1
USAGE instructions for the Independent JPEG Group's JPEG software
2
=================================================================
4
This file describes usage of the JPEG conversion programs cjpeg and djpeg,
5
as well as the utility programs jpegtran, rdjpgcom and wrjpgcom. (See
6
the other documentation files if you wish to use the JPEG library within
9
If you are on a Unix machine you may prefer to read the Unix-style manual
10
pages in files cjpeg.1, djpeg.1, jpegtran.1, rdjpgcom.1, wrjpgcom.1.
15
These programs implement JPEG image compression and decompression. JPEG
16
(pronounced "jay-peg") is a standardized compression method for full-color
17
and gray-scale images. JPEG is designed to handle "real-world" scenes,
18
for example scanned photographs. Cartoons, line drawings, and other
19
non-realistic images are not JPEG's strong suit; on that sort of material
20
you may get poor image quality and/or little compression.
22
JPEG is lossy, meaning that the output image is not necessarily identical to
23
the input image. Hence you should not use JPEG if you have to have identical
24
output bits. However, on typical real-world images, very good compression
25
levels can be obtained with no visible change, and amazingly high compression
26
is possible if you can tolerate a low-quality image. You can trade off image
27
quality against file size by adjusting the compressor's "quality" setting.
32
We provide two programs, cjpeg to compress an image file into JPEG format,
33
and djpeg to decompress a JPEG file back into a conventional image format.
35
On Unix-like systems, you say:
36
cjpeg [switches] [imagefile] >jpegfile
38
djpeg [switches] [jpegfile] >imagefile
39
The programs read the specified input file, or standard input if none is
40
named. They always write to standard output (with trace/error messages to
41
standard error). These conventions are handy for piping images between
44
On most non-Unix systems, you say:
45
cjpeg [switches] imagefile jpegfile
47
djpeg [switches] jpegfile imagefile
48
i.e., both the input and output files are named on the command line. This
49
style is a little more foolproof, and it loses no functionality if you don't
50
have pipes. (You can get this style on Unix too, if you prefer, by defining
51
TWO_FILE_COMMANDLINE when you compile the programs; see install.doc.)
54
cjpeg [switches] -outfile jpegfile imagefile
56
djpeg [switches] -outfile imagefile jpegfile
57
This syntax works on all systems, so it is useful for scripts.
59
The currently supported image file formats are: PPM (PBMPLUS color format),
60
PGM (PBMPLUS gray-scale format), BMP, Targa, and RLE (Utah Raster Toolkit
61
format). (RLE is supported only if the URT library is available.)
62
cjpeg recognizes the input image format automatically, with the exception
63
of some Targa-format files. You have to tell djpeg which format to generate.
65
JPEG files are in the defacto standard JFIF file format. There are other,
66
less widely used JPEG-based file formats, but we don't support them.
68
All switch names may be abbreviated; for example, -grayscale may be written
69
-gray or -gr. Most of the "basic" switches can be abbreviated to as little as
70
one letter. Upper and lower case are equivalent (-BMP is the same as -bmp).
71
British spellings are also accepted (e.g., -greyscale), though for brevity
72
these are not mentioned below.
77
The basic command line switches for cjpeg are:
79
-quality N Scale quantization tables to adjust image quality.
80
Quality is 0 (worst) to 100 (best); default is 75.
81
(See below for more info.)
83
-grayscale Create monochrome JPEG file from color input.
84
Be sure to use this switch when compressing a grayscale
85
BMP file, because cjpeg isn't bright enough to notice
86
whether a BMP file uses only shades of gray. By
87
saying -grayscale, you'll get a smaller JPEG file that
88
takes less time to process.
90
-optimize Perform optimization of entropy encoding parameters.
91
Without this, default encoding parameters are used.
92
-optimize usually makes the JPEG file a little smaller,
93
but cjpeg runs somewhat slower and needs much more
94
memory. Image quality and speed of decompression are
95
unaffected by -optimize.
97
-progressive Create progressive JPEG file (see below).
99
-targa Input file is Targa format. Targa files that contain
100
an "identification" field will not be automatically
101
recognized by cjpeg; for such files you must specify
102
-targa to make cjpeg treat the input as Targa format.
