Abstract: A method and apparatus for sensing audio frequency-pressure modulation of the moisture content of the atmosphere caused by a moisture-laden audio source, such as a live human voice, is disclosed. A first light emitter is provided to generate a first light beam. An audio source that emits moisture, and the environment, modulate the first light beam to generate a first modulated light beam. A first detector is provided to detect and demodulate the first modulated light beam to recover an uncompensated audio signal. A second light emitter is provided to generate a second light beam. The second light beam is modulated by the environment and not by the audio source to produce a second modulated light beam. A second detector is provided to detect and demodulate the second modulated light beam to recover the signal component caused by environmental conditions.