Video MPEG-2 coding
There are two distinguishing features of a video clip, both of which are utilised by MPEG in its data compression technique. The first is that a video piece is a sequence of still images and as such can be compressed using the same technique as that used by JPEG. This is known as spatial inter-frame compression. The second feature is, in general, successive images of a video piece differ very little, making it possible to dispense with the unchanging or redundant part and send only the difference. This type of compression which is time-related, inter-frame compression known as temporal DCT compression; DCT (discrete cosine transform) being the name of the mathematical process used.
Video MPEG coding consists of three major parts: data preparation, compression (temporal and spatial) and quantisation (Figure 4.2).
Video data preparation
The purpose of video data preparation is to ensure a raw-coded sample of the picture frame organised in a way that is suitable for data compres- sion. The video information enters the video encoder in the form of line- scanned coded samples of luminance Y, and chrominance CR and CB. Video preparation involves regrouping these samples into 8 X 8 blocks to be used in spatial redundancy removal. These blocks are then rearranged into 16 X 16 macroblocks to be used in temporal redundancy removal. The macroblocks are then grouped into slices which are the basic units for data compression. The make up of a macroblock is determined by the chosen MPEG-2 profile. Using 4:2:0 sampling, a macroblock will consist of four
blocks of luminance and one block of each of the chrominance components CR and CB. Theoretically, a slice may range from one macroblock up to the whole picture. But in practice a slice will cover a complete picture row or part of a picture row.