This chapter has introduced block-based video compression techniques currently used in many video coding standards. It described how the codecs use knowledge about the human vision system to achieve good compression efficiency. The human vision system is more sensitive to luminance than chrominance information, and is also more sensitive to low-frequency information than high-frequency information. The compression techniques therefore operate in the frequency domain, to ensure that psycho-visual effects can be taken into account. The discrete cosine transform (DCT) is commonly used in coding standards. DCT coefficients are quantized and run-level coded to achieve compression. In addition, similarities between successive video frames are exploited by using a technique called motion compensation (MC) to form a prediction of the current frame from previous frames in the video sequence.
The chapter also discussed some more advanced video coding techniques that are used in standards such as H.264 to achieve compression and functionality superior to earlier standard codecs. These include an alternative transform to the DCT, more MB modes to enable more accurate representation of the motion within the video sequence, and an adaptive deblocking filter.
A brief comparison was made between the different codecs standardized by the ITU and the ISO (MPEG).
Finally, different methods of assessing video quality were discussed. Subjective evaluation is capable of providing the ...