Microphones

Microphone array design is critical to robot audio performance. A single microphone captures sound from all directions equally, making it impossible to focus on a specific speaker in a noisy room. Arrays of 2, 4, 6, or more microphones spaced across the robot's body enable beamforming — the computational process of combining signals from multiple microphones to create a directional listening pattern that enhances sound from the desired direction while suppressing noise from other directions. The spacing between microphones determines the frequency range over which beamforming is effective: wider spacing improves low-frequency directionality, while closely spaced microphones handle high-frequency beamforming. Many robots combine microphones at different spacings to cover the full speech frequency range (roughly 100 Hz to 8 kHz).

Far-field voice capture — recognizing commands spoken from several meters away — is one of the most challenging audio processing tasks. The robot must distinguish the user's voice from background noise (television, music, conversations), echo from its own speaker output, and the sound of its own motors and mechanisms. Advanced echo cancellation algorithms subtract the robot's known speaker output from the microphone signal, while noise reduction algorithms trained on thousands of hours of real-world audio data suppress environmental interference. The quality of these processing algorithms, combined with the physical microphone array design, determines whether a robot reliably responds to voice commands from across the room or requires users to speak loudly from close range.