Computer Science; Information Technology; Network Design
Digital signal processors (DSPs) are microprocessors designed for a special function. DSPs are used with analog signals to continuously monitor their output, often performing additional functions such as filtering or measuring the signal. One of the strengths of DSPs is that they can process more than one instruction or piece of data at a time.
Digital signal processors (DSPs) are microprocessors designed for a special function. Semiconductor intellectual property (SIP) blocks designed for use as DSPs generally have to work in a very low latency environment. They are constantly processing streams of video or audio. They also need to keep power consumption to a minimum, particularly in the case of mobile devices, which rely heavily on DSPs. To make this possible, DSPs are designed to work efficiently on both fixed-point arithmetic and the more computationally intensive floating-point arithmetic. The latter, however, is not needed in most DSP applications. DSPs tend to use chip architectures that allow them to fetch multiple instructions at once, such as the Harvard architecture. Many DSPs are required to accept analog data as input, convert this to digital data, perform some operation, and then convert the digital signals back to analog for output. This gives DSPs a pipelined architecture. They use a multistep process in which the output of one step is the input needed by the next step.
An example of the type of work performed by a DSP can be seen in a multiplier-accumulator. This is a piece of hardware that performs a two-step operation. First, it receives two values as inputs and multiplies one value by the other. Next, the multiplier-accumulator takes the result of the first step and adds it to the value stored in the accumulator. At the end, the accumulator's value can be passed along as output. Because DSPs rely heavily on multiplier-accumulator operations, these are part of the instruction set hardwired into such chips. DSPs must be able to carry out these operations quickly in order to keep up with the continuous stream of data that they receive.
The processing performed by DSPs can sometimes seem mysterious. However, in reality it often amounts to the performance of fairly straightforward mathematical operations on each value in the stream of data. Each piece of analog input is converted to a digital value. This digital value is then added to, multiplied by, subtracted from, or divided by another value. The result is a modified data stream that can then be passed to another process or generated as output.
There are many applications in which DSPs have become an integral part of daily life. The basic purpose of a DSP is to accept as input some form of analog information from the real world. This could include anything from an audible bird call to a live video-feed broadcast from the scene of a news event.
Digital signal processing is also critical to many military applications, particularly those that rely on the use of sonar or radar. As with ultrasound devices, radar and sonar send out analog signals in the form of energy waves. These waves bounce off features in the environment and back to the radar- or sonar-generating device. The device uses DSPs to receive this analog information and convert it into digital data. The data can then be analyzed and converted into graphical displays that humans can easily interpret. DSPs must be able to minimize delays in processing, because a submarine using sonar to navigate underwater cannot afford to wait to find out whether obstacles are in its path.
A type of digital signal processing that many people have encountered at one time or another in their lives is the fingerprint scanner used in many security situations to verify one's identity. These scanners allow a person to place his or her finger on the scanner, and the scanner receives the analog input of the person's fingerprint. This input is then converted to a digital format and compared to the digital data on file for the person to see whether they match. As biometric data becomes increasingly important for security applications, the importance of digital signal processing will likely grow.
—Scott Zimmer, JD
Binh, Le Nguyen. Digital Processing: Optical Transmission and Coherent Receiving Techniques. Boca Raton: CRC, 2013. Print.
Iniewski, Krzysztof. Embedded Systems: Hardware, Design, and Implementation. Hoboken: Wiley, 2013. Print.
Kuo, Sen M., Bob H. Lee, and Wenshun Tian. Real-Time Digital Signal Processing: Fundamentals, Implementations and Applications. 3rd ed. Hoboken: Wiley, 2013. Print.
Snoke, David W. Electronics: A Physical Approach. Boston: Addison, 2014. Print.
Sozan´ski, Krzysztof. Digital Signal Processing in Power Electronics Control Circuits. New York: Springer, 2013. Print.
Tan, Li, and Jean Jiang. Digital Signal Processing: Fundamentals and Applications. 2nd ed. Boston: Academic, 2013. Print.