Databases Reference
In-Depth Information
720
×
480
Horizontal filter
and subsample
Select odd
field
720
×
240
360
×
240
CCIR 601
Y
SIF
Select odd
field
Horizontal filter
and subsample
Vertical filter
and subsample
360
×
480
360
×
240
180
×
240
180
×
120
CCIR 601
U
,
V
SIF
F I GU R E 19 . 8
Generation of an SIF frame.
Y
U
V
4:2:2 CCIR-601
MPEG-SIF
F I GU R E 19 . 9
CCIR 601 to MPEG-SIF.
In the following we describe several of the video coding standards in existence today. Our
order of description follows the historical development of the standards. As each standard
has built upon features of previous standards this seems like a logical plan of attack. As in
the case of image compression, most of the standards for video compression are based on
the discrete cosine transform (DCT). The standard for teleconferencing applications, ITU-T
recommendation H.261, is no exception. Most systems currently in use for videoconferencing
use proprietary compression algorithms. However, in order for the equipment from different
manufacturers to communicate with each other, these systems also offer the option of using
H.261. We will describe the compression algorithm used in the H.261 standard in the next
section. We will follow that with a description of the MPEG algorithms used in Video CDs,
DVDs, and HDTV, and a discussion of the latest joint offering from ITU and MPEG.
We will also describe a new approach towards compression of video for videophone appli-
cations called three-dimensional model-based coding. This approach is far from maturity, and
