Analog-to-digital conversion

The graph of the figure illustrates a method of analog-to-digital (A/D) conversion. A voice signal, or continuously variable signal, is digitized, or converted into the string of pulses, whose amplitudes can achieve only a finite number of states.

Resolution
The number of states is always a power of 2, such that it is represented as a binary number code. Fidelity becomes better as the exponent increases. The number of states is called as sampling resolution, or simply resolution.You may think that the resolution would have to be quite large for good reproduction to be possible. But actually a resolution of 23 = 8  is good enough for the voice transmission, and is standard resolution for the commercial digital voice circuits. A resolution of 24 = 16 is adequate for the compact disks which is used in advanced hi-fi systems.

Sampling  rate
The efficiency with which the signal can be digitized depends on frequency at which sampling is done. Generally the sampling  rate should have a frequency at least twice the highest data frequency.For the audio signal with the components  as high as 3 kHz, minimum sampling rate for the effective digitization is 6 kHz, or a sample every 167 microseconds (µs). Ideally, sampling rate should be higher; the commercial voice standard is 8 kHz, or one samples every 125 µs.For music and hi-fi digital transmission, standard sampling rate is 44.1 kHz, or one samples every 22.7 µs. This is based on the maximum audio frequency of 20 kHz, the estimated upper limit of human hearing range.