Databases Reference
In-Depth Information
T A B L E 19 . 3
After filtering the rows.
110
165
140
112
108
159
135
114
110
188
187
112
112
109
111
116
This filter is either switched on or off for each macroblock. The conditions for turning the
filter on or off are not specified by the recommendations.
19.5.3 The Transform
The transformoperation is performedwith aDCTon an 8
8 block of pixels or pixel differences.
If the motion compensation operation does not provide a close match, then the transform
operation is performed on an 8
×
8 block of pixels. If the transform operation is performed
on a block level, either a block or the difference between the block and its predicted value
is quantized and transmitted to the receiver. The receiver performs the inverse operations to
reconstruct the image. The receiver operation is also simulated at the transmitter, where the
reconstructed images are obtained and stored in a frame store. The encoder is said to be in
intra
mode if it operates directly on the input image without the use of motion compensation.
Otherwise, it is said to be in
inter
mode.
×
19.5.4 Quantization and Coding
Depending on how good or poor the prediction is, we can get a wide variation in the char-
acteristics of the coefficients that are to be quantized. In the case of an intra block, the DC
coefficients will take on much larger values than the other coefficients. Where there is little
motion from frame to frame, the difference between the block being encoded and the prediction
will be small, leading to small values for the coefficients.
In order to deal with this wide variation, we need a quantization strategy that can be
rapidly adapted to the current situation. The H.261 algorithm does this by switching between
32 different quantizers, possibly fromonemacroblock to the next. One quantizer is reserved for
the intra DC coefficient, while the remaining 31 quantizers are used for the other coefficients.
T A B L E 19 . 4
Final block.
110
165
140
112
108
167
148
113
110
161
154
113
112
109
111
116
