Digital signal processing is done by specially constructed DSPs (digital signal processors), whose architecture is designed to suit the needs of this task. The aim of DSPs is to convert continuous physical signals to digital ones. These processors achieve the goal with high efficiency (at a high speed or low power consumption), they can be integrated into PCs via an interface card or can be used alongside other processors as co-processors. The memory architecture of DSPs allows for simultaneous access of data and commands, thus increasing processing speed and power efficiency.
The aim of this thesis is to examine a DSP as a co-processor carrying out an implementation of an acoustical beamforming method. The objective of this task is to make sensors such as an array of microphones sensitive to direction, and to manipulate their characteristics in order to filter out data coming from specific points. The task is done by an SPM34CP card made by RTD Embedded Technologies with a C6678 digital signal processor made by Texas Instruments. First, I am going to examine the structure of the signal processing card, the functions of its peripherals and the connection to the host computer. After establishing the data flow between the host and the DSP, I am going to implement the delay and sum beamforming algorithm.