Apparatus and method for power efficient signal conditioning for a voice recognition system
A disclosed method includes monitoring an audio signal energy level while having a noise suppressor deactivated to conserve battery power, buffering the audio signal in response to a detected increase in the audio energy level, activating and running a voice activity detector on the audio signal in response to the detected increase in the audio energy level and activating and running a noise estimator in response to voice being detected in the audio signal by the voice activity detector. The method may further include activating and running the noise suppressor only if the noise estimator determines that noise suppression is required. The method activates and runs a noise type classifier to determine the noise type based on information received from the noise estimator and selects a noise suppressor algorithm, from a group of available noise suppressor algorithms, where the selected noise suppressor algorithm is the most power consumption efficient.
1. A computer-implemented method when executed on data processing hardware of a computing device causes the data processing hardware to perform operations comprising:
receiving an audio signal detected by a first microphone in a group of microphones of the computing device while a second microphone in the group of microphones is powered off;
while the second microphone is powered off, determining an audio signal energy level of the audio signal detected by the first microphone has deviated from a baseline audio signal energy level by more than a threshold amount;
in response to determining that the audio signal energy level of the audio signal detected by the first microphone has deviated from the baseline audio signal energy level by more than the threshold amount:
triggering the second microphone to power on; and
triggering a voice activity detector to power on;
performing, using the voice activity detector that is powered on, voice activity detection on the audio signal to determine whether speech is detected in the audio signal detected by the first microphone;
buffering a voice signal based on audio signals detected by the first microphone and the second microphone; and
in response to determining that speech is detected in the audio signal detected by the first microphone:
estimating a signal-to-noise ratio (SNR) of the buffered voice signal; and
performing, based on the estimated SNR of the buffered voice signal, noise suppression on the buffered voice signal to provide a noise suppressed voice signal.
2. The computer-implemented method of claim 1 , wherein the operations further comprise, in response to determining that speech is not detected in the audio signal detected by the first microphone, powering the second microphone off and maintaining the first microphone powered on.
3. The computer-implemented method of claim 1 , wherein the operations further comprise performing speech recognition on the noise suppressed voice signal.
4. The computer-implemented method of claim 1 , wherein the operations further comprise:
receiving a subsequent audio signal;
determining that an audio signal energy level of the subsequent audio signal has deviated from the baseline audio signal energy level by more than the threshold amount;
performing, using the voice activity detector, voice activity detection on the subsequent audio signal to determine that speech is detected in the subsequent audio signal;
buffering a subsequent voice signal based on the subsequent audio signal; and
in response to determining that speech is detected in the subsequent audio signal:
estimating a signal-to-noise ratio (SNR) of the buffered subsequent voice signal; and
determining not to apply noise suppression on the buffered subsequent voice signal based on the estimated SNR of the buffered subsequent voice signal.
5. The computer-implemented method of claim 4 , wherein the operations further comprise, after determining not to apply noise suppression on the buffered subsequent voice signal, performing speech recognition on the buffered subsequent voice signal.
6. The computer-implemented method of claim 1 , wherein:
the group of microphones comprises the first microphone, the second microphone, and at least one other microphone; and
triggering the second microphone to power on comprises triggering the second microphone and the at least one other microphone in the group of microphones to power on.
7. The computer-implemented method of claim 1 , wherein the computing device comprises a battery-powered, voice-enabled electronic device.
8. The computer-implemented method of claim 1 , wherein the computing device comprises a voice-enabled smart phone.
9. A computing device comprising:
data processing hardware; and
memory hardware in communication with the data processing hardware and storing instructions that when executed on the data processing hardware cause the data processing hardware to perform operations comprising:
receiving an audio signal detected by a first microphone in a group of microphones of the computing device while a second microphone in the group of microphones is powered off;
while the second microphone is powered off, determining an audio signal energy level of the audio signal detected by the first microphone has deviated from a baseline audio signal energy level by more than a threshold amount;
in response to determining that the audio signal energy level of the audio signal detected by the first microphone has deviated from the baseline audio signal energy level by more than the threshold amount:
triggering the second microphone to power on; and
triggering a voice activity detector to power on;
performing, using the voice activity detector that is powered on, voice activity detection on the audio signal to determine whether speech is detected in the audio signal detected by the first microphone;
buffering a voice signal based on audio signals detected by the first microphone and the second microphone; and
in response to determining that speech is detected in the audio signal detected by the first microphone:
estimating a signal-to-noise ratio (SNR) of the buffered voice signal; and
performing, based on the estimated SNR of the buffered voice signal, noise suppression on the buffered voice signal to provide a noise suppressed voice signal.
10. The computing device of claim 9 , wherein the operations further comprise, in response to determining that speech is not detected in the audio signal detected by the first microphone, powering the second microphone off and maintaining the first microphone powered on.
11. The computing device of claim 9 , wherein the operations further comprise performing speech recognition on the noise suppressed voice signal.
12. The computing device of claim 9 , wherein the operations further comprise:
receiving a subsequent audio signal;
determining that an audio signal energy level of the subsequent audio signal has deviated from the baseline audio signal energy level by more than the threshold amount;
performing, using the voice activity detector, voice activity detection on the subsequent audio signal to determine that speech is detected in the subsequent audio signal;
buffering a subsequent voice signal based on the subsequent audio signal; and
in response to determining that speech is detected in the subsequent audio signal:
estimating a signal-to-noise ratio (SNR) of the buffered subsequent voice signal; and
determining not to apply noise suppression on the buffered subsequent voice signal based on the estimated SNR of the buffered subsequent voice signal.
13. The computing device of claim 12 , wherein the operations further comprise, after determining not to apply noise suppression on the buffered subsequent voice signal, performing speech recognition on the buffered subsequent voice signal.
14. The computing device of claim 9 , wherein:
the group of microphones comprises the first microphone, the second microphone, and at least one other microphone; and
triggering the second microphone to power on comprises triggering the second microphone and the at least one other microphone in the group of microphones to power on.
15. The computing device of claim 9 , wherein the computing device comprises a battery-powered, voice-enabled electronic device.
16. The computing device of claim 9 , wherein the computing device comprises a voice-enabled smart phone.