What you have learned

In order to digitally process continuous-time analog signals, such as those that occur naturally, they must be sampled and quantized.

Quantization approximates a real-number sample with a finite set of values.

Quantization is divided into two parts, usually with a N-bit binary representation in the middle. The first step is analog-to-digital conversion, and the second is digital-to-analog conversion.

An N-bit binary number can represent 2^N values.

There are two basic number representations: fixed-point and floating-point. Floating point has the advantage of providing a large dynamic range, at the expense of a more complicated implementation.