TABLE OF CONTENTS In our treatment of the subject of DSP thus far, we have assumed that the signals we have been processing and analysing have been sampled appropriately by some unspecified means. In practice, when dealing with real systems, we need to think carefully about how to convert analog electrical waveforms into accurate and highly resolved discrete signal equivalents. We have already mentioned in passing the Shannon sampling theorem, which states that, to ensure fidelity, an analog signal containing frequencies extending to f 0 Hz must be sampled at a rate of at least 2 f 0 Hz. In this chapter we will look at this statement a little more rigorously, and establish precisely why it is true. Following this, we will review the methods by which analog signals are digitised, a process called analog to digital conversion (ADC). Since digital signals are often (although not always), converted back into analog form (think of a CD player), we will also examine the principles of digital to analog conversion (DAC). Finally in this chapter, we will consider some practical ADC/DAC circuit arrangements, including in our discussions the various algorithmic approaches that are often adopted to acquire, process and transmit data in off-line and real-time DSP environments.
By now we know that computers and other digital systems work with discrete, digitised data. Strictly speaking however, there is a subtle but important distinction between a signal that is merely discrete and one that is...
