Amazon Technologies, Inc.

David Berol

Philippe Andre Lantin

Ori Neidich

Systems and methods for acoustic event detection model selection are disclosed. For example, models may be calibrated utilizing audio samples from sound output by specific audio-output devices. Thereafter, a determination may be made as to which audio-input devices are near a give audio-output device and a model selector may select a model trained to detect an acoustic event produced by the audio-output device. The proximate audio-input device(s) may be configured with the selected model to detect the acoustic event.

Acoustic event detection model selection

Described herein is a system for recalibrating an audio configuration for mobile or moving devices. The system may configure a multi-device output group to generate synchronous output audio using multiple devices. For example, the output group may include a first device generating a first portion of output audio corresponding to a first channel and a second device generating a second portion of the output audio corresponding to a second channel. If the second device detects motion and/or movement indicating a change in its location, the system may recalibrate the output group to continue generating the output audio without the second device. For example, the first device or a new device can generate the second portion of the output audio instead of the second device. When the second device returns, the system can recalibrate the output group to include the second device again.

Audio output configuration for moving devices

A user may access multiple virtual assistants via a voice-enabled device. The device may receive a command from the user, detect a wakeword corresponding to one of the assistants, and send audio data to a command processing system corresponding to the selected assistant. The device transmits encrypted audio data to one or more systems and, upon detecting a wakeword or wake command corresponding to one of the systems, the device may provide an encryption key to that particular system. The system may decrypt and process the audio data without additional latency introduced by having to wait for the audio data to arrive.

Growing community of inventors

Sunnyvale, CA, United States of America

David Berol