-transitory computer readable medium according to claim 15 , wherein the estimating of the probability that voice is present at the one or more subsequent periods is also based on the one or more residual echo parameters.

17. A device comprising a processing circuit that is configured to:
detect a voice trigger or receive an indication regarding an occurrence of a voice trigger;
search, based at least on a timing of occurrence of the voice trigger, for at least one of a noise period and a voice period;
when finding a noise period then update one or more noise parameters based on one or more features of sensed audio signals received during the noise period;
when finding a voice period then update one or more voice parameters based on one or more features of sensed audio signals received during the voice period;
estimate, based on the one or more noise parameters and the one or more voice parameters, a probability that voice is present at one or more subsequent periods;
cancel noise at the one or more subsequent periods, based on the voice presence probability; and
forget the one or more noise parameters and the one or more voice parameters following a last interaction of a user with the device.

18. The device according to claim 17 , wherein the voice trigger is a wake up word spoken by a user.

19. The method according to claim 1 , wherein the voice trigger is a wake up word spoken by a user.

Assignees

• Dsp Group Ltd

Inventors

• Erell Adoram

Classifications

• G10L15/22 Primary

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title

• H04R2410/01

  Noise reduction using microphones having different directional characteristics · CPC title

• G10L21/0216 Primary

  characterised by the method used for estimating noise · CPC title

• G10L2021/02082

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

• H04R3/005

  for combining the signals of two or more microphones (specially adapted for hearing aids H04R25/407) · CPC title