Design Decisions for Audio Localization Implementation 1.3 2004/12/15 10:00:34 US/Central 2004/12/15 15:55:31 US/Central Elizabeth Gregory lizzardg@rice.edu Joseph Cole joecole@rice.edu Elizabeth Gregory lizzardg@rice.edu Joseph Cole joecole@rice.edu beamforming design In this section, we will go over the design decisions we had to make in the course of our project, including determining how many microphones we should use, as well as the spacing of the array. With the theory out of the way, we have to face the real world and set more constraints.

The Number of Microphones In this project, we are using the TI TMS320C6211 DSK board. This board has two channels that sample at 48 kHz and another channel that samples at 8 kHz. Since we are not interpolating our signals, the sampling frequency is increasingly critical, so we can only use two microphones. shows what happens to the beampattern when we use reduced sampling frequency.

Beampattern with Reduced Sampling Frequency

Array Spacing In order to find the best array spacing (theoretically), we have to go through a few calculations. In our most extreme case, the signals will be hitting our array from 180°. This gives us perfect destructive interference, as our phase difference will be . Therefore, we only need half of the wavelength: d λ 2 We can find λ by dividing the frequency of our sine wave (500 Hz) by the speed of sound (346.287 m/s): λ c f 346.287 500 0.69 and our array spacing d ends up being 0.345 m.