Sony DTV-01, Manual

Page: 9 / 132

12
Data compression is accomplished by elimination of the high frequency
components designated by the zeros.
10 5 1 0 0 0 0 0
5 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
8 X 8 Pixel Block
LFreq HFref
HFref
Spacial Redundancy
Within a Frame, there are many redundant pixels. An example would
be a blue sky. This type of redundancy within the Horizotal and Vertical
plane of a frame is called “Spacial Redundancy”.
One pixel could be stored with information to repeat for the remaining
pixels. This would eliminate the need to store every pixel in the frame.
Frames making use of this technique are called “Intra-Frames”.
Temporal Redundancy
Within Video scenes, there are many redundant frames. An example
would be an anchor person reporting the news. With the exception of
lip movement, the other portions of the frame remain unchanged over
time. This type of redundancy over time is called “Temporal Redun-
dancy”.
The first frame could be stored as the reference or non changing por-
tions while remaining frames carry the lip motion information. This would
eliminate the need to store several full frames.
Frames using Temporal redundancy which predict information based on
preceding frames are called “Inter-Frames” or Motion predicted Images.
Predictive & Motion Encoding
The next step in the MPEG process is termed “Predictive & Motion
Encoding” and it takes advantage of both Spacial and Temporal redun-
dancy to achieve compression.
I-Pictures
To begin the process of using Spacial and Temporal redundancy tech-
niques in compression, we need a reference or a start frame which does
not depend on previous or preceding frames.
This start frame would make use of Spacial redundancy within itself and
is termed an I picture.
I pictures are Intra-Frames and have zero dependency on previous or
preceding frames. They do however provide information to preceding
frames.
The other pictures types used by MPEG are called P-Pictures (Predic-
tive) and B-Pictures (Bi-Directional).
I Pictures carry the most amount of data content. They are 3 times the
size of a P-Picture and 5 to 6 times the size of a B-Picture.