Suppressing key phrase detection in generated audio using self-trigger detector
An example apparatus for suppression of key phrase detection includes an audio receiver to receive generated audio from a loopback endpoint and captured audio from a microphone. The apparatus includes a self-trigger detector to detect a key phrase in the generated audio. The apparatus also includes a detection suppressor to suppress detection of the detected key phrase in the captured audio at a second key detector for a predetermined time in response to detecting the key phrase in the generated audio.
1. An apparatus for suppression of key phrase detection, comprising:
an audio receiver to receive generated audio from a loopback endpoint and captured audio from a microphone;
a self-trigger detector to detect a key phrase in the generated audio from the loopback endpoint; and
a detection suppressor to suppress detection of the detected key phrase in the captured audio from the microphone at a key phrase detector for a predetermined time in response to detecting the key phrase in the generated audio from the loopback endpoint.
2. The apparatus of claim 1 , wherein the generated audio comprises a reference signal.
3. The apparatus of claim 1 , wherein the loopback endpoint is to receive the generated audio from a playback endpoint.
4. The apparatus of claim 1 , wherein the captured audio comprises an echo corresponding to the generated audio.
5. The apparatus of claim 1 , wherein the predetermined time is based on the detected key phrase.
6. The apparatus of claim 1 , comprising a processor to perform echo cancellation on the captured audio.
7. The apparatus of claim 1 , comprising a processor to perform direct current (DC) removal on the captured audio to remove DC bias.
8. The apparatus of claim 1 , comprising a processor to increase gain on the captured audio.
9. The apparatus of claim 1 , comprising a processor to perform beamforming on the captured audio.
10. The apparatus of claim 1 , comprising a processor to perform noise reduction on the captured audio.
11. A method for suppressing key phrase detection, comprising:
receiving, via a processor, generated audio from a loopback endpoint and captured audio from a microphone;
detecting, via the processor, a key phrase in the generated audio from the loopback endpoint; and
suppressing, via the processor, detection of the detected key phrase in the captured audio from the microphone for a predetermined time in response to detecting the key phrase in the generated audio from the loopback endpoint.
12. The method of claim 11 , wherein the generated audio comprises a reference signal.
13. The method of claim 11 , wherein the loopback endpoint is to receive the generated audio from a playback endpoint.
14. The method of claim 11 , wherein the captured audio comprises an echo corresponding to the generated audio.
15. The method of claim 11 , wherein the predetermined time is based on the detected key phrase.
16. The method of claim 11 , comprising performing echo cancellation on the captured audio.
17. The method of claim 11 , comprising performing direct current (DC) removal on the captured audio to remove DC bias.
18. The method of claim 11 , comprising increasing gain on the captured audio.
19. The method of claim 11 , comprising beamforming the captured audio.
20. The method of claim 11 , comprising performing noise reduction on the captured audio.
21. At least one non-transitory computer readable medium for suppressing detection of key phrases having instructions stored therein that, in response to being executed on a computing device, cause the computing device to:
receive generated audio from a loopback endpoint and audio from a microphone;
detect a key phrase in the generated audio from the loopback endpoint; and
suppress detection of the detected key phrase in the audio from the microphone at a key phrase detector for a predetermined time in response to detecting the key phrase in the generated audio from the loopback endpoint.
22. The at least one non-transitory computer readable medium of claim 21 , wherein the generated audio comprises a reference signal.
23. The at least one non-transitory computer readable medium of claim 21 , wherein the loopback endpoint is to receive the generated audio from a playback endpoint.
24. The at least one non-transitory computer readable medium of claim 21 , wherein the captured audio comprises an echo corresponding to the generated audio.
25. The at least one non-transitory computer readable medium of claim 21 , wherein the predetermined time is based on the detected key phrase.