Quantization Noise
Analog signals are digitally encoded for transmission across the PSTN.
The encoding process introduces quantization error.
Quantization error can be reduced by using more bits to represent each point or by using a non-linear quantization technique.
The USA uses a ?255 non-linear quantization technique with 8 bits per sample. This provides quantization error (in a certain limited way) equivalent to a 13 to 14 bit linear technique.