Penguin
Recent Changes
Blame: pnmtojbig(1)
EditPageHistoryDiffInfoLikePages
Annotated edit history of pnmtojbig(1) version 1, including all changes. View license author blame.
Rev Author # Line
1 perry 1 PNMTOJBIG
2 !!!PNMTOJBIG
3 NAME
4 SYNOPSIS
5 DESCRIPTION
6 OPTIONS
7 FORMATS
8 STANDARDS
9 AUTHOR
10 SEE ALSO
11 LICENSE
12 ----
13 !!NAME
14
15
16 pnmtojbig - PNM to JBIG file converter
17 !!SYNOPSIS
18
19
20 __pnmtojbg__ [[ ''options'' ] [[ ''input-file'' | - [[
21 ''output-file'' ]]
22 !!DESCRIPTION
23
24
25 Reads in a PBM or PGM format image from a file or standard
26 input, compresses it, and outputs the image as a ''JBIG''
27 bi-level image entity (BIE) file.
28
29
30 ''JBIG'' is a highly effective lossless compression
31 algorithm for bi-level images (one bit per pixel), which is
32 particularly suitable for scanned document
33 pages.
34
35
36 A ''JBIG'' encoded image can be stored in several
37 resolutions (progressive mode). These resolution layers can
38 be stored all in one single BIE or they can be stored in
39 several separate BIE files. All resolution layers except the
40 lowest one are stored merely as differences to the next
41 lower resolution layer, because this requires less space
42 than encoding the full image completely every time. Each
43 resolution layer has twice the number of horizontal and
44 vertical pixels than the next lower layer. ''JBIG'' files
45 can also store several bits per pixel as separate bitmap
46 planes, and __pnmtojbig__ can read a PGM file and
47 transform it into a multi-bitplane BIE.
48 !!OPTIONS
49
50
51 __-__ A single hyphen instead of an input file name will
52 cause __pnmtojbg__ to read the data from standard input
53 instead from a file.
54
55
56 __-q__ Encode the image in one single resolution layer
57 (sequential mode). This is usually the most efficient
58 compression method. By default, the number of resolution
59 layers is chosen automatically such that the lowest layer
60 image is not larger than 640 480 pixels.
61
62
63 __-x__ ''number''
64
65
66 Specify the maximal horizontal size of the lowest resolution
67 layer. The default is 640 pixels.
68
69
70 __-y__ ''number''
71
72
73 Specify the maximal vertical size of the lowest resolution
74 layer. The default is 480 pixels.
75
76
77 __-l__ ''number''
78
79
80 Select the lowest resolution layer that will be written to
81 the BIE. It is possible to store the various resolution
82 layers of a ''JBIG'' image in progressive mode into
83 different BIEs. Options __-l__ and __-h__ allow to
84 select the resolution-layer interval that will appear in the
85 created BIE. The lowest resolution layer has number 0 and
86 this is also the default value. By default all layers will
87 be written.
88
89
90 __-h__ ''number''
91
92
93 Select the highest resolution layer that will be written to
94 the BIE. By default all layers will be written. See also
95 option __-l.__
96
97
98 __-b__ Use binary values instead of Gray code words in
99 order to encode pixel values in multiple bitplanes. This
100 option has only an effect if the input is a PGM file and if
101 more than one bitplane is produced. Note that the decoder
102 has to make the same selection but cannot determine from the
103 BIE, whether Gray or binary code words were used by the
104 encoder.
105
106
107 __-d__ ''number''
108
109
110 Specify the total number of differential resolution layers
111 into which the input image will be split in addition to the
112 lowest layer. Each additional layer reduces the size of
113 layer 0 by 50 %. This option overrides options __-x__ and
114 __-y__ which are usually a more comfortable way of
115 selecting the number of resolution layers.
116
117
118 __-s__ ''number''
119
120
121 The ''JBIG'' algorithm splits each image into a number of
122 horizontal stripes. This option specifies that each stripe
123 shall have ''number'' lines in layer 0. The default value
124 is selected so that approximately 35 stripes will be used
125 for the whole image.
126
127
128 __-m__ ''number''
129
130
131 Select the maximum horizontal offset of the adaptive
132 template pixel. The ''JBIG'' encoder uses a number of
133 neighbour pixels in order to get statistical a priori
134 knowledge of the probability, whether the next pixel will be
135 black or white. One single pixel out of this template of
136 context neighbor pixels can be moved around. Especially for
137 dithered images it can be a significant advantage to have
138 one neighbor pixel which has a distance large enough to
139 cover the period of a dither function. By default, the
140 adaptive template pixel can be moved up to 8 pixels away.
141 This encoder supports up to 23 pixels, however as decoders
142 are only required to support at least a distance of 16
143 pixels by the standard, no higher value than 16 for
144 ''number'' is recommended in order to maintain
145 interoperability with other ''JBIG'' implementations. The
146 maximal vertical offset of the adaptive template pixel is
147 always zero.
