1CJPEG(1) General Commands Manual CJPEG(1)
2
6 cjpeg - compress an image file to a JPEG file
7
9 cjpeg [ options ] [ filename ]
10
12 cjpeg compresses the named image file, or the standard input if no file
13 is named, and produces a JPEG/JFIF file on the standard output. The
14 currently supported input file formats are: PPM (PBMPLUS color format),
15 PGM (PBMPLUS grayscale format), BMP, GIF, and Targa.
16
18 All switch names may be abbreviated; for example, -grayscale may be
19 written -gray or -gr. Most of the "basic" switches can be abbreviated
20 to as little as one letter. Upper and lower case are equivalent (thus
21 -BMP is the same as -bmp). British spellings are also accepted (e.g.,
22 -greyscale), though for brevity these are not mentioned below.
23 The basic switches are:
25 -quality N[,...]
27 Scale quantization tables to adjust image quality. Quality is 0
28 (worst) to 100 (best); default is 75. (See below for more
29 info.)
30 -grayscale
32 Create monochrome JPEG file from color input. By saying
33 -grayscale, you'll get a smaller JPEG file that takes less time
34 to process.
35 -rgb Create RGB JPEG file. Using this switch suppresses the conver‐
37 sion from RGB colorspace input to the default YCbCr JPEG col‐
38 orspace.
39 -optimize
41 Perform optimization of entropy encoding parameters. Without
42 this, default encoding parameters are used. -optimize usually
43 makes the JPEG file a little smaller, but cjpeg runs somewhat
44 slower and needs much more memory. Image quality and speed of
45 decompression are unaffected by -optimize.
46 -progressive
48 Create progressive JPEG file (see below).
49 -targa Input file is Targa format. Targa files that contain an "iden‐
51 tification" field will not be automatically recognized by cjpeg;
52 for such files you must specify -targa to make cjpeg treat the
53 input as Targa format. For most Targa files, you won't need
54 this switch.
55 The -quality switch lets you trade off compressed file size against
57 quality of the reconstructed image: the higher the quality setting, the
58 larger the JPEG file, and the closer the output image will be to the
59 original input. Normally you want to use the lowest quality setting
60 (smallest file) that decompresses into something visually indistin‐
61 guishable from the original image. For this purpose the quality set‐
62 ting should generally be between 50 and 95 (the default is 75) for pho‐
63 tographic images. If you see defects at -quality 75, then go up 5 or
64 10 counts at a time until you are happy with the output image. (The
65 optimal setting will vary from one image to another.)
66 -quality 100 will generate a quantization table of all 1's, minimizing
68 loss in the quantization step (but there is still information loss in
69 subsampling, as well as roundoff error.) For most images, specifying a
70 quality value above about 95 will increase the size of the compressed
71 file dramatically, and while the quality gain from these higher quality
72 values is measurable (using metrics such as PSNR or SSIM), it is rarely
73 perceivable by human vision.
74 In the other direction, quality values below 50 will produce very small
76 files of low image quality. Settings around 5 to 10 might be useful in
77 preparing an index of a large image library, for example. Try -quality
78 2 (or so) for some amusing Cubist effects. (Note: quality values below
79 about 25 generate 2-byte quantization tables, which are considered op‐
80 tional in the JPEG standard. cjpeg emits a warning message when you
81 give such a quality value, because some other JPEG programs may be un‐
82 able to decode the resulting file. Use -baseline if you need to ensure
83 compatibility at low quality values.)
84 The -quality option has been extended in this version of cjpeg to sup‐
86 port separate quality settings for luminance and chrominance (or, in
87 general, separate settings for every quantization table slot.) The
88 principle is the same as chrominance subsampling: since the human eye
89 is more sensitive to spatial changes in brightness than spatial changes
90 in color, the chrominance components can be quantized more than the lu‐
91 minance components without incurring any visible image quality loss.
92 However, unlike subsampling, this feature reduces data in the frequency
93 domain instead of the spatial domain, which allows for more fine-
94 grained control. This option is useful in quality-sensitive applica‐
95 tions, for which the artifacts generated by subsampling may be unac‐
96 ceptable.
