Noise classification for event detection
In one aspect, a network microphone device includes a plurality of microphones and is configured to detect sound via the one or more microphones. The network microphone device may capture sound data based on the detected sound in a first buffer, and capture metadata associated with the detected sound in a second buffer. The network microphone device may classify one or more noises in the detected sound and cause the network microphone device to perform an action based on the classification of the respective one or more noises.
1. A method comprising:
detecting sound via one or more microphones of a network microphone device (NMD), wherein the detected sound includes a voice utterance;
capturing first sound data in a first buffer of the NMD based on the detected sound;
analyzing, via the NMD, the first sound data to detect a wake word;
based on the analyzed first sound data, detecting the wake word;
after detecting the wake word, transmitting at least the voice utterance to one or more remote computing devices associated with a voice assistant service;
detecting additional sound via the one or more microphones;
capturing second sound data in the first buffer based on the detected additional sound;
analyzing, via the NMD, the second sound data to detect the wake word, wherein the wake word is not detected based on the analyzed second sound data;
capturing metadata associated with the detected additional sound in a second buffer of the NMD;
processing the metadata to classify one or more noises in the detected additional sound; and
causing the NMD to perform an action based on the classification of the respective one or more noises.
2. The method of claim 1 , wherein:
the second sound data transmitted to the one or more servers comprises recorded audio; and
the metadata comprises spectral information that is temporally disassociated from the recorded audio.
3. The method of claim 1 , wherein processing the metadata comprises transmitting the metadata to one or more other remote servers for analyzing the metadata.
4. The method of claim 1 , wherein processing the metadata comprises locally analyzing the metadata and classifying the one or more noises via the NMD.
5. The method of claim 1 , wherein classifying the one or more noises comprises comparing the metadata to reference metadata associated with known noise events.
6. The method of claim 1 , wherein causing the NMD to perform an action comprises at least one of: playing back a sound via the NMD, sending a notification to a user's mobile computing device, or flashing a light.
7. The method of claim 1 , wherein causing the NMD to perform an action is based on at least one of a sound pressure level or a directionality of the sound.
8. A network microphone device (NMD), comprising:
one or more processors;
one or more microphones;
a first buffer;
a second buffer;
a tangible, non-transitory, computer-readable medium storing instructions executable by the one or more processors to cause the NMD to perform operations comprising:
detecting sound via one or more microphones of the NMD, wherein the detected sound includes a voice utterance;
capturing first sound data in the first buffer based on the detected sound;
analyzing the first sound data to detect a wake word;
based on the analyzed first sound data, detecting the wake word;
after detecting the wake word, transmitting at least the voice utterance to one or more remote computing devices associated with a voice assistant service;
detecting additional sound via the one or more microphones;
capturing second sound data in the first buffer based on the detected additional sound;
analyzing the second sound data to detect the wake word, wherein the wake word is not detected based on the analyzed second sound data;
capturing metadata associated with the detected additional sound in the second buffer;
processing the metadata to classify one or more noises in the detected additional sound; and
performing an action based on the classification of the respective one or more noises.
9. The NMD of claim 8 , wherein:
the second sound data transmitted to the one or more servers comprises recorded audio; and
the metadata comprises spectral information that is temporally disassociated from the recorded audio.
10. The NMD of claim 8 , wherein processing the metadata comprises transmitting the metadata to one or more other remote servers for analyzing the metadata.
11. The NMD of claim 8 , wherein processing the metadata comprises locally analyzing the metadata and classifying the one or more noises via the NMD.
12. The NMD of claim 8 , wherein classifying the one or more noises comprises comparing the metadata to reference metadata associated with known noise events.
13. The NMD of claim 8 , wherein performing an action comprises at least one of: playing back a sound via the NMD, sending a notification to a user's mobile computing device, or flashing a light.
14. The NMD of claim 8 , wherein causing the NMD to perform an action is based on at least one of a sound pressure level or a directionality of the sound.
15. Tangible, non-transitory, computer-readable medium storing instructions executable by one or more processors to cause a network microphone device (NMD) to perform operations comprising:
detecting sound via one or more microphones of the NMD, wherein the detected sound includes a voice utterance;
capturing first sound data in a first buffer of the NMD based on the detected sound;
analyzing the first sound data to detect a wake word;
based on the analyzed first sound data, detecting the wake word;
after detecting the wake word, transmitting at least the voice utterance to one or more remote computing devices associated with a voice assistant service;
detecting additional sound via the one or more microphones;
capturing second sound data in the first buffer based on the detected additional sound;
analyzing the second sound data to detect the wake word, wherein the wake word is not detected based on the analyzed second sound data;
capturing metadata associated with the detected additional sound in a second buffer of the NMD;
processing the metadata to classify one or more noises in the detected additional sound; and
performing an action based on the classification of the respective one or more noises.
16. The tangible, non-transitory, computer-readable medium of claim 15 , wherein:
the second sound data transmitted to the one or more servers comprises recorded audio; and
the metadata comprises spectral information that is temporally disassociated from the recorded audio.
17. The tangible, non-transitory, computer-readable medium of claim 15 , wherein processing the metadata comprises transmitting the metadata to one or more other remote servers for analyzing the metadata.
18. The tangible, non-transitory, computer-readable medium of claim 15 , wherein processing the metadata comprises locally analyzing the metadata and classifying the one or more noises via the NMD.
19. The tangible, non-transitory, computer-readable medium of claim 15 , wherein classifying the one or more noises comprises comparing the metadata to reference metadata associated with known noise events.
20. The tangible, non-transitory, computer-readable medium of claim 15 , wherein performing an action comprises at least one of: playing back a sound via the NMD, sending a notification to a user's mobile computing device, or flashing a light.