In this case, we investigated the possibility of using the acousto-optic effect for localizing sound source. The main idea here is that the laser beam of the LDV integrates the pressure along the line followed by the light. If the pressure along this line is constant, the integral yields a large value. Otherwise, if the pressure fluctuates, the integrated pressure will tend to cancel out, thus, yielding a smaller value. This feature can be used to localize noise sources in the far field. The measurement principle proposed to exploit the spatial filtering performed by the laser beam is illustrated in the picture below.
As can be seen, the acousto-optic beamformer, that is, the LDV together with the reflecting point, are hit by plane waves emitted by the noise source. When the laser beam is parallel to the frontwave, the output of the LDV is maximum. This makes it possible to locate the source. Furthermore, the proposed acousto-optic beamformer renders total immunity to spatial aliasing. This property cannot be achieved with conventional beamforming systems. You can find more information in my list of publications.
Experimental setup and results
The proposed acousto-optic beamformer was built in an anechoic room where a loudspeaker was located 5 m far away from the beamforming system. A picture of the experimental setup can be seen below.
All the equipment was covered with as much absorbing material in order to reduce the scattering effects at high frequencies. An accelerometer was used to monitor the mechanical vibration of the reflecting point. The results obtained with this setup can be seen in the figure below.
As can be seen, the loudspeaker was located at 0 degrees and the secundary lobes did not pollute the map at high frequencies, where otherwise spatial aliasing artifacts would have appeared if we had used a conventional beamforming technique. For ease of comparison, a line array of 19 microphones was mounted on top of the acousto-optic beamformer with approximately the same length as the laser beam. The results obtained using conventional delay-and-sum beamforming (DSB) can be seen in the following picture.
In this case, clear traces of spatial aliasing can be seen at high frequencies. The comparison between these two beamforming systems confirms the immunity of the acousto-optic beamformer to spatial aliasing.
Separating the Radetzky March from Prelude to Act 1 of Carmen
Here you can listen to some audio samples that illustrate the spatial filtering performed by the acousto-optic beamformer. The sample titled 'A single microphone' is a recording of the sound pressure measured with a microphone located at the center of the beamforming system. As you can hear, there are two different songs playing at the same time. These songs were played by two loudspeaker placed at two different positions (about 60 degrees separation from the beamformer point of view). The other two audio samples show the signals retrieved with the acousto-optic beamformer. Can you distinguish the two songs better by listening to these two samples?
(Note that the signals have been high-pass filtered in order to compensate for the bad performance of the beamforming system at low frequencies)