103
For most Targa files, you won't need this switch.
105
The -quality switch lets you trade off compressed file size against quality of
106
the reconstructed image: the higher the quality setting, the larger the JPEG
107
file, and the closer the output image will be to the original input. Normally
108
you want to use the lowest quality setting (smallest file) that decompresses
109
into something visually indistinguishable from the original image. For this
110
purpose the quality setting should be between 50 and 95; the default of 75 is
111
often about right. If you see defects at -quality 75, then go up 5 or 10
112
counts at a time until you are happy with the output image. (The optimal
113
setting will vary from one image to another.)
115
-quality 100 will generate a quantization table of all 1's, minimizing loss
116
in the quantization step (but there is still information loss in subsampling,
117
as well as roundoff error). This setting is mainly of interest for
118
experimental purposes. Quality values above about 95 are NOT recommended for
119
normal use; the compressed file size goes up dramatically for hardly any gain
120
in output image quality.
122
In the other direction, quality values below 50 will produce very small files
123
of low image quality. Settings around 5 to 10 might be useful in preparing an
124
index of a large image library, for example. Try -quality 2 (or so) for some
125
amusing Cubist effects. (Note: quality values below about 25 generate 2-byte
126
quantization tables, which are considered optional in the JPEG standard.
127
cjpeg emits a warning message when you give such a quality value, because some
128
other JPEG programs may be unable to decode the resulting file. Use -baseline
129
if you need to ensure compatibility at low quality values.)
131
The -progressive switch creates a "progressive JPEG" file. In this type of
132
JPEG file, the data is stored in multiple scans of increasing quality. If the
133
file is being transmitted over a slow communications link, the decoder can use
134
the first scan to display a low-quality image very quickly, and can then
135
improve the display with each subsequent scan. The final image is exactly
136
equivalent to a standard JPEG file of the same quality setting, and the total
137
file size is about the same --- often a little smaller. CAUTION: progressive
138
JPEG is not yet widely implemented, so many decoders will be unable to view a
139
progressive JPEG file at all.
141
Switches for advanced users:
143
-dct int Use integer DCT method (default).
144
-dct fast Use fast integer DCT (less accurate).
145
-dct float Use floating-point DCT method.
146
The float method is very slightly more accurate than
147
the int method, but is much slower unless your machine
148
has very fast floating-point hardware. Also note that
149
results of the floating-point method may vary slightly
150
across machines, while the integer methods should give
151
the same results everywhere. The fast integer method
152
is much less accurate than the other two.
154
-restart N Emit a JPEG restart marker every N MCU rows, or every
155
N MCU blocks if "B" is attached to the number.
156
-restart 0 (the default) means no restart markers.
158
-smooth N Smooth the input image to eliminate dithering noise.
159
N, ranging from 1 to 100, indicates the strength of
160
smoothing. 0 (the default) means no smoothing.
162
-maxmemory N Set limit for amount of memory to use in processing
163
large images. Value is in thousands of bytes, or
164
millions of bytes if "M" is attached to the number.
165
For example, -max 4m selects 4000000 bytes. If more
166
space is needed, temporary files will be used.
168
-verbose Enable debug printout. More -v's give more printout.
169
or -debug Also, version information is printed at startup.
171
The -restart option inserts extra markers that allow a JPEG decoder to
172
resynchronize after a transmission error. Without restart markers, any damage
173
to a compressed file will usually ruin the image from the point of the error
174
to the end of the image; with restart markers, the damage is usually confined
175
to the portion of the image up to the next restart marker. Of course, the
176
restart markers occupy extra space. We recommend -restart 1 for images that
177
will be transmitted across unreliable networks such as Usenet.
179
The -smooth option filters the input to eliminate fine-scale noise. This is
180
often useful when converting dithered images to JPEG: a moderate smoothing
181
factor of 10 to 50 gets rid of dithering patterns in the input file, resulting
182
in a smaller JPEG file and a better-looking image. Too large a smoothing
183
factor will visibly blur the image, however.
185
Switches for wizards:
187
-baseline Force baseline-compatible quantization tables to be
188
generated. This clamps quantization values to 8 bits
189
even at low quality settings. (This switch is poorly
190
named, since it does not ensure that the output is
191
actually baseline JPEG. For example, you can use
192
-baseline and -progressive together.)
