Non-uniform coding schemes, such as μ-law (or U-law) and A-law, are companding techniques which are used to preserve dynamic range of a signal whilst reducing the number of bits used. Companding involves compressing the data before transmission and then expanding the data at the other end.
This method of compression is lossy, meaning some information is lost. However, this loss of detail is generally not as noticeable compared to linear compression because the human ear perceives sound ‘volume’ in a logarithmic fashion: the absolute power difference between two quiet sounds is more noticeable than two loud sounds with the same absolute power difference. Thus we can represent higher values with less precision than lower values with less perceived loss of quality.
These algorithms are used in telephony systems around the world. μ-law is used in North America and Japan, and it has a higher dynamic range than A-law, at the expense of worse distortion of small signals. A-law is commonly used in Europe and takes precedence over μ-law with international calls if either side of a call uses it. Other countries also use these algorithms.
(Above image was by Ozhiker - Own work This W3C-unspecified plot was created with Gnuplot., CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=8241430)
Here are some useful resources with more information
Comments