97 The -quality option accepts a comma-separated list of parameters, which
99 respectively refer to the quality levels that should be assigned to the
100 quantization table slots. If there are more q-table slots than parame‐
101 ters, then the last parameter is replicated. Thus, if only one quality
102 parameter is given, this is used for both luminance and chrominance
103 (slots 0 and 1, respectively), preserving the legacy behavior of cjpeg
104 v6b and prior. More (or customized) quantization tables can be set
105 with the -qtables option and assigned to components with the -qslots
106 option (see the "wizard" switches below.)
107 JPEG files generated with separate luminance and chrominance quality
109 are fully compliant with standard JPEG decoders.
110 CAUTION: For this setting to be useful, be sure to pass an argument of
112 -sample 1x1 to cjpeg to disable chrominance subsampling. Otherwise,
113 the default subsampling level (2x2, AKA "4:2:0") will be used.
114 The -progressive switch creates a "progressive JPEG" file. In this
116 type of JPEG file, the data is stored in multiple scans of increasing
117 quality. If the file is being transmitted over a slow communications
118 link, the decoder can use the first scan to display a low-quality image
119 very quickly, and can then improve the display with each subsequent
120 scan. The final image is exactly equivalent to a standard JPEG file of
121 the same quality setting, and the total file size is about the same ---
122 often a little smaller.
123 Switches for advanced users:
125 -arithmetic
127 Use arithmetic coding. Caution: arithmetic coded JPEG is not
128 yet widely implemented, so many decoders will be unable to view
129 an arithmetic coded JPEG file at all.
130 -dct int
132 Use accurate integer DCT method (default).
133 -dct fast
135 Use less accurate integer DCT method [legacy feature]. When the
136 Independent JPEG Group's software was first released in 1991,
137 the compression time for a 1-megapixel JPEG image on a main‐
138 stream PC was measured in minutes. Thus, the fast integer DCT
139 algorithm provided noticeable performance benefits. On modern
140 CPUs running libjpeg-turbo, however, the compression time for a
141 1-megapixel JPEG image is measured in milliseconds, and thus the
142 performance benefits of the fast algorithm are much less notice‐
143 able. On modern x86/x86-64 CPUs that support AVX2 instructions,
144 the fast and int methods have similar performance. On other
145 types of CPUs, the fast method is generally about 5-15% faster
146 than the int method.
147 For quality levels of 90 and below, there should be little or no
149 perceptible quality difference between the two algorithms. For
150 quality levels above 90, however, the difference between the
151 fast and int methods becomes more pronounced. With quality=97,
152 for instance, the fast method incurs generally about a 1-3 dB
153 loss in PSNR relative to the int method, but this can be larger
154 for some images. Do not use the fast method with quality levels
155 above 97. The algorithm often degenerates at quality=98 and
156 above and can actually produce a more lossy image than if lower
157 quality levels had been used. Also, in libjpeg-turbo, the fast
158 method is not fully accelerated for quality levels above 97, so
159 it will be slower than the int method.
160 -dct float
162 Use floating-point DCT method [legacy feature]. The float
163 method does not produce significantly more accurate results than
164 the int method, and it is much slower. The float method may
165 also give different results on different machines due to varying
166 roundoff behavior, whereas the integer methods should give the
167 same results on all machines.
168 -icc file
170 Embed ICC color management profile contained in the specified
171 file.
172 -restart N
174 Emit a JPEG restart marker every N MCU rows, or every N MCU
175 blocks if "B" is attached to the number. -restart 0 (the de‐
176 fault) means no restart markers.
177 -smooth N
179 Smooth the input image to eliminate dithering noise. N, ranging
180 from 1 to 100, indicates the strength of smoothing. 0 (the de‐
181 fault) means no smoothing.
182 -maxmemory N
184 Set limit for amount of memory to use in processing large im‐
185 ages. Value is in thousands of bytes, or millions of bytes if
186 "M" is attached to the number. For example, -max 4m selects
187 4000000 bytes. If more space is needed, an error will occur.
188 -outfile name
190 Send output image to the named file, not to standard output.
191 -memdst
193 Compress to memory instead of a file. This feature was imple‐
194 mented mainly as a way of testing the in-memory destination man‐
195 ager (jpeg_mem_dest()), but it is also useful for benchmarking,
196 since it reduces the I/O overhead.
