Analog-to-Digital

Unfinished report...

In Nature all objects are "digital" at sub-atomic level, i.e. they are not continuous or "analog". However for practical purposes most of them are thought as of analog.

For one particular example, let us look at a traditional B&W photograph. It cosists of many tiny silver crystals (the black "paint"), but to the naked eye these are non-existrent. Instead, the eye (or more precisely the brain) sees a continuous (in space) distribution of continuous gray levels. However, if this picture is to be displayed on any digital device - such as CRT - it has to be turned into digital form. It has to be rasterized and color levels quatized.

Let us see how a 1" by 1" B&W photograph is turned into N by N pixels @ C bpp. The process of rasterization breaks the (space - here plane) continuity into pixels. It defines the resolution (usually dots-per-inch, dpi). The bigger the N is - the higher the resolution will be from the same original., so smaller details can be preserved. If our picture is rasterized at 100dpi, we will get a 100 by 100 px (or 1x10^4px) to work with. So now that we have defined the pixels, we have to define the color of each of them - it can only be one numeric value. We also need to specify the desired color depth (usually in bits-per-pixel, bpp). Let us choose a 256 gray-levels color depth. Thus we need 8 bits (2^8 =256) to encode the color of each pixel. By quantization all the input colors (thousands of them) are stripped down to a mere 256 gray levels by on of the many quantization algorithms. As a result we will have 100 by 100 px, 8 bpp digital image ready to display.