ISO/IEC 10918-1 : 1993(E)
Control procedure for encoding a minimum coded unit (MCU)
The minimum coded unit is defined in A.2. Within a given MCU the data units are coded in the order in which they occur
in the MCU. The control procedure for encoding a MCU is shown in Figure E.5.
N = 0
N = N + 1
Encode data unit
N = Nb
Figure E.5 Control procedure for encoding a minimum coded unit (MCU)
Figure E.5 [D65], = 12 cm = 469.%
In Figure E.5, Nb refers to the number of data units in the MCU. The order in which data units occur in the MCU is
defined in A.2. The data unit is an 8
8 block for DCT-based processes, and a single sample for lossless processes.
The procedures for encoding a data unit are specified in Annexes F, G, and H.
Decoder control procedures
Control procedure for decoding compressed image data
Figure E.6 shows the decoding process control for compressed image data.
Decoding control centers around identification of various markers. The first marker must be the SOI (Start Of Image)
marker. The "Decoder_setup" procedure resets the restart interval (Ri = 0) and, if the decoder has arithmetic decoding
capabilities, sets the conditioning tables for the arithmetic coding to their default values. (See F.22.214.171.124.4 and F.126.96.36.199.1.)
The next marker is normally a SOF
(Start Of Frame) marker; if this is not found, one of the marker segments listed in
Table E.1 has been received.
CCITT Rec. T.81 (1992 E)