Beamforming of radio and sound waves has many applications. In the case of sound waves, we can think of different festivals, concerts, where sound is directed towards the audience but there may be a need for beamforming even in a room using a home cinema system. Besides that, there are many possibilities of applications.
We can use various methods for beamforming, but if we wish to achieve a precise control, we need more loudspeakers assembled as an array. The shape of the array and its exact structure are variable, among others the array curvature can be used to direct the sound. This kind of loudspeaker arrays are commercially available too, even smaller ones intended for home use.
Usable results can be obtained by rotating the loudspeakers, but the desired radiation pattern can be best achieved by delaying the input signals of the loudspeakers (delay-and-sum method) or filtering them (by choosing any kind of filter) prior to sending it to the loudspeaker. Besides the signal processing method, the results will also depend on the type of the loudspeakers, their radiation patterns and the distances between them.
This thesis presents the steps of loudspeaker line array beamforming using MATLAB for calculations and filter design. First, I outline some of the methods used in the literature: the delay-and-sum method and the least squares FIR filter design method in frequency- and time-domain. Then I compare the methods and based on the results I choose FIR filter design in time-domain. After that, I measure the radiation patterns of two different loudspeakers and I simulate the achievable patterns using the results of the measurements. After choosing the better-performing loudspeaker I construct a loudspeaker line array from eight loudspeakers and I measure the radiation pattern of the complete array. In this way I can achieve more accurate results in filter design because I can take into account the positions of the speakers in the array as well. Using the measurement results I realize beamforming with delay-and-sum and time-domain FIR filter designing methods and I evaluate the results by simulations and real measurements. Finally, I present a comparison of the results achieved by the two beamforming methods.