A device and system to remotely detect vocalizations of speech. The skin located on the throat region of a speaking person or a reflective layer on the skin on the throat region vibrates in response to vocalizations of speech by the person. The vibrating skin or reflective layer is reflective of impinging radiation. A laser Doppler vibrometer transmits radiation onto the vibrating skin or the covering reflective layer and receives reflected radiation from the vibrating skin or reflective layer. The laser Doppler vibrometer generates voltage output signals that are representative of the speech causing the vibrations. A target tracker directs the impinging radiation and detects the reflected radiation to pass between the throat region and the laser Doppler vibrometer and includes a processor that removes non-speech signal artifacts from the voltage output signals. An interconnected audio speaker reproduces the speech from the voltage output signals.