Chapter 3: Principles of Video Compression

Overview

The statistical analysis of video signals indicates that there is a strong correlation both between successive picture frames and within the picture elements themselves. Theoretically, decorrelation of these signals can lead to bandwidth compression without significantly affecting image resolution. Moreover, the insensitivity of the human visual system to loss of certain spatio-temporal visual information can be exploited for further reduction. Hence, subjectively lossy compression techniques can be used to reduce video bit rates while maintaining an acceptable image quality.

For coding still images, only the spatial correlation is exploited. Such a coding technique is called intraframe coding and is the basis for JPEG coding. If temporal correlation is exploited as well, then it is called interframe coding. Interframe predictive coding is the main coding principle that is used in all standard video codecs, such as H.261, H.263, MPEG-1, 2 and 4. It is based on three fundamental redundancy reduction principles:

1. Spatial redundancy reduction: to reduce spatial redundancy among the pixels within a picture (similarity of pixels, within the frames), by employing some data compressors, such as transform coding.

2. Temporal redundancy reduction: to remove similarities between the successive pictures, by coding their differences.

3. Entropy coding: to reduce the redundancy between the compressed data symbols, using variable length coding techniques.

A detailed description of these redundancy reduction techniques is given in the following sections.