Abstract: Techniques for using both speaker-identification information and other characteristics associated with received voice commands to determine how and whether to respond to the received voice commands. A user may interact with a device through speech by providing voice commands. After beginning an interaction with the user, the device may detect subsequent speech, which may originate from the user, from another user, or from another source. The device may then use speaker-identification information and other characteristics associated with the speech to attempt to determine whether or not the user interacting with the device uttered the speech. The device may then interpret the speech as a valid voice command and may perform a corresponding operation in response to determining that the user did indeed utter the speech. If the device determines that the user did not utter the speech, however, then the device may refrain from taking action on the speech.

Abstract: Natural language controlled devices may be configured to activate command recognition in response to one or more wake words. Techniques are provided to allow for multiple operating modes in which different recognition parameters are employed in recognizing wake words that activate the natural language control functionality of a computing device.

Abstract: An audio device may be configured to produce output audio and to capture input audio for speech recognition. In some cases, a second device may also be used to capture input audio to improve isolation of input audio with respect to the output audio. In addition, acoustic echo cancellation (AEC) may be used to remove components of output audio from input signals of the first and second devices. AEC may be implemented by an adaptive filter based on dynamically optimized filter coefficients. The filter coefficients may be analyzed to detect situations in which the first and second devices are too close to each other, and the user may then be prompted to increase the distance between the two devices.

Abstract: Residual echo that remains after an echo cancellation process may interfere with speech recognition. If near-end speech is detected in an audio input signal, a controller may attenuate the audio playback signal. Too much attenuation may disturb playback, whereas too little attenuation may not improve speech recognition. Accordingly, features are disclosed for attenuating an audio playback signal based at least in part on residual echo level.