148
149
150 __-t__ ''number''
151
152
153 Encode only the specified number of most significant bit
154 planes. This option allows to reduce the depth of an input
155 PGM file if not all bits per pixel are needed in the
156 output.
157
158
159 __-o__ ''number''
160
161
162 ''JBIG'' separates an image into several horizontal
163 stripes, resolution layers and planes, were each plane
164 contains one bit per pixel. One single stripe in one plane
165 and layer is encoded as a data unit called stripe data
166 entity (SDE) inside the BIE. There are 12 different possible
167 orders in which the SDEs can be stored inside the BIE and
168 ''number'' selects which one shall be used. The order of
169 the SDEs is only relevant for applications that want to
170 decode a ''JBIG'' file which has not yet completely
171 arrived from e.g. a slow network connection. For instance
172 some applications prefer that the outermost of the three
173 loops (stripes, layers, planes) is over all layers so that
174 all data of the lowest resolution layer is transmitted
175 first.
176 The following values for ''number'' select these loop
177 arrangements for writing the SDEs (outermost loop
178 first):
179
180
181 0 planes, layers, stripes
182 2 layers, planes, stripes
183 3 layers, stripes, planes
184 4 stripes, planes, layers
185 5 planes, stripes, layers
186 6 stripes, layers, planes
187
188
189 All loops count starting with zero, however by adding 8 to
190 the above order code, the layer loop can be reversed so that
191 it counts down to zero and then higher resolution layers
192 will be stored before lower layers. Default order is 3 which
193 writes at first all planes of the first stripe and then
194 completes layer 0 before continuing with the next layer and
195 so on.
196
197
198 __-p__ ''number''
199
200
201 This option allows to activate or deactivate various
202 optional algorithms defined in the ''JBIG'' standard.
203 Just add the numbers of the following options which you want
204 to activate in order to get the ''number''
205 value:
206
207
208 4 deterministic prediction (DPON)
209 8 layer 0 typical prediction (TPBON)
210 16 diff. layer typ. pred. (TPDON)
211 64 layer 0 two-line template (LRLTWO)
212
213
214 Except for special applications (like communication with
215 ''JBIG'' subset implementations) and for debugging
216 purposes you will normally not want to change anything here.
217 The default is 28, which provides the best compression
218 result.
219
220
221 __-c__ The adaptive template pixel movement is determined
222 as suggested in annex C of the standard. By default the
223 template change takes place directly in the next line which
224 is most effective. However a few conformance test examples
225 in the standard require the adaptive template change to be
226 delayed until the first line of the next stripe. This option
227 selects this special behavior, which is normally not
228 required except in order to pass some conformance test
229 suite.
230
231
232 __-v__ After the BIE has been created, a few technical
233 details of the created file will be listed (verbose
234 mode).
235 !!FORMATS
236
237
238 Most of the format __pnmtojbig__ creates is defined by
239 the ''JBIG'' standard.
240
241
242 The standard, however, does not specify which values in the
243 BIE mean white and which mean black. It contains a
244 recommendation that for a single plane image zero mean
245 background and one mean foreground, but the Netpbm formats
246 have no concept of foreground and background. And the
247 standard says nothing about values for multiple plane
248 BIEs.
249
250
251 __pnmtojbig__ follows Markus Kuhn's implementation of the
252 standard in the __pbmtojbg__ program that comes with his
253 ''JBIG'' library: If the BIE is a single plane BIE, zero
254 means white and one means black. If it is a multiple plane
255 BIE, zero means black and the maximal value is
256 white.
257 !!STANDARDS
258
259
260 This program implements the ''JBIG'' image coding
261 algorithm as specified in ISO/IEC 11544:1993 and ITU-T
262 T.82(1993).
263 !!AUTHOR
264
265
266 __pnmtojbig__ is based on the ''JBIG'' library by
267 Markus Kuhn, part of his __JBIG-KIT__ package. The
268 __pbmtojbg__ program is part of the ''JBIG-KIT''
269 package. The most recent version of that library and tools
270 set is freely available on the Internet from anonymous ftp
271 server ftp.informatik.uni-erlangen.de in directory
272 pub/doc/ISO/JBIG/.
273
274
275 __pnmtojbig__ is part of the Netpbm package of graphics
276 tools.
277 !!SEE ALSO
278
279
280 pbm(5),pgm(5),jbigtopbm(1)
281 !!LICENSE
282
283
284 If you use __pnmtojbig__, you are using various patents,
285 particularly on its arithmetic encoding method, and in all
286 probability you do not have a license from the patent owners
287 to do so.
288 ----
This page is a man page (or other imported legacy content). We are unable to automatically determine the license status of this page.