video compression

Video compression is a lossy compression that takes advantage of inertia and varying eye sensitivity for brightness and color resolution. It is based on local and temporal redundancies. In local redundancies, two adjacent pixels are identical; in temporal redundancies, the pixels in two consecutive images are the same or they differ only slightly. The temporal and local redundancies, like the minimal differences between pixels, offer the possibility of data reduction.

With respect to the sensitivity between brightness and color, video compression uses colormodels in which colors and brightness are separate, as in the YUV color model or the YCbCr color model. Since in the latter color models the color resolution is much lower than the resolution of the brightness values, the color data can already be reduced by color subsampling. The color information is further reduced, for example, by a DCT transformation, by requantization and by the picture groups corresponding to the Group of Pictures( GOP) in MPEG compression.

Two international standardization institutes, the ITU-T and ISO/ IEC, develop standards for video compression. At the ITU-T, they are called H-Recommendations and have the designations H.261, H. 262, H. 263, H. 264/ AVC, and H.265. The ISO-IEC standards have the designations MPEG-x: MPEG-1, MPEG-2, and MPEG-4.

Most of the ITU recommendations are for real-time applications such as video conferencing and telephony, while the ISO/IEC standards target video storage, broadcast video such as DVB-T, DVB-S, and DVB-C, and video streaming over the Internet.