194
-qtables file Use the quantization tables given in the specified
197
-qslots N[,...] Select which quantization table to use for each color
200
-sample HxV[,...] Set JPEG sampling factors for each color component.
202
-scans file Use the scan script given in the specified text file.
204
The "wizard" switches are intended for experimentation with JPEG. If you
205
don't know what you are doing, DON'T USE THEM. These switches are documented
206
further in the file wizard.doc.
211
The basic command line switches for djpeg are:
213
-colors N Reduce image to at most N colors. This reduces the
214
or -quantize N number of colors used in the output image, so that it
215
can be displayed on a colormapped display or stored in
216
a colormapped file format. For example, if you have
217
an 8-bit display, you'd need to reduce to 256 or fewer
218
colors. (-colors is the recommended name, -quantize
219
is provided only for backwards compatibility.)
221
-fast Select recommended processing options for fast, low
222
quality output. (The default options are chosen for
223
highest quality output.) Currently, this is equivalent
224
to "-dct fast -nosmooth -onepass -dither ordered".
226
-grayscale Force gray-scale output even if JPEG file is color.
227
Useful for viewing on monochrome displays; also,
228
djpeg runs noticeably faster in this mode.
230
-scale M/N Scale the output image by a factor M/N. Currently
231
the scale factor must be 1/1, 1/2, 1/4, or 1/8.
232
Scaling is handy if the image is larger than your
233
screen; also, djpeg runs much faster when scaling
236
-bmp Select BMP output format (Windows flavor). 8-bit
237
colormapped format is emitted if -colors or -grayscale
238
is specified, or if the JPEG file is gray-scale;
239
otherwise, 24-bit full-color format is emitted.
241
-gif Select GIF output format. Since GIF does not support
242
more than 256 colors, -colors 256 is assumed (unless
243
you specify a smaller number of colors). If you
244
specify -fast, the default number of colors is 216.
246
-os2 Select BMP output format (OS/2 1.x flavor). 8-bit
247
colormapped format is emitted if -colors or -grayscale
248
is specified, or if the JPEG file is gray-scale;
249
otherwise, 24-bit full-color format is emitted.
251
-pnm Select PBMPLUS (PPM/PGM) output format (this is the
252
default format). PGM is emitted if the JPEG file is
253
gray-scale or if -grayscale is specified; otherwise
256
-rle Select RLE output format. (Requires URT library.)
258
-targa Select Targa output format. Gray-scale format is
259
emitted if the JPEG file is gray-scale or if
260
-grayscale is specified; otherwise, colormapped format
261
is emitted if -colors is specified; otherwise, 24-bit
262
full-color format is emitted.
264
Switches for advanced users:
266
-dct int Use integer DCT method (default).
267
-dct fast Use fast integer DCT (less accurate).
268
-dct float Use floating-point DCT method.
269
The float method is very slightly more accurate than
270
the int method, but is much slower unless your machine
271
has very fast floating-point hardware. Also note that
272
results of the floating-point method may vary slightly
273
across machines, while the integer methods should give
274
the same results everywhere. The fast integer method
275
is much less accurate than the other two.
277
-dither fs Use Floyd-Steinberg dithering in color quantization.
278
-dither ordered Use ordered dithering in color quantization.
279
-dither none Do not use dithering in color quantization.
280
By default, Floyd-Steinberg dithering is applied when
281
quantizing colors; this is slow but usually produces
282
the best results. Ordered dither is a compromise
283
between speed and quality; no dithering is fast but
284
usually looks awful. Note that these switches have
285
no effect unless color quantization is being done.
286
Ordered dither is only available in -onepass mode.
288
-map FILE Quantize to the colors used in the specified image
289
file. This is useful for producing multiple files
290
with identical color maps, or for forcing a predefined
291
set of colors to be used. The FILE must be a GIF
292
or PPM file. This option overrides -colors and
295
-nosmooth Use a faster, lower-quality upsampling routine.
297
-onepass Use one-pass instead of two-pass color quantization.
