Identifying sound from a source of interest based on multiple audio feeds

• US 9,691,413 B2
• Filed: 10/06/2015
• Issued: 06/27/2017
• Est. Priority Date: 10/06/2015
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Zad Issa, Syavosh
• Primary Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US14/876,666
• Publication Number:
  US 20170098457A1
• Time in Patent Office: 630 Days
• Field of Search: 704208, 704233, 704205, 704225, 704226
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2021/02165 Two microphones, one receiving mainly the noise signal and the other one mainly the speech signal
• CPC Class Codes: G10L 2025/783 based on threshold decision
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/034 Automatic adjustment
• CPC Class Codes: G10L 21/0388 Details of processing therefor
• CPC Class Codes: G10L 21/055 for synchronising with other signals, e.g. video signals
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: H04R 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407

Abstract

Methods and systems for identifying sound from a source of interest are provided for herein. In some embodiments, a first audio feed is captured by a first microphone and a second audio feed is captured by a second microphone. The first microphone may be located closer in proximity to the source of interest than the second microphone. The first audio feed can be processed utilizing the second audio feed to produce a first processed audio feed that can enable identification of sound originating from the source of interest. In some embodiments, the second audio feed can be additionally processed utilizing the first audio feed to produce a second processed audio feed. In such embodiments, frequencies from the first processed audio feed can be compared against frequencies of the second processed audio feed to identify sound originating from the source of interest. Other embodiments may be described and/or claimed herein.

Claims

1. A sound processing system comprising:

a first audio capture device and a second audio capture device, wherein the first audio capture device is located in closer proximity to a point of interest than the second audio capture device;

a voice activity detection module to;

receive first and second audio feeds respectively captured by the first and second audio capture devices;

attenuate at least a portion of the first audio feed based on a corresponding portion of the second audio feed to generate a first attenuated audio feed;

attenuate at least a portion of the second audio feed based on a corresponding portion of the first audio feed to generate a second attenuated audio feed;

compare frequency bands of the first attenuated audio feed with corresponding frequency bands of the second attenuated audio feed; and

determine a source confidence level based on a number of the frequency bands from the first attenuated audio feed that exceed a predefined threshold of difference from the corresponding frequency bands of the second attenuated audio feed, wherein the source confidence level is indicative of whether sound is originating from the point of interest.

2. The sound processing system of claim 1, wherein a higher value for the source confidence level is more indicative of sound within the first attenuated audio feed originating from the point of interest than a lower value for the source confidence level.

3. The sound processing system of claim 1, wherein to attenuate at least the portion of the first audio feed based on the corresponding portion of the second audio feed is to attenuate one or more frequencies contained within the first audio feed that are contained within the second audio feed, and wherein to attenuate at least the portion of the second audio feed based on the corresponding portion of the first audio feed is to attenuate one or more frequencies contained within the second audio feed that are contained within the first audio feed.

4. The sound processing system of claim 1, wherein the voice activity detection module is further to:

time synchronize the first audio feed with the second audio feed prior to attenuating at least the portion of the first audio feed; and

time synchronize the second audio feed with the first audio feed prior to attenuating at least the portion of the second audio feed.

5. The sound processing system of claim 1, wherein to time synchronize the first audio feed with the second audio feed is to apply a first delay to the first audio feed, the first delay reflecting the amount of time it takes for sound to travel from the first audio capture device to the second audio capture device, and wherein to time synchronize the second audio feed with the first audio feed is to apply a second delay to the second audio feed, the second delay reflecting the amount of time it takes for sound to travel from the second audio capture device to the first audio capture device.

6. The sound processing system of claim 1, further comprising:

a voice recognition module to;

receive the first attenuated audio feed;

monitor the first attenuated audio feed to identify one or more triggers contained within the first attenuated audio feed; and

cause one or more actions to occur in response to identifying the one or more triggers.

7. The sound processing system of claim 6, wherein the voice activity detection module is further to:

output the first attenuated audio feed to the voice recognition engine in response to a determination that the source confidence level exceeds a preconfigured limit.

8. The sound processing system of claim 7, wherein the preconfigured limit varies based upon a power level of a computing device that hosts the sound processing system.

