Quantization

Quantization (also called discretization) is the process of forcing data points representing a real-valued quantity to conform to a set of discrete values. For example, a logger may report the wind speeds measured by an anemometer to a single decimal place; these wind speed values would be quantized to the nearest multiple of 0.1.

The graph below shows a real-valued signal (blue) that is rounded to the nearest integer to produce a quantized signal (orange).

Because computers have finite precision, all numbers used by a computer are technically quantized by the digitalization process, even if the level of quantization is so small as to be utterly negligible. Therefore when we refer to quantized wind data, we generally refer to data that has been further rounded by some process other than digitalization to the extent that the distribution of the values is noticeably 'chunky'. For example, the graph below shows the histogram of some wind speed data that has been rounded to the nearest integer value of m/s. The bin size of the histogram is 0.5 m/s, which shows the empty bins where there is no data because of the quantization.

If your data shows this kind of severe quantization and you wish to 'repair' it by randomizing the data in a way that is faithful to the statistics of the data, use the Fix Quantization window.