Abstract: A method of processing digital video information in real time involves reading-in digital data representing pixels of successive video frames, e.g., from a digitizer. The pixel frame data is initially stored and subjected to an idealization treatment to choose and pre-select representative values for pixel characteristics which are then adopted in the further processing. The idealized frame data can then be subjected to compression by deriving further data representing blocks of pixels (mini-blocks) and blocks of mini-blocks (super-blocks) and by generating code words to describe the changes in the pixel data of successive frames. The compressed data is finally stored and/or transmitted.