Before a signal can be converted from analogue into digital form, there must some steps be taken. Sampling is the first step, in which, instances of the analogue signal, with a certain repetitive speed is gathered, stored, and forwarded to the next step device.

Sound, in nature, is a series of curved waves like ripples on a pond. If you look at these waves from another angle, (the pond's surface, to continue the metaphor) sound is a series of curved, regular disturbances, moving through some physical medium (water or air, for instance). Notice, the word curved appears in both of these viewpoints. That's a problem because computers understand discrete numbers, not curves. To get around this, digital recording devices measure a sound's characteristics at regular intervals. When we record sound with a computer, we're approximating a group of curves by recording measurements of them, taken at regular intervals. This regular measurement, called sampling, gives a rough, stair-stepped approximation of a curve. The recording's fidelity (meaning, how closely it resembles the original sound) is mainly a result of two factors that we can control:

1. How often we take the measurements sampling rate, and

(2) How precisely we record the measurements sampling resolution.