Compression starts within an individual video frame.

Spatial Compression

Each video frame has lots of redundancy in it. Blue sky, grass, walls, furniture, uniforms, and so on. Spatial compression seeks to eliminate this redundancy and save bandwidth by keeping only the key information. For example, a compression system might chop a video frame into 4 pixel by 4 pixel and 8 pixel by 8 pixel blocks of information. If there is blue sky in those blocks, it might not be necessary to include all of the information from each of the 64 pixels. But the information for one pixel could be sent down the line with a code to repeat the same thing for the other 63 pixels. This would eliminate redundancy and reduce bandwidth requirements.

Entropy Reduction

Systems will also take information that is not very important and throw it away. This is called entropy reduction.

Entropy Encoding

Another technique uses shorter bit strings to represent commonly used colors and shades and longer bit strings to represent uncommonly used colors and shades. It would be like the computer saying: “Use this 6-bit string to mean this color of red that would normally need 8 bits,” and “use this 10-bit string to mean this color of puce that normally requires 8 bits.” This technique is called entropy encoding.

A number of different mathematical formulas, called algorithms, can be used in video compression. In order to make the system work, the transmitting end and the receiving end of the system must be using the same formula. These formulas are always a compromise. If you don't send enough information down the line, the final picture will not be as sharp and clear as it should be. On the other hand, if you are trying to save bandwidth, you don't want to send any more information than you have to.

Some compression is done by using 4 by 4 and 8 by 8 blocks of pixels.

Entropy encoding.