9. The sound processing system of claim 1, wherein the voice activity detection module is further to:

determine a noise confidence level based on a number of the frequency bands from the first audio feed that are within a predefined threshold of difference from the corresponding frequency bands of the second audio feed, wherein a higher value for the noise confidence level is more indicative of sound within the first audio feed being noise than a lower value for the noise confidence level.

10. The sound processing system of claim 1, further comprising an acoustic echo cancellation (AEC) module that is to:

reduce an amount of echo contained within the first attenuated audio feed.

11. One or more computer storage hardware media device having computer-executable instructions embodied thereon that, when executed, by one or more processors of a computing device, causes the one or more processors to:

perform a method for processing sound, the method comprising;

filtering a first audio feed utilizing a second audio feed to produce a filtered audio feed, wherein the first audio feed is captured by a first microphone and the second audio feed is captured by a second microphone, the first microphone being closer in proximity to an audio source of interest than the second microphone; and

identifying whether the first audio feed contains sound originating from a direction of the source of interest based on frequencies contained within the filtered audio feed.

12. The one or more computer storage media of claim 11, wherein the filtered audio feed is a first filtered audio feed the method further comprising:

filtering the second audio feed utilizing the first audio feed to produce a second filtered audio feed, wherein identifying whether the first audio feed contains sound originating from the direction of the source of interest includes comparing frequency bands of the first filtered audio feed with corresponding frequency bands of the second filtered audio feed; and

determining a source confidence level based on a number of the frequency bands from the first filtered audio feed that exceed a predefined threshold of difference from the corresponding frequency bands of the second filtered audio feed.

13. The one or more computer storage media of claim 12, the method further comprising sending the filtered audio feed to a voice recognition engine of the computing device in response to the source confidence level exceeding a preconfigured limit.

14. The one or more computer storage media of claim 13, wherein the preconfigured limit varies based upon a power level of the computing device.

15. The one or more computer storage media of claim 12,

wherein filtering the first audio feed utilizing the second audio feed further comprises filtering frequencies from the first audio feed that are contained within the second audio feed, and

wherein filtering the second audio feed utilizing the first audio feed further comprises filtering frequencies from the second audio feed that are contained within the first audio feed.

16. A computer-implemented method for voice activity detection comprising:

receiving a first audio feed captured by a first microphone of a computing device and a second audio feed captured by a second microphone of the computing device, wherein the first microphone is closer in proximity to a source of interest than the second microphone; and

processing the first audio feed utilizing the second audio feed to enable identification of sound originating from a direction of the source of interest.

17. The computer-implemented method of claim 16, wherein processing the first audio feed utilizing the second audio feed comprises:

filtering frequencies of the first audio feed based on corresponding frequencies of the second audio feed to produce a filtered audio feed.

18. The computer-implemented method of claim 16, wherein processing the first audio feed utilizing the second audio feed comprises:

attenuating frequencies of the first audio feed based on corresponding frequencies of the second audio feed to produce an attenuated audio feed.

19. The computer-implemented method of claim 16, wherein processing the first audio feed utilizing the second audio feed comprises:

filtering frequencies of the first audio feed based on corresponding frequencies of the second audio feed to produce a first filtered audio feed;

filtering frequencies of the second audio feed based on corresponding frequencies of the first audio feed to produce a second filtered audio feed;

comparing frequency bands of the first filtered audio feed with corresponding frequency bands of the second filtered audio feed; and

determining a source confidence level based on a number of the frequency bands from the first filtered audio feed that exceed a predefined threshold of difference from the corresponding frequency bands of the second filtered audio feed, wherein a higher value for the source confidence level is more indicative of sound within the first audio feed originating from the direction of the source of interest than a lower value for the source confidence level.

20. The computer-implemented method of claim 19, wherein the source of interest is a user of the computing device, the method further comprising:

sending the first filtered audio feed to a voice recognition engine of the computing device in response to a determination that the value for the source confidence level exceeds a preconfigured limit, wherein the preconfigured limit is based upon a current power level of the computing device, and wherein a higher preconfigured limit reduces the amount of the first audio feed that is output to the voice recognition engine.