CCITT GROUP 3 Modified Huffman Run Length ENCODEING
This compression method is originally a fax machine compression coding method set by CCITT, which is FACSIMile Compatible CCITT Group3. However, this compression method reduces the special code on some fax machines. This compression method can only compress the handling black and white image data. This compression method is to use a BITS number unfixed encoding to replace a string white point or black spot that continuously appears in black and white image data.
All encoded BITS values are from the following three forms:
Terminating Codes
Make-Up Codes
Additional make-up cots
(1) Terminating Codes represents the encoding of the 0-63 consecutive white point or black. Within this range, there is no corresponding encoding without a white point or black.
(2) Make-Up Codes represents the 64-1728 encoding of the continuous white or black spots. Within this range, there is a corresponding encoding every 64 points or black dots.
(3) Additional make-up Codes represents the 1792-2560 encoding of continuous white or black spots. Within this range, a corresponding encoding is shared every 64 points or black spots.
Table 1a. Terminating White Codes
Code LNG RUN
---------------------------
00110101 8 0
000111 6 1
0111 4 2
1000 4 3
1011 4 4
1100 4 5
1110 4 6
1111 4 7
10011 5 8
10100 5 9
00111 5 10
01000 5 11
001000 6 12
000011 6 13
110100 6 14
110101 6 15
101010 6 16
101011 6 17
0100111 7 18
0001100 7 19
0001000 7 20
0010111 7 21
0000011 7 22
0000100 7 23
0101000 7 24
0101011 7 25
0010011 7 26
0100100 7 27
0011000 7 28
00000010 8 29
00000011 8 30
00011080 8 31
00011011 8 32
00010010 8 33
00010011 8 34
00010100 8 35
00010101 8 3600010110 8 37
00010111 8 38
00101000 8 39
00101001 8 40
00101010 8 41
00101011 8 42
00101100 8 43
00101101 8 44
00000100 8 45
00000101 8 46
00001010 8 47
00001011 8 48
01010010 8 49
01010011 8 50
01010100 8 51
01010101 8 52
00100100 8 53
00100101 8 54
01011000 8 55
01011001 8 56
01011010 8 57
01011011 8 58
01001010 8 59
01001011 8 60
00110010 8 61
00110011 8 62
00110100 8 63
Table 1b. Make Up White Codes
Code LNG RUN
---------------------------
11011 5 64
10010 5 128
010111 6 192
0110111 7 256
00110110 8 320
00110111 8 384
01100100 8 448
01100101 8 512
01101000 8 576
01100111 8 640
011001100 9 704
011001101 9 768
011010010 9 832
011010011 9 896
011010100 9 960
011010101 9 1024
011010110 9 1088
011010111 9 1152
011011000 9 1216
011011001 9 1280
011011010 9 1344
011011011 9 1408
010011000 9 1472
010011001 9 1536
010011010 9 1600
011000 6 1664
010011011 9 1728
Table 2a. Terminating Black Codes
Code LNG RUN
--------------------------- 0000110111 10 0
010 3 1
11 2 2
10 2 3
011 3 4
0011 4 5
0010 4 6
00011 5 7
000101 6 8
000100 6 9
0000100 7 10
0000101 7 11
0000111 7 12
00000100 8 13
00000111 8 14
000011000 9 15
0000010111 10 16
0000011000 10 17
0000001000 10 18
00001100111 11 19
00001101000 11 20
00001101100 11 21
00000110111 11 22
00000101000 11 23
00000010111 11 24
00000011000 11 25
000011001010 12 26
000011001011 12 27
000011001100 12 28
000011001101 12 29
000001101000 12 30
000001101001 12 31
000001101010 12 32
000001101011 12 33
000011010010 12 34
000011010011 12 35
000011010100 12 36
000011010101 12 37
000011010110 12 38
000011010111 12 39
000001101100 12 40
000001101101 12 41
000011011010 12 42
000011011011 12 43
000001010100 12 44
000001010101 12 45
000001010110 12 46
000001010111 12 47
000001100100 12 48
000001100101 12 49
000001010010 12 50
000001010011 12 51
000000100100 12 52
000000110111 12 53
000000111000 12 54
000000100111 12 55
000000101000 12 56
000001011000 12 57
000001011001 12 58
000000101011 12 59
000000101100 12 60
000001011010 12 61000001100110 12 62
000001100111 12 63
Table 2b. Make Up Black Codes
Code LNG RUN
---------------------------
0000001111 10 64
000011001000 12 128
000011001001 12 192
000001011011 12 256
000000110011 12 320
000000110100 12 384
000000110101 12 448
0000001101100 13 512
0000001101101 13 576
0000001001010 13 640
0000001001011 13 704
0000001001100 13 768
0000001001101 13 832
0000001110010 13 896
0000001110011 13 960
0000001110100 13 1024
0000001110101 13 1088
0000001110110 13 1152
0000001110111 13 1216
0000001010010 13 1280
0000001010011 13 1344
0000001010100 13 1408
0000001010101 13 1472
0000001011010 13 1536
0000001011011 13 1600
0000001100100 13 1664
0000001100101 13 1728
Table 3. Extended Make Up Codes (Black and White)
Code LNG RUN
---------------------------
00000001000 11 1792
00000001100 11 1856
00000001101 11 1920
000000010010 12 1984
000000010011 12 2048
000000010100 12 2112
000000010101 12 2176
000000010110 12 2240
000000010111 12 2304
000000011100 12 2368
000000011101 12 2432
000000011110 12 2496
0000000111112 12 2560
Anyone who wants to use this compression method to handle black and white image data must now save these three forms, so that the program can refer to these forms to identify the correct encoding, prepared to compress data.
