WIND AVOIDANCE AUDIO OPTIMIZATION FOR VOICE

1 . An image capture device, comprising:

a first microphone;

a second microphone; and

a processor configured to:

obtain a first microphone signal from the first microphone;

obtain a second microphone signal from the second microphone;

determine coherence values between the first microphone signal and the second microphone signal across a frequency band, wherein the frequency band comprises a voice sub-band and non-voice sub-bands, and wherein the voice sub-band and the non-voice sub-bands each comprise frequency bins and a coherence value is determined for each frequency bin;

determine that wind is present based on the determined coherence values for each frequency bin;

select non-voice sub-band frequency bins from the first microphone signal and the second microphone signal based on a lowest energy value of each respective non-voice sub-band frequency bin;

select voice sub-band frequency bins from a predetermined microphone signal; and

output a composite signal that comprises the selected non-voice sub-band frequency bins and the selected voice sub-band frequency bins.

2 . The image capture device of claim 1 , wherein the voice sub-band ranges from 300 Hz to 8000 Hz.

3 . The image capture device of claim 1 , wherein the predetermined microphone signal is the first microphone signal.

4 . The image capture device of claim 1 , wherein the predetermined microphone signal is the second microphone signal.

5 . The image capture device of claim 1 , wherein coherence values of a subset of the frequency bins are averaged to create a wind meter value that indicates a presence of wind.

6 . The image capture device of claim 1 , wherein the lowest energy value corresponds to a high coherence value.

7 . The image capture device of claim 1 , wherein each frequency bin is 93.75 Hz.

8 . An image capture device, comprising:

a first microphone;

a second microphone; and

a processor configured to:

obtain a first microphone signal from the first microphone;

obtain a second microphone signal from the second microphone;

determine coherence values between the first microphone signal and the second microphone signal across a frequency band, wherein the frequency band comprises a voice sub-band and non-voice sub-bands, and wherein the voice sub-band and the non-voice sub-bands each comprise frequency bins and a coherence value is determined for each frequency bin;

determine that wind is present based on the determined coherence values for each frequency bin;

select non-voice sub-band frequency bins from the first microphone signal and the second microphone signal based on a lowest energy value of each respective non-voice sub-band frequency bin;

select voice sub-band frequency bins from the first microphone signal based on an average energy per microphone in the voice sub-band; and

output a composite signal that comprises the selected non-voice sub-band frequency bins and the selected voice sub-band frequency bins.

9 . The image capture device of claim 8 , wherein the processor is further configured to:

select voice sub-band frequency bins from the second microphone signal based on the average energy per microphone in the voice sub-band; and

apply a smoothing algorithm to the voice sub-band.

10 . The image capture device of claim 8 , wherein the voice sub-band frequency bins of the first microphone signal are selected for a minimum duration.

11 . The image capture device of claim 10 , wherein the minimum duration is 5 milliseconds.

12 . The image capture device of claim 8 , wherein the non-voice sub-band range is below 300 Hz.

13 . The image capture device of claim 8 , wherein coherence values of a subset of the frequency bins are averaged to create a wind meter value that indicates an absence of wind.

14 . The image capture device of claim 8 , wherein the lowest energy value corresponds to a high coherence value.

15 . The image capture device of claim 8 , wherein each frequency bin is 93.75 Hz.

16 . A method, comprising:

obtaining a first microphone signal from a first microphone;

obtaining a second microphone signal from a second microphone;

determining coherence values between the first microphone signal and the second microphone signal across a frequency band, wherein the frequency band comprises a voice sub-band and non-voice sub-bands, and wherein the voice sub-band and the non-voice sub-bands each comprise frequency bins and a coherence value is determined for each frequency bin;

determining that wind is present based on the determined coherence values for each frequency bin;

selecting non-voice sub-band frequency bins from the first microphone signal and the second microphone signal based on a lowest energy value of each respective non-voice sub-band frequency bin;

selecting voice sub-band frequency bins from the first microphone signal based on a lowest coherence value; and

outputting a composite signal that comprises the selected non-voice sub-band frequency bins and the selected voice sub-band frequency bins.

17 . The method of claim 16 , further comprising:

selecting voice sub-band frequency bins from the second microphone signal based on the lowest coherence value; and

applying a smoothing algorithm to the voice sub-band.

18 . The method of claim 16 , wherein the voice sub-band frequency bins of the first microphone signal are selected for a minimum duration.

19 . The method of claim 18 , wherein the minimum duration is 5 milliseconds.

20 . The method of claim 16 , wherein the non-voice sub-band range is above 8000 Hz.