Smart Speaker

Overview

A smart speaker is a network-connected speaker built around a voice assistant: it listens for a wake word, streams the request to a cloud service, and plays back the answer or music through its own drivers. The acoustics and electronics share one cylindrical body. The Enclosure Assembly forms a sealed chamber for the Driver Array, a wrap-around mesh grille lets sound and the user's voice pass while hiding the internals, and a weighted base keeps the unit planted. Around the top edge sits the Microphone Array, a ring of far-field microphones that pick up speech from across a room, and just below the top surface the Light Ring glows to show when the speaker is listening, thinking, or muted.

The brains live on the Main Board, which carries an application SoC, an always-on co-processor, system memory, flash, and a Wi-Fi/Bluetooth radio. Touch and physical controls on the Control Panel handle volume and microphone mute without needing voice.

How it works

The six Microphone Array microphones each hear the room slightly out of phase. A beamforming algorithm on the SoC combines the channels to steer a virtual "beam" toward whoever is speaking and reject noise and the speaker's own playback. A low-power co-processor MCU on the main board runs the wake-word model continuously; only after it triggers does the SoC wake fully and stream audio over Wi-Fi to the assistant service. The returned audio is decoded by the SoC, sent to the Amplifier Board Class-D amplifier, and reproduced by the woofer and tweeter, with the passive radiator extending the bass that a small sealed box cannot produce on its own.

Power enters through the DC Barrel Jack from an external adapter and feeds the internal supply that rails the boards. The light ring mirrors assistant state in real time, and pressing mic-mute on the control panel cuts the microphone array's supply so the device cannot listen.