298
The one-pass method is faster and needs less memory,
299
but it produces a lower-quality image. -onepass is
300
ignored unless you also say -colors N. Also,
301
the one-pass method is always used for gray-scale
302
output (the two-pass method is no improvement then).
304
-maxmemory N Set limit for amount of memory to use in processing
305
large images. Value is in thousands of bytes, or
306
millions of bytes if "M" is attached to the number.
307
For example, -max 4m selects 4000000 bytes. If more
308
space is needed, temporary files will be used.
310
-verbose Enable debug printout. More -v's give more printout.
311
or -debug Also, version information is printed at startup.
316
Color GIF files are not the ideal input for JPEG; JPEG is really intended for
317
compressing full-color (24-bit) images. In particular, don't try to convert
318
cartoons, line drawings, and other images that have only a few distinct
319
colors. GIF works great on these, JPEG does not. If you want to convert a
320
GIF to JPEG, you should experiment with cjpeg's -quality and -smooth options
321
to get a satisfactory conversion. -smooth 10 or so is often helpful.
323
Avoid running an image through a series of JPEG compression/decompression
324
cycles. Image quality loss will accumulate; after ten or so cycles the image
325
may be noticeably worse than it was after one cycle. It's best to use a
326
lossless format while manipulating an image, then convert to JPEG format when
327
you are ready to file the image away.
329
The -optimize option to cjpeg is worth using when you are making a "final"
330
version for posting or archiving. It's also a win when you are using low
331
quality settings to make very small JPEG files; the percentage improvement
332
is often a lot more than it is on larger files. (At present, -optimize
333
mode is always selected when generating progressive JPEG files.)
335
GIF input files are no longer supported, to avoid the Unisys LZW patent.
336
Use a Unisys-licensed program if you need to read a GIF file. (Conversion
337
of GIF files to JPEG is usually a bad idea anyway.)
342
To get a quick preview of an image, use the -grayscale and/or -scale switches.
343
"-grayscale -scale 1/8" is the fastest case.
345
Several options are available that trade off image quality to gain speed.
346
"-fast" turns on the recommended settings.
348
"-dct fast" and/or "-nosmooth" gain speed at a small sacrifice in quality.
349
When producing a color-quantized image, "-onepass -dither ordered" is fast but
350
much lower quality than the default behavior. "-dither none" may give
351
acceptable results in two-pass mode, but is seldom tolerable in one-pass mode.
353
If you are fortunate enough to have very fast floating point hardware,
354
"-dct float" may be even faster than "-dct fast". But on most machines
355
"-dct float" is slower than "-dct int"; in this case it is not worth using,
356
because its theoretical accuracy advantage is too small to be significant
359
Two-pass color quantization requires a good deal of memory; on MS-DOS machines
360
it may run out of memory even with -maxmemory 0. In that case you can still
361
decompress, with some loss of image quality, by specifying -onepass for
362
one-pass quantization.
364
To avoid the Unisys LZW patent, djpeg produces uncompressed GIF files. These
365
are larger than they should be, but are readable by standard GIF decoders.
368
HINTS FOR BOTH PROGRAMS
370
If more space is needed than will fit in the available main memory (as
371
determined by -maxmemory), temporary files will be used. (MS-DOS versions
372
will try to get extended or expanded memory first.) The temporary files are
373
often rather large: in typical cases they occupy three bytes per pixel, for
374
example 3*800*600 = 1.44Mb for an 800x600 image. If you don't have enough
375
free disk space, leave out -progressive and -optimize (for cjpeg) or specify
376
-onepass (for djpeg).
378
On MS-DOS, the temporary files are created in the directory named by the TMP
379
or TEMP environment variable, or in the current directory if neither of those
380
exist. Amiga implementations put the temp files in the directory named by
381
JPEGTMP:, so be sure to assign JPEGTMP: to a disk partition with adequate free
384
The default memory usage limit (-maxmemory) is set when the software is
385
compiled. If you get an "insufficient memory" error, try specifying a smaller
386
-maxmemory value, even -maxmemory 0 to use the absolute minimum space. You
387
may want to recompile with a smaller default value if this happens often.