During the process of compressed compressed data, the first step is to store a white point encoding first. If the first bit value of the image data is a black point, the continuous white point encoding (00110101) of the number of points is 0, and then the coding of the black point is started. This is a special provision of this compression method: the first coding value of each line must be encoded in a continuous white point. The next question is how to replace image data in encoding within the above table? Below, the first two lines of one image data are as follows: The first column 15 hollow, 625 white points, 3360 black dots
Postal code: 00110100000000110010 000000000111111
0 white points 15 black spots 576 white points 49 white points 2560 black spots
011001101 000001101010 000000
768 black spots 32 black spots 6 bits
The second list 18 white points, 1486 black spots, 2496 white points
coding:
0100111 0000000010101011 0000001111 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000101011010101
18 white points 1472 black spots 14 black spots 2496 white dots 0 white dots
From the coded data listed above, you can summarize the following four coding principles:
(1) The first encoding of each line is compressed, and it must be encoded in a continuous white point. Even if the first point in each line is a black point, you must first use 00110101 (0 white points) as the first encoding of the compressed data, then start the first encoding of the data, and then store the true representative graphic black. Point code;
(2) The total number of BITS of each line is required to be 8 multiple. If the total number of BITS is not 8 times, it is necessary to compress the end of the data per line, plus a bits having a value of 0, is a multiple of the BITS total number of 8. For example, the total number of BITS of the first row of the above instance is 66, and the six bits are lacking, in order to become a multiple of 8, the end of the first column compressed data, adding six values 0, which enables the total number of BITS to reach 72, enough to be 8 times. As for the total number of BITS of the second line of compressed data, it is exactly the multiple of 8, so there is no need to add some extra BITS.
(3) There are three types of encoding.
(A) Number of points does not exceed 63, and you can refer to the Terminating Codes form to find the corresponding encoding. This coding is only one, which instead of a string of the same points.
(B) If the number is between 64-1728, it is found according to Terminating Codes and Make-Up Codes forms. If the number of points is not equal to the value of the Make-Up Codes table, it is necessary to use two encodings to replace a string of the same points.
(C) The number of points is more than 1792, and you must refer to three forms at the same time to find the code you want. At this time, the encoding of a string of the same value may require three. If you have a lot of points than 2560, you need to encode, just three.
(4) The encoding in the Additional Mak-Up Codes table can represent a continuous white point or black point, so the encoding of this table is to bring the encoding of two tables, in order to let the decompression program know what exactly The read code is a representative white point or a black point. For example, the second row of the above example has a string of dotted points of 2496, just the same value as the 12th point in the Additional MAK-UP CODES table. Therefore, after 000000011110 (2496 consecutive white points or black spots), plus 00110101 (0 white points), indicating that 000000011110 represents 2496 continuous white points. The above is taken from the "Mystery of Exploring Image Files" Li Zhenhui and RFC804
my experience:
When a continuous white point or black spot is only used as Make-Up Codes or
Additional make-up CODES When these two forms, must be added 00110101 (0 white points) or 0000110111 (0 black points)
E.g:
1792 white points
Postal code: 000000001010101010101011010
1792 white points 0 white points 5 bits
128 black spots
Code: 00110 000011011100
0 white dots 128 black dots 0 black dots 2 bits
2624 white points
Code 0000000111111010101010101010101010101010101010101010101010101110101010
2560 white point 64 white dots 0 white dots 7 bits
2625 white points
Code 0000000111111 11011 000111 0
2560 white points 64 white points 1 white point 1 bits
