Multi-step validation of wakeup phrase processing
Various embodiments describe a multi-step process for determining whether the keyword has been spoken. The first tier uses an onboard digital signal processor (DSP) to process converted audio signals. If the DSP analysis cannot confirm the keyword due to noise in the signal, a second tier of processing may use an applications processor (AP) and a more stringent audio filter to try to verify the keyword. If the AP analysis cannot confirm the keyword, a third tier of processing may use server side processor and an even more stringent audio filter to try to verify the keyword. Certain signal to noise ratios (SNRs) of the detected and converted audio signal are used to decide when and if to try the second and third tier of processing.
1. A method of verifying a spoken keyword for a voice enabled communication device, comprising:
detecting, via at least one microphone sensor, an analog audio signal indicative of a spoken word or phrase;
converting the detected analog audio signal to a digital signal;
inputting the digital signal to a digital signal processor (DSP) within the communication device for processing to determine a signal to noise ratio (SNR) of the digital signal and whether the digital signal matches a known keyword;
when the DSP processing determines the digital signal matches the known keyword, causing the communication device to wake up and present an acknowledgement;
when the DSP processing could not positively determine that the digital signal matches the keyword, determining whether the SNR of the digital signal exceeds a first SNR threshold (T1) and, if so, causing the communication device to remain in a sleep state, and, if not, inputting the digital signal to an applications processor (AP) on the communications device for more stringent processing to determine if the digital signal matches the known keyword;
when the AP processing determines the digital signal matches the known keyword, causing the communication device to wake up and present an acknowledgement;
when the AP processing could not positively determine that the digital signal matches the keyword, determining whether the SNR of the digital signal exceeds a second SNR threshold (T2), where T2 is less than T1, and, if so, causing the communication device to remain in the sleep state, and, if not, sending the digital signal to a server processor for even more stringent analysis to determine if the digital signal matches the known keyword; and
receiving a pass or fail result from the server processor wherein if the result is fail, causing the communication device to remain in the sleep state and if the result is pass, causing the communication device to wake up and present an acknowledgement.
2. The method of claim 1 , wherein the acknowledgement presented by the communication device is an audible tone.
3. The method of claim 1 , wherein the acknowledgement presented by the communication device is a visual light indicator.
4. The method of claim 1 , wherein the T1 threshold is 6 dB.
5. The method of claim 4 , wherein the T2 threshold is 3 dB.
6. A communication device, comprising:
a memory; and
logic, at least a portion of which is implemented in circuitry coupled to the memory, the logic to:
detect, via at least one microphone sensor, an analog audio signal indicative of a spoken word or phrase;
convert the detected analog audio signal to a digital signal;
input the digital signal to a digital signal processor (DSP) within the communication device for processing to determine a signal to noise ratio (SNR) of the digital signal and whether the digital signal matches a known keyword;
when the DSP processing determines the digital signal matches the known keyword, cause the communication device to wake up and present an acknowledgement;
when the DSP processing could not positively determine that the digital signal matches the keyword, determine whether the SNR of the digital signal exceeds a first SNR threshold (T1) and, if so, cause the communication device to remain in a sleep state, and, if not, input the digital signal to an applications processor (AP) on the communications device for more stringent processing to determine if the digital signal matches the known keyword;
when the AP processing determines the digital signal matches the known keyword, cause the communication device to wake up and present an acknowledgement;
when the AP processing could not positively determine that the digital signal matches the keyword, determine whether the SNR of the digital signal exceeds a second SNR threshold (T2), where T2 is less than T1, and, if so, cause the communication device to remain in the sleep state, and, if not, send the digital signal to a server processor for even more stringent analysis to determine if the digital signal matches the known keyword; and
receive a pass or fail result from the server processor wherein if the result is fail, cause the communication device to remain in the sleep state and if the result is pass, cause the communication device to wake up and present an acknowledgement.
7. The communication device of claim 6 , wherein the acknowledgement presented by the communication device is an audible tone.
8. The communication device of claim 6 , wherein the acknowledgement presented by the communication device is a visual light indicator.
9. The communication device of claim 6 , wherein the T1 threshold is 6 dB.
10. The communication device of claim 9 , wherein the T2 threshold is 3 dB.
11. At least one non-transitory computer-readable medium comprising a set of instructions that, in response to being executed by a processor circuit, cause the processor circuit to:
detect, via at least one microphone sensor, an analog audio signal indicative of a spoken word or phrase;
convert the detected analog audio signal to a digital signal;
input the digital signal to a digital signal processor (DSP) within the communication device for processing to determine a signal to noise ratio (SNR) of the digital signal and whether the digital signal matches a known keyword;
when the DSP processing determines the digital signal matches the known keyword, cause the communication device to wake up and present an acknowledgement;
when the DSP processing could not positively determine that the digital signal matches the keyword, determine whether the SNR of the digital signal exceeds a first SNR threshold (T1) and, if so, cause the communication device to remain in a sleep state, and, if not, input the digital signal to an applications processor (AP) on the communications device for more stringent processing to determine if the digital signal matches the known keyword;
when the AP processing determines the digital signal matches the known keyword, cause the communication device to wake up and present an acknowledgement;
when the AP processing could not positively determine that the digital signal matches the keyword, determine whether the SNR of the digital signal exceeds a second SNR threshold (T2), where T2 is less than T1, and, if so, cause the communication device to remain in the sleep state, and, if not, send the digital signal to a server processor for even more stringent analysis to determine if the digital signal matches the known keyword; and
receive a pass or fail result from the server processor wherein if the result is fail, cause the communication device to remain in the sleep state and if the result is pass, cause the communication device to wake up and present an acknowledgement.
12. The at least one non-transitory computer-readable medium of claim 11 , wherein the acknowledgement presented by the communication device is an audible tone.
13. The at least one non-transitory computer-readable medium of claim 11 , wherein the acknowledgement presented by the communication device is a visual light indicator.
14. The at least one non-transitory computer-readable medium of claim 11 , wherein the T1 threshold is 6 dB.
15. The at least one non-transitory computer-readable medium of claim 14 , wherein the T2 threshold is 3 dB.