Microphone array

One microphone tells you that a sound happened. Several microphones, placed apart, tell you where it came from.

Sound reaches the nearer microphone a fraction of a second before the far one. Software measures that tiny gap and turns it into a direction, the same way your two ears let you point at a voice behind you without looking.

XPENG says its humanoid robot uses nine microphones to estimate the direction of speech, then checks lip movement on camera to decide which person is actually talking.