389
On machines that have "environment" variables, you can define the environment
390
variable JPEGMEM to set the default memory limit. The value is specified as
391
described for the -maxmemory switch. JPEGMEM overrides the default value
392
specified when the program was compiled, and itself is overridden by an
393
explicit -maxmemory switch.
395
On MS-DOS machines, -maxmemory is the amount of main (conventional) memory to
396
use. (Extended or expanded memory is also used if available.) Most
397
DOS-specific versions of this software do their own memory space estimation
398
and do not need you to specify -maxmemory.
403
jpegtran performs various useful transformations of JPEG files.
404
It can translate the coded representation from one variant of JPEG to another,
405
for example from baseline JPEG to progressive JPEG or vice versa. It can also
406
perform some rearrangements of the image data, for example turning an image
407
from landscape to portrait format by rotation.
409
jpegtran works by rearranging the compressed data (DCT coefficients), without
410
ever fully decoding the image. Therefore, its transformations are lossless:
411
there is no image degradation at all, which would not be true if you used
412
djpeg followed by cjpeg to accomplish the same conversion. But by the same
413
token, jpegtran cannot perform lossy operations such as changing the image
416
jpegtran uses a command line syntax similar to cjpeg or djpeg.
417
On Unix-like systems, you say:
418
jpegtran [switches] [inputfile] >outputfile
419
On most non-Unix systems, you say:
420
jpegtran [switches] inputfile outputfile
421
where both the input and output files are JPEG files.
423
To specify the coded JPEG representation used in the output file,
424
jpegtran accepts a subset of the switches recognized by cjpeg:
425
-optimize Perform optimization of entropy encoding parameters.
426
-progressive Create progressive JPEG file.
427
-restart N Emit a JPEG restart marker every N MCU rows, or every
428
N MCU blocks if "B" is attached to the number.
429
-scans file Use the scan script given in the specified text file.
430
See the previous discussion of cjpeg for more details about these switches.
431
If you specify none of these switches, you get a plain baseline-JPEG output
432
file. The quality setting and so forth are determined by the input file.
434
The image can be losslessly transformed by giving one of these switches:
435
-flip horizontal Mirror image horizontally (left-right).
436
-flip vertical Mirror image vertically (top-bottom).
437
-rotate 90 Rotate image 90 degrees clockwise.
438
-rotate 180 Rotate image 180 degrees.
439
-rotate 270 Rotate image 270 degrees clockwise (or 90 ccw).
440
-transpose Transpose image (across UL-to-LR axis).
441
-transverse Transverse transpose (across UR-to-LL axis).
443
The transpose transformation has no restrictions regarding image dimensions.
444
The other transformations operate rather oddly if the image dimensions are not
445
a multiple of the iMCU size (usually 8 or 16 pixels), because they can only
446
transform complete blocks of DCT coefficient data in the desired way.
448
jpegtran's default behavior when transforming an odd-size image is designed
449
to preserve exact reversibility and mathematical consistency of the
450
transformation set. As stated, transpose is able to flip the entire image
451
area. Horizontal mirroring leaves any partial iMCU column at the right edge
452
untouched, but is able to flip all rows of the image. Similarly, vertical
453
mirroring leaves any partial iMCU row at the bottom edge untouched, but is
454
able to flip all columns. The other transforms can be built up as sequences
455
of transpose and flip operations; for consistency, their actions on edge
456
pixels are defined to be the same as the end result of the corresponding
457
transpose-and-flip sequence.
459
For practical use, you may prefer to discard any untransformable edge pixels
460
rather than having a strange-looking strip along the right and/or bottom edges
461
of a transformed image. To do this, add the -trim switch:
462
-trim Drop non-transformable edge blocks.
463
Obviously, a transformation with -trim is not reversible, so strictly speaking
464
jpegtran with this switch is not lossless. Also, the expected mathematical
465
equivalences between the transformations no longer hold. For example,
466
"-rot 270 -trim" trims only the bottom edge, but "-rot 90 -trim" followed by
467
"-rot 180 -trim" trims both edges.
469
Another not-strictly-lossless transformation switch is:
470
-grayscale Force grayscale output.
