Headset activation validation based on audio data

1. A method comprising:

detecting an activation event on a device;

responsive to detecting the activation event:

starting supply of a power to a microphone array on the device;

capturing, via the microphone array on the device over a plurality of time periods, sound from a local area of the device;

storing each portion of the sound captured during a respective time period of the plurality of time period as respective audio data in a buffer;

performing a respective validation of a plurality of iterative validations of the activation event by analyzing the respective audio data stored in the buffer; and

performing an action based on a result of the plurality of iterative validations, the action comprises stopping supply of the power to the microphone array when a result of each validation of the plurality of iterative validations is an invalid activation state.

2. The method of claim 1 , wherein the activation event is a press of a soft-touch button on the device.

3. The method of claim 1 , wherein the action further comprises switching the device into a high-power state from a low-power state when the result of the plurality of iterative validations is a valid activation state.

4. The method of claim 3 , further comprising:

transmitting the respective audio data stored in the buffer to a second device when the result of the plurality of iterative validations is the valid activation state.

5. The method of claim 1 , further comprising:

responsive to detecting the activation event on the device, switching the device into a high-power state from a low-power state, wherein a first power is supplied to the microphone array in the low-power state, a second power is supplied to the microphone array in the high-power state, and the first power is lower than the second power.

6. The method of claim 1 , wherein performing the plurality of iterative validations of the activation event by analyzing the captured sound comprises:

determining a proximity of the device to a user of the device based on the captured sound; and

determining the result of the plurality of iterative validations based in part on the determined proximity.

7. The method of claim 6 , wherein determining the proximity of the device to the user comprises:

identifying a portion of the captured sound as speech generated by the user; and

analyzing an intensity of the speech to estimate the proximity of the device to the user.

8. The method of claim 1 , wherein performing the plurality of iterative validations of the activation event by analyzing the captured sound comprises:

determining a movement of the device relative to a user of the device based on the captured sound; and

determining the result of the plurality of iterative validations based in part on the determined movement.

9. The method of claim 8 , wherein the result of the plurality of iterative validations is determined to be the invalid activation state based on the determined movement of the device being away from the user of the device.

10. The method of claim 8 , wherein determining the movement of the device relative to the user comprises:

identifying a portion of the captured sound as speech generated by the user; and

analyzing an intensity of the speech over time to determine the movement of the device relative to the user.

11. The method of claim 8 , wherein determining the movement of the device relative to the user comprises:

determining an acoustic receptive field of the microphone array using the captured sound; and

analyzing the acoustic receptive field over time to determine the movement of the device relative to the user.

12. The method of claim 11 , wherein the microphone array includes a first microphone configured to monitor a first region in the acoustic receptive field, wherein the first region is a null region when the device is in a first position relative to the user, and analyzing the acoustic receptive field over time comprises determining whether the first region transitions between an active region and the null region over time.

13. A headset comprising:

a frame configured to be worn on a head of a user;

a soft-touch button on the frame and configured to detect an activation event;

a microphone array configured to capture sound; and

a controller configured to:

start supply of a power to the microphone array responsive to detection of the activation event,

instruct the microphone array to capture sound from a local area of the headset over a plurality of time periods responsive to detection of the activation event,

store each portion of the sound captured during a respective time period of the plurality of time period as respective audio data in a buffer,

perform a respective validation of a plurality of iterative validations of the activation event by analyzing the respective audio data stored in the buffer, and

perform an action based on a result of the plurality of iterative validations, the action comprises stopping supply of the power to the microphone array when the result of each validation of the plurality of iterative validations is an invalid activation state.

14. The headset of claim 13 , wherein the soft-touch button is a capacitive touch sensor configured to determine the activation event based on a proximity of skin of the user to the capacitive touch sensor.

15. The headset of claim 13 , wherein the controller is further configured to:

switch the headset into a high-power state from a low-power state when the result of the plurality of iterative validations is a valid activation state.

16. The headset of claim 13 , wherein the controller is further configured to:

switch the headset into a high-power state from a low-power state based on the activation event, wherein a first power is supplied to the microphone array in the low-power state, a second power is supplied to the microphone array in the high-power state, and the first power is lower than the second power.

17. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform operations comprising:

detecting an activation event on a device;

responsive to detecting the activation event:

starting supply of a power to a microphone array on the device;

instructing the microphone array to capture sound from a local area of the device over a plurality of time periods;

storing each portion of the sound captured during a respective time period of the plurality of time period as respective audio data in a buffer;

performing a respective validation of a plurality of iterative validations of the activation event by analyzing the respective audio data stored in the buffer; and

performing an action based on a result of the plurality of iterative validations, the action comprises stopping supply of the power to the microphone array when a result of each validation of the plurality of iterative validations is an invalid activation state.