I bounced a laser beam off of a window in my house and recovered the audio from inside the room via the beam deflection. I used a Hamamatsu S7815 amplified photodiode and connected it with a 9V battery to my stereo's microphone input jack. The audio quality was very low -- probably due to the double-pane windows in my house. Speech was just barely intelligible. I also tested the procedure of bouncing a laser beam off of a framed picture that is hanging on the wall inside the room to be monitored. The reflected beam will hit a wall somewhere else in the room, and the dot can be monitored by a telescope from remote. The goal would be to measure the beam wobble via the telescope and recover the audio without needing a stringent geometric relation to the target room. This didn't work at all, but I think with a sensitive detector, it has potential. More about laser microphones: http://www.williamson-labs.com/laser-mic.htm