A Simple Approach to Digital Signal Processing

3

Converting Analog to Digital

We have seen that signals in the real world are analog. In this chapter we shall first look at how we convert them into digital patterns that can be used by a DSP and then how we rebuild the analog output signal.

The first part of the process that we shall look at is called sampling. To illustrate this operation we shall use the changing weight of a baby, as shown in Figure 3.1. The information was actually gathered from the fortnightly visit to the clinic where the baby was weighed, as shown in Table 3.1. Plotting these values against time (Figure 3.2), we see we have a set of snapshots of the baby's progress. These snapshots are called samples of the input signal (the baby's weight). In other words, sampling an input signal is a method of recording an instantaneous value of that signal.

There are many other examples that we can use. We often see daily figures for rainfall, the number of hours of sunshine and the depth of snow in ski resorts. What is notable is that the sampling period is always a fixed unit: fortnightly for the child and daily for the rain, sunshine, and depth of snow. This allows us to understand the information given and to possibly make a decision on what to do next.

Let us take another example: A stockbroker wants to predict what is likely to happen to his portfolio of shares. We shall assume that the stock index is going to vary continuously over time, as shown in Figure 3.3. Also assume that the stock index was published only ...