197 -report
199 Report compression progress.
200 -strict
202 Treat all warnings as fatal. Enabling this option will cause
203 the compressor to abort if an LZW-compressed GIF input image
204 contains incomplete or corrupt image data.
205 -verbose
207 Enable debug printout. More -v's give more output. Also, ver‐
208 sion information is printed at startup.
209 -debug Same as -verbose.
211 -version
213 Print version information and exit.
214 The -restart option inserts extra markers that allow a JPEG decoder to
216 resynchronize after a transmission error. Without restart markers, any
217 damage to a compressed file will usually ruin the image from the point
218 of the error to the end of the image; with restart markers, the damage
219 is usually confined to the portion of the image up to the next restart
220 marker. Of course, the restart markers occupy extra space. We recom‐
221 mend -restart 1 for images that will be transmitted across unreliable
222 networks such as Usenet.
223 The -smooth option filters the input to eliminate fine-scale noise.
225 This is often useful when converting dithered images to JPEG: a moder‐
226 ate smoothing factor of 10 to 50 gets rid of dithering patterns in the
227 input file, resulting in a smaller JPEG file and a better-looking im‐
228 age. Too large a smoothing factor will visibly blur the image, how‐
229 ever.
230 Switches for wizards:
232 -baseline
234 Force baseline-compatible quantization tables to be generated.
235 This clamps quantization values to 8 bits even at low quality
236 settings. (This switch is poorly named, since it does not en‐
237 sure that the output is actually baseline JPEG. For example,
238 you can use -baseline and -progressive together.)
239 -qtables file
241 Use the quantization tables given in the specified text file.
242 -qslots N[,...]
244 Select which quantization table to use for each color component.
245 -sample HxV[,...]
247 Set JPEG sampling factors for each color component.
248 -scans file
250 Use the scan script given in the specified text file.
251 The "wizard" switches are intended for experimentation with JPEG. If
253 you don't know what you are doing, don't use them. These switches are
254 documented further in the file wizard.txt.
255
257 This example compresses the PPM file foo.ppm with a quality factor of
258 60 and saves the output as foo.jpg:
259 cjpeg -quality 60 foo.ppm > foo.jpg
261
263 Color GIF files are not the ideal input for JPEG; JPEG is really in‐
264 tended for compressing full-color (24-bit) images. In particular,
265 don't try to convert cartoons, line drawings, and other images that
266 have only a few distinct colors. GIF works great on these, JPEG does
267 not. If you want to convert a GIF to JPEG, you should experiment with
268 cjpeg's -quality and -smooth options to get a satisfactory conversion.
269 -smooth 10 or so is often helpful.
270 Avoid running an image through a series of JPEG compression/decompres‐
272 sion cycles. Image quality loss will accumulate; after ten or so cy‐
273 cles the image may be noticeably worse than it was after one cycle.
274 It's best to use a lossless format while manipulating an image, then
275 convert to JPEG format when you are ready to file the image away.
276 The -optimize option to cjpeg is worth using when you are making a "fi‐
278 nal" version for posting or archiving. It's also a win when you are
279 using low quality settings to make very small JPEG files; the percent‐
280 age improvement is often a lot more than it is on larger files. (At
281 present, -optimize mode is always selected when generating progressive
282 JPEG files.)
283
285 JPEGMEM
286 If this environment variable is set, its value is the default
287 memory limit. The value is specified as described for the
288 -maxmemory switch. JPEGMEM overrides the default value speci‐
289 fied when the program was compiled, and itself is overridden by
290 an explicit -maxmemory.
291
293 djpeg(1), jpegtran(1), rdjpgcom(1), wrjpgcom(1)
294 ppm(5), pgm(5)
295 Wallace, Gregory K. "The JPEG Still Picture Compression Standard",
296 Communications of the ACM, April 1991 (vol. 34, no. 4), pp. 30-44.
297
299 Independent JPEG Group
300 This file was modified by The libjpeg-turbo Project to include only in‐
302 formation relevant to libjpeg-turbo, to wordsmith certain sections, and
303 to describe features not present in libjpeg.
304
306 Not all variants of BMP and Targa file formats are supported.
307 The -targa switch is not a bug, it's a feature. (It would be a bug if
309 the Targa format designers had not been clueless.)
310 30 November 2021 CJPEG(1)