background image
ISO/IEC 10918-1 : 1993(E)
The procedure for encoding a MCU (see Figure E.5) repetitively invokes the procedure for coding a data unit. For
DCT-based encoders the data unit is an 8
8 block of samples.
Only a portion of each 8
8 block is coded in each scan, the portion being determined by the scan header parameters Ss,
Se, Ah, and Al (see B.2.3). The procedures used to code portions of each 8
8 block are described in this annex. Note,
however, that where these procedures are identical to those used in the sequential DCT-based mode of operation, the
sequential procedures are simply referenced.
G.1.1.1.1
Spectral selection control
In spectral selection the zig-zag sequence of DCT coefficients is segmented into bands. A band is defined in the scan
header by specifying the starting and ending indices in the zig-zag sequence. One band is coded in a given scan of the
progression. DC coefficients are always coded separately from AC coefficients, and only scans which code DC
coefficients may have interleaved blocks from more than one component. All other scans shall have only one component.
With the exception of the first DC scans for the components, the sequence of bands defined in the scans need not follow
the zig-zag ordering. For each component, a first DC scan shall precede any AC scans.
G.1.1.1.2
Successive approximation control
If successive approximation is used, the DCT coefficients are reduced in precision by the point transform (see A.4)
defined in the scan header (see B.2.3). The successive approximation bit position parameter Al specifies the actual point
transform, and the high four bits (Ah) if there are preceding scans for the band contain the value of the point transform
used in those preceding scans. If there are no preceding scans for the band, Ah is zero.
Each scan which follows the first scan for a given band progressively improves the precision of the coefficients by one bit,
until full precision is reached.
G.1.1.2 Coding models for progressive DCT-based encoders
If successive approximation is used, the DCT coefficients are reduced in precision by the point transform (see A.4)
defined in the scan header (see B.2.3). These models also apply to the progressive DCT-based encoders, but with the
following changes.
G.1.1.2.1
Progressive encoding model for DC coefficients
If Al is not zero, the point transform for DC coefficients shall be used to reduce the precision of the DC coefficients. If Ah
is zero, the coefficient values (as modified by the point transform) shall be coded, using the procedure described in Annex
F. If Ah is not zero, the least significant bit of the point transformed DC coefficients shall be coded, using the procedures
described in this annex.
G.1.1.2.2
Progressive encoding model for AC coefficients
If Al is not zero, the point transform for AC coefficients shall be used to reduce the precision of the AC coefficients. If Ah
is zero, the coefficient values (as modified by the point transform) shall be coded using modifications of the procedures
described in Annex F. These modifications are described in this annex. If Ah is not zero, the precision of the coefficients
shall be improved using the procedures described in this annex.
G.1.2
Progressive encoding procedures with Huffman coding
G.1.2.1 Progressive encoding of DC coefficients with Huffman coding
The first scan for a given component shall encode the DC coefficient values using the procedures described in F.1.2.1. If
the successive approximation bit position parameter Al is not zero, the coefficient values shall be reduced in precision by
the point transform described in Annex A before coding.
In subsequent scans using successive approximation the least significant bits are appended to the compressed bit stream
without compression or modification (see G.1.2.3), except for byte stuffing.
G.1.2.2 Progressive encoding of AC coefficients with Huffman coding
In spectral selection and in the first scan of successive approximation for a component, the AC coefficient coding model is
similar to that used by the sequential procedures. However, the Huffman code tables are extended to include coding of
runs of End-Of-Bands (EOBs). See Table G.1.
CCITT Rec. T.81 (1992 E)
121
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186]