Digital Video Compression

Video Frames
-A video is a sequence of pictures called frames.

-Why subsampling ? Human eye is more sensitive to luminance than colour.
Loss in colour precision

Transform coding
Spatial redundancy
Why Macroblock is necessary.
How transform coding can be used to obtain compression ?

Concentrates the energy of the signal and allows quantization
Compare quantization in the pixel domain vs transform domain to show the potential of transform coding.

RLE - Run Length Encoding
Coefficients are sparse and the matrix has a high number of zeros.
Coefficient ZigZag scanning as a tool to improve the RLE encoding performance.

Inter frame prediction
Temporal redundancy
Despite the high compression obtained by the transform coding in the intra frame prediction, the inter frame redudancy is still significant. As a result, exploiting the correlation between frames is a obvious step.
Motion Estimation
Last Update: 30th June 2010
Online since: 30th June 2010