471
This option discards the chrominance channels if the input image is YCbCr
472
(ie, a standard color JPEG), resulting in a grayscale JPEG file. The
473
luminance channel is preserved exactly, so this is a better method of reducing
474
to grayscale than decompression, conversion, and recompression. This switch
475
is particularly handy for fixing a monochrome picture that was mistakenly
476
encoded as a color JPEG. (In such a case, the space savings from getting rid
477
of the near-empty chroma channels won't be large; but the decoding time for
478
a grayscale JPEG is substantially less than that for a color JPEG.)
480
jpegtran also recognizes these switches that control what to do with "extra"
481
markers, such as comment blocks:
482
-copy none Copy no extra markers from source file. This setting
483
suppresses all comments and other excess baggage
484
present in the source file.
485
-copy comments Copy only comment markers. This setting copies
486
comments from the source file, but discards
487
any other inessential data.
488
-copy all Copy all extra markers. This setting preserves
489
miscellaneous markers found in the source file, such
490
as JFIF thumbnails and Photoshop settings. In some
491
files these extra markers can be sizable.
492
The default behavior is -copy comments. (Note: in IJG releases v6 and v6a,
493
jpegtran always did the equivalent of -copy none.)
495
Additional switches recognized by jpegtran are:
500
These work the same as in cjpeg or djpeg.
503
THE COMMENT UTILITIES
505
The JPEG standard allows "comment" (COM) blocks to occur within a JPEG file.
506
Although the standard doesn't actually define what COM blocks are for, they
507
are widely used to hold user-supplied text strings. This lets you add
508
annotations, titles, index terms, etc to your JPEG files, and later retrieve
509
them as text. COM blocks do not interfere with the image stored in the JPEG
510
file. The maximum size of a COM block is 64K, but you can have as many of
511
them as you like in one JPEG file.
513
We provide two utility programs to display COM block contents and add COM
514
blocks to a JPEG file.
516
rdjpgcom searches a JPEG file and prints the contents of any COM blocks on
517
standard output. The command line syntax is
518
rdjpgcom [-verbose] [inputfilename]
519
The switch "-verbose" (or just "-v") causes rdjpgcom to also display the JPEG
520
image dimensions. If you omit the input file name from the command line,
521
the JPEG file is read from standard input. (This may not work on some
522
operating systems, if binary data can't be read from stdin.)
524
wrjpgcom adds a COM block, containing text you provide, to a JPEG file.
525
Ordinarily, the COM block is added after any existing COM blocks, but you
526
can delete the old COM blocks if you wish. wrjpgcom produces a new JPEG
527
file; it does not modify the input file. DO NOT try to overwrite the input
528
file by directing wrjpgcom's output back into it; on most systems this will
529
just destroy your file.
531
The command line syntax for wrjpgcom is similar to cjpeg's. On Unix-like
533
wrjpgcom [switches] [inputfilename]
534
The output file is written to standard output. The input file comes from
535
the named file, or from standard input if no input file is named.
537
On most non-Unix systems, the syntax is
538
wrjpgcom [switches] inputfilename outputfilename
539
where both input and output file names must be given explicitly.
541
wrjpgcom understands three switches:
542
-replace Delete any existing COM blocks from the file.
543
-comment "Comment text" Supply new COM text on command line.
544
-cfile name Read text for new COM block from named file.
545
(Switch names can be abbreviated.) If you have only one line of comment text
546
to add, you can provide it on the command line with -comment. The comment
547
text must be surrounded with quotes so that it is treated as a single
548
argument. Longer comments can be read from a text file.
550
If you give neither -comment nor -cfile, then wrjpgcom will read the comment
551
text from standard input. (In this case an input image file name MUST be
552
supplied, so that the source JPEG file comes from somewhere else.) You can
553
enter multiple lines, up to 64KB worth. Type an end-of-file indicator
554
(usually control-D or control-Z) to terminate the comment text entry.
556
wrjpgcom will not add a COM block if the provided comment string is empty.
557
Therefore -replace -comment "" can be used to delete all COM blocks from a
560
These utility programs do not depend on the IJG JPEG library. In
561
particular, the source code for rdjpgcom is intended as an illustration of
562
the minimum amount of code required to parse a JPEG file header correctly.
added
removed
src/3rdparty/libjpeg/usage.doc
