Granted Patent B2

US 10,230,346 · App. 13/346,719 · Granted Mar 12, 2019

Abstract

Claims (43)

1. A method, comprising:

receiving a plurality of signals associated with a plurality of subbands, each of the plurality of signals being associated with one of the plurality of subbands;

detecting, at an acoustic voice activity detector, a signal from an omnidirectional microphone;

performing an operation using the plurality of signals to determine a plurality of values, each of the plurality of values being a binary value associated with one of the plurality of signals;

comparing each of the plurality of values using autocorrelation and cross-correlation to determine whether speech is detected in the signal, and, if speech is detected in the signal and scratch noise is not detected in the signal, generating a first signal;

receiving, at a vibration voice activity detector, the signal and, a second signal generated by a skin surface microphone, and generating, by a contact detector that receives the first signal and the second signal, a third signal indicating that speech is detected if the second signal exceeds a threshold, the contact detector configured to determine a state of contact with a skin surface; and

generating a final voice activity detected signal indicating that speech is detected if either the first signal is generated, or the second signal is generated, or both.

2. The method of claim 1 , wherein each of the plurality of signals is associated with at least one of the plurality of subbands, the plurality of subbands being configured for a format frequency that corresponds to the speech.

3. The method of claim 1 , wherein the omnidirectional microphone comprises a dual omnidirectional microphone array.

4. The method of claim 1 , wherein each of the plurality of signals comprises a subbanded signal, the subbanded signal being used to evaluate a frequency range of substantially 0 to 1500 Hz.

5. The method of claim 1 , wherein the operation comprises generating a signal-to-noise ratio.

6. The method of claim 1 , wherein performing the operation further comprises:

determining a ratio;

referencing a threshold; and

comparing the ratio to the threshold to generate a result, the result being configured to indicate whether the speech is detected on one of the plurality of subbands.

7. The method of claim 1 , wherein at least two of the plurality of signals indicates detection of the speech.

8. A method, comprising:

receiving a plurality of signals, each of the plurality of signals being associated with one of a plurality of subbands;

detecting, at an acoustic voice activity detector, a signal from a microphone associated with an omnidirectional microphone array;

performing an operation using the plurality of signals to determine a binary value associated with each of the plurality of subbands;

performing another operation using the signal from the microphone associated with the omnidirectional microphone array after performing the operation;

evaluating, using autocorrelation and a cross-correlation, a result associated with the operation to determine if speech is detected, and, if speech is detected in the signal and scratch noise is not detected in the signal, generating a first signal;

comparing another result of the another operation to generate a ratio, the ratio being compared to a threshold to determine whether the signal is devoiced;

receiving, at a vibration voice activity detector, the first signal and a second signal generated by a skin surface microphone, and generating, by a contact detector that receives the first signal and the second signal, a third signal indicating that speech is detected if the second signal exceeds a threshold, the contact detector configured to determine a state of contact with a skin surface; and

generating a final voice activity detected signal indicating that speech is detected if either the first signal is generated, or the second signal is generated, or both.

9. The method of claim 8 , further comprising evaluating the binary value associated with each of the plurality of values to determine a final acoustic voice activity detector.

10. The method of claim 8 , wherein the microphone is an omnidirectional microphone.

11. The method of claim 8 , wherein the omnidirectional microphone array comprises a dual omnidirectional microphone array.

12. The method of claim 8 , wherein the microphone comprises a skin surface microphone.

13. A method, comprising:

detecting, at an acoustic voice activity detector, a signal associated with a subband from a microphone;

performing an operation on data associated with the signal, the operation generating a value associated with the subband;

determining whether the value distinguishes the signal from noise by using the value to determine a signal-to-noise ratio and comparing the value to a threshold;

comparing, using autocorrelation and cross-correlation, the value to a threshold to determine whether speech is detected, and, if speech is detected in the signal and scratch noise is not detected in the signal, generating a first signal;

receiving, at a vibration voice activity detector, the signal and a second signal generated by a skin surface microphone, and generating, by a contact detector that receives the first signal and the second signal, a third signal indicating that speech is detected if the second signal exceeds a threshold, the contact detector configured to determine a state of contact with a skin surface; and

generating a final voice activity detected signal indicating that speech is detected if either the first signal is generated, or the second signal is generated, or both.

14. The method of claim 13 , wherein the microphone comprises one or more omnidirectional microphones.

15. The method of claim 13 , further comprising determining a final acoustic voice activity detection value by comparing a plurality of values generated from a plurality of signals detected by the microphone.

16. The method of claim 15 , wherein speech is detected if two or more of the plurality of values are equal to or exceed a threshold configured to distinguish the signal from the noise.

17. The method of claim 13 , wherein speech is detected in the subband if the value is equal to or greater than a threshold configured to identify the signal.

18. The method of claim 13 , wherein the subband is configured to correspond to a format frequency used to distinguish the signal from noise.

19. The method of claim 13 , wherein the microphone comprises a dual omnidirectional microphone array.

20. The method of claim 13 , wherein the microphone comprises an omnidirectional microphone array.

Assignments (17)