Conventionally, digital color images are represented by setting specific values of the color space coordinates for each pixel. Color spaces with decoupled luminance and chrominance coordinates (YUV type) allow the number of bits required for acceptable color description of an image to be reduced. This reduction is based on greater sensitivity of the human eye to changes in luminance than to changes in chrominance. The idea behind this approach is to set individual value of luminance component to each pixel, while assigning the same color (chrominance components) to certain groups of pixels (sometimes called macropixels
) in accordance with some specific rules. This process is called downsampling
and there are different sampling formats depending on the underlying scheme.