Audio device with wakeup word detection
An audio device with at least one microphone adapted to receive sound from a sound field and create an output, and a processing system that is responsive to the output of the microphone. The processing system is configured to use a signal processing algorithm to detect a wakeup word, and modify the signal processing algorithm that is used to detect the wakeup word if the sound field changes.
1. An audio device, comprising:
a plurality of microphones that are configurable into a microphone array adapted to receive sound from a sound field and create a plurality of microphone outputs, wherein the sound field comprises human voices and noise and the sound field is characterized by a sound pressure level, and
a processing system that is responsive to the plurality of outputs of the microphones comprising:
a signal processing algorithm to detect a wake up word, the signal processing algorithm comprising a beamformer that uses multiple microphone outputs,
wherein the signal processing algorithm is modified if the sound field changes,
wherein modifying the signal processing algorithm comprises changing the quantity of microphone outputs used by the beamformer, and;
wherein a change in the sound field comprises a change in the sound pressure level or a change in the number of human voices that comprise the sound field.
2. The audio device of claim 1 , wherein the processing system is configured to use more microphone outputs in the beamformer as the sound pressure level of the sound field increases.
3. The audio device of claim 1 , further comprising determining a rate of false wake up word detections.
4. The audio device of claim 3 , wherein the processing system is configured to use more microphones in the beamformer if a rate of false wake up word detections is above a threshold rate.
5. The audio device of claim 3 , wherein the processing system is configured to use more microphones in the beamformer if a rate of false wake up word detections is above a threshold rate and the sound pressure level of the sound field is above a threshold level.
6. The audio device of claim 5 , wherein the processing system is configured to use fewer microphones in the beamformer if the sound pressure level of the sound field decreases below the threshold level and the rate of false wake up word detections is below its threshold rate.
7. The audio device of claim 5 , wherein the processing system is configured to use fewer microphones in the beamformer if the sound pressure level of the sound field decreases below the threshold level, the rate of false wake up word detections is below its threshold rate, and at least one hysteresis rule is met.
8. The audio device of claim 7 , wherein the at least one hysteresis rule comprises at least one of a sound field sound pressure level and a timer.
9. The audio device of claim 1 , wherein modifying the signal processing algorithm that is used to detect the wakeup word if the sound field changes comprises modifying the complexity of the signal processing algorithm as the sound field changes.
10. The audio device of claim 9 , wherein the complexity of the signal processing algorithm is increased if the sound pressure level of the sound field increases.
11. The audio device of claim 9 , wherein modifying the complexity of the signal processing algorithm as the sound field changes comprises applying noise mitigation techniques.
12. The audio device of claim 9 , wherein modifying the complexity of the signal processing algorithm as the sound field changes comprises applying blind source separation.
13. The audio device of claim 1 , wherein the audio device comprises headphones.
14. The audio device of claim 13 , wherein the headphones have earcups with an outside and an inside, and at least some of the microphones are adapted to receive sound outside of an earcup.
15. The audio device of claim 1 , wherein the wakeup word detector is implemented with a low-power digital signal processor.
16. The audio device of claim 1 , wherein the processing system is further configured to monitor a microphone output to detect an indication of echo.
17. The audio device of claim 16 , wherein modifying the signal processing algorithm that is used to detect the wakeup word if the sound field changes comprises enabling an echo cancelling algorithm after an indication of echo has been detected.
18. The audio device of claim 17 , comprising a plurality of microphones that are configurable into a microphone array, and wherein:
the processing system is configured to use more microphones in the beamformer when the echo conditions are high.
19. The audio device of claim 1 , wherein the processing system is further configured to enable an automated echo cancelling algorithm when audio is playing through the audio device.