Digital signal processing (DSP) generally involves the mathematical processing of data, usually in the form of voltage level samples. Various characteristics, including intensity and interference, are typically represented by numbers which a computer can change to modify the signal. In addition to changing it, the numbers can be used to gather information, such as from communications, medical, or scientific equipment. Computer processors and other electronics are often used to incorporate DSP into various applications.
One component that is often used is the digital signal converter. It can convert signals from analog to digital formats, which is sometimes performed on input signals to sample them and calculate a numerical value. Voltage is usually the characteristic used to identify the traits of the signal. Time intervals for sampling are typically spaced evenly, while the voltage level at a particular moment in time is typically converted into a number. The actual signal is typically better reconstructed when the sampling occurs at a relatively fast rate in relation to time.
The next step is often digital to analog conversion; a signal returned to an analog, mathematically processed form is typically changed, and sometimes amplified for use with loudspeakers or headphones, for example. A digital cable signal is often used for television, but telephones can incorporate features such as data compression, echo reduction, and filtering with DSP. Scientific data acquisition instruments, and those used for recording earthquakes, typically use digital signal processing as well. Other systems such as military radar, medical imaging systems, and equipment used to process spacecraft photographs also do.
Filtering is a common operation; aspects of a signal can be changed, such as the actual waveform as well as its frequency. Digital signal processing sometimes adds a delay so a processor can compare parts of the data to make changes. Sudden impulses can also be filtered out by using sophisticated mathematical techniques. These sometimes include operations called Discrete Fourier Transform (DFT), inverse DFT, and Fast Fourier Transform.
Digital signal processing first began in the 1960s when digital computers were developed. It was advanced mostly through radar systems, oil exploration, medical imaging, and in space technology. Personal computers incorporated DSP starting in the 1980s, while mobile telephones and voice mail systems generally spurred its development through the 1990s. Scientists and engineers usually require skills in digital signal processing even while studying as undergraduate students; they also often need to be familiar with communications, probability, and statistical concepts.