Cryptography Reference
In-Depth Information
Theoretically there can be as many enhancement layers as possible, each
refines the previous layer to achieve an improved video quality. However,
multi-layer scalability results in heavy overheads and thus reduces the coding
e
ciency compared to non-scalable video. Therefore only two layers (base
layer and enhancement layer) are usually used in MPEG-2 scalable video
coding.
MPEG-2 Interlaced Coding
MPEG-2 supports both progressive scanning and interlaced scanning. Inter-
laced scanning is the scanning method used in the analogue television systems
where odd lines of a frame are scanned first as one field (odd field), and even
lines (field) are then scanned after the odd field.
An interlaced video coding can use two different picture structures: frame
structure and field structure. In the frame structure, lines of two fields are
coded together as a frame, the same way as in sequential scanning. One picture
header is used for two fields. In the field structure, the two fields of a frame are
coded independently, and the odd field is followed by the even field. However
cross field predictive coding is possible as the two fields can be treated as two
independent frames. The interlaced video sequence can switch between frame
structure and field structure on a picture-by-pictures basics.
Alternate Zigzag Scanning
In interlaced scanning, adjacent scan lines come from different fields, and
vertical resolution is reduced when there is motion in the scene. Therefore
the frequency distribution of the DCT coe
cients will have different pattern
as is the case of non-interlaced video. A new zigzag scanning is designed to
optimize the coding e
ciency of the interlaced video, so as to make sure that
the lower frequency coe
cients will be scanned first.
Other Extensions of MPEG-2
MPEG-2 provides error concealment tools for the decoder in the case of packet
loss during transmission. There are a number of ways to conceal errors due
to packet loss, such as spatial neighborhood block averaging, temporal neigh-
borhood block averaging, and temporal motion vectors averaging.
Another extension of MPEG-2 is the use of variable bit rate coding (VBR),
which allows encoder to adjust the bit rate or quality level based on the spa-
tial and temporal complexity of each video segment. More complicated video
segments require more bits to code, and less complicated video segments can
be coded using a lower bit rate. Variable bit rate coding is an important ad-
vancement in video/audio encoding technology, which has found good success
in Super Video CD.
