ISO/IEC 10918-1 : 1993(E)
c
b
x
a
TISO1680-93/d106
Figure H.1 Relationship between sample and prediction samples
Figure H.1 [D106] = 4.5 cm = 176 %
Define Px to be the prediction and Ra, Rb, and Rc to be the reconstructed samples immediately to the left, immediately
above, and diagonally to the left of the current sample. The allowed predictors, one of which is selected in the scan
header, are listed in Table H.1.
Table H.1 Predictors for lossless coding
Selection-value
Prediction
0
No prediction (See Annex J)
1
Px
=
Ra
2
Px
=
Rb
3
Px
=
Rc
4
Px
=
Ra
+
Rb Rc
5
Px
=
Ra
+
((Rb Rc)/2)
a)
6
Px
=
Rb
+
((Ra Rc)/2)
a)
7
Px
=
(Ra
+
Rb)/2
a)
Shift right arithmetic operation
Selection-value 0 shall only be used for differential coding in the hierarchical mode of operation. Selections 1, 2 and 3 are
one-dimensional predictors and selections 4, 5, 6, and 7 are two-dimensional predictors.
The one-dimensional horizontal predictor (prediction sample Ra) is used for the first line of samples at the start of the scan
and at the beginning of each restart interval. The selected predictor is used for all other lines. The sample from the line
above (prediction sample Rb) is used at the start of each line, except for the first line. At the beginning of the first line and
at the beginning of each restart interval the prediction value of 2
P 1
is used, where P is the input precision.
If the point transformation parameter (see A.4) is non-zero, the prediction value at the beginning of the first lines and the
beginning of each restart interval is 2
P Pt 1
, where Pt is the value of the point transformation parameter.
Each prediction is calculated with full integer arithmetic precision, and without clamping of either underflow or overflow
beyond the input precision bounds. For example, if Ra and Rb are both 16-bit integers, the sum is a 17-bit integer. After
dividing the sum by 2 (predictor 7), the prediction is a 16-bit integer.
CCITT Rec. T.81 (1992 E)
133