Methods and apparatus for detecting a voice command

At least one non-transitory computer readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform a method of monitoring an acoustic environment of a device to receive acoustic input from the environment of the device while the device is operating in a low power mode, the method comprising:

• performing at least one processing stage on the acoustic input, the at least one processing stage comprising:
• performing automatic speech recognition (ASR), while in the low power mode, to recognize at least one word or phrase in the acoustic input;
• determining, while in the low power mode, whether the recognized at least one word or phrase matches a trigger word or phrase;
• in response to determining that the recognized at least one word or phrase matches the trigger word or phrase, requesting, via a network while in the low power mode, that at least one server perform recognition on at least a portion of the acoustic input to determine whether the acoustic input includes one or more voice commands;
• in response to receiving an indication from the at least one server that the acoustic input includes at least one voice command, selecting whether to remain in the lower power mode or to transition to a higher power mode, and
• responding to the at least one voice command recognized by the at least one server, in either the lower power mode or the higher power mode, depending on the selecting.

A device comprising:

• at least one processor configured to process acoustic input from an environment of the device while the device is operating in a low power mode, the processing comprising:
• performing automatic speech recognition (ASR), while in the low power mode, to recognize at least one word or phrase in the acoustic input;
• determining, while in the low power mode, whether the recognized at least one word or phrase matches a trigger word or phrase;
• in response to determining that the recognized at least one first word or phrase matches the trigger word or phrase, requesting that at least one server perform recognition on at least a portion of the acoustic input to determine whether the acoustic input includes one or more voice commands;
• in response to receiving an indication from the at least one server that the acoustic input includes at least one voice command, selecting whether to remain in the lower power mode or to transition to a higher power mode;
• and
• performing at least one task to carry out the at least one voice command in the acoustic input recognized by the at least one server, in either the lower power mode or the higher power mode, depending on the selecting.