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, Targa, and RLE (Utah Raster Tool‐
16 kit format). (RLE is supported only if the URT library is available.)
17
19 All switch names may be abbreviated; for example, -grayscale may be
20 written -gray or -gr. Most of the "basic" switches can be abbreviated
21 to as little as one letter. Upper and lower case are equivalent (thus
22 -BMP is the same as -bmp). British spellings are also accepted (e.g.,
23 -greyscale), though for brevity these are not mentioned below.
24 The basic switches are:
26 -quality N[,...]
28 Scale quantization tables to adjust image quality. Quality is 0
29 (worst) to 100 (best); default is 75. (See below for more
30 info.)
31 -grayscale
33 Create monochrome JPEG file from color input. Be sure to use
34 this switch when compressing a grayscale BMP file, because cjpeg
35 isn't bright enough to notice whether a BMP file uses only
36 shades of gray. By saying -grayscale, you'll get a smaller JPEG
37 file that takes less time to process.
38 -rgb Create RGB JPEG file. Using this switch suppresses the conver‐
40 sion from RGB colorspace input to the default YCbCr JPEG col‐
41 orspace.
42 -optimize
44 Perform optimization of entropy encoding parameters. Without
45 this, default encoding parameters are used. -optimize usually
46 makes the JPEG file a little smaller, but cjpeg runs somewhat
47 slower and needs much more memory. Image quality and speed of
48 decompression are unaffected by -optimize.
49 -progressive
51 Create progressive JPEG file (see below).
52 -targa Input file is Targa format. Targa files that contain an "iden‐
54 tification" field will not be automatically recognized by cjpeg;
55 for such files you must specify -targa to make cjpeg treat the
56 input as Targa format. For most Targa files, you won't need
57 this switch.
58 The -quality switch lets you trade off compressed file size against
60 quality of the reconstructed image: the higher the quality setting, the
61 larger the JPEG file, and the closer the output image will be to the
62 original input. Normally you want to use the lowest quality setting
63 (smallest file) that decompresses into something visually indistin‐
64 guishable from the original image. For this purpose the quality set‐
65 ting should generally be between 50 and 95 (the default is 75) for pho‐
66 tographic images. If you see defects at -quality 75, then go up 5 or
67 10 counts at a time until you are happy with the output image. (The
68 optimal setting will vary from one image to another.)
69 -quality 100 will generate a quantization table of all 1's, minimizing
71 loss in the quantization step (but there is still information loss in
72 subsampling, as well as roundoff error.) For most images, specifying a
73 quality value above about 95 will increase the size of the compressed
74 file dramatically, and while the quality gain from these higher quality
75 values is measurable (using metrics such as PSNR or SSIM), it is rarely
76 perceivable by human vision.
77 In the other direction, quality values below 50 will produce very small
79 files of low image quality. Settings around 5 to 10 might be useful in
80 preparing an index of a large image library, for example. Try -quality
81 2 (or so) for some amusing Cubist effects. (Note: quality values below
82 about 25 generate 2-byte quantization tables, which are considered
83 optional in the JPEG standard. cjpeg emits a warning message when you
84 give such a quality value, because some other JPEG programs may be
85 unable to decode the resulting file. Use -baseline if you need to
86 ensure compatibility at low quality values.)
87 The -quality option has been extended in this version of cjpeg to sup‐
89 port separate quality settings for luminance and chrominance (or, in
90 general, separate settings for every quantization table slot.) The
91 principle is the same as chrominance subsampling: since the human eye
92 is more sensitive to spatial changes in brightness than spatial changes
93 in color, the chrominance components can be quantized more than the
94 luminance components without incurring any visible image quality loss.
95 However, unlike subsampling, this feature reduces data in the frequency
96 domain instead of the spatial domain, which allows for more fine-
97 grained control. This option is useful in quality-sensitive applica‐
98 tions, for which the artifacts generated by subsampling may be unac‐
99 ceptable.
100 The -quality option accepts a comma-separated list of parameters, which
102 respectively refer to the quality levels that should be assigned to the
103 quantization table slots. If there are more q-table slots than parame‐
104 ters, then the last parameter is replicated. Thus, if only one quality
105 parameter is given, this is used for both luminance and chrominance
106 (slots 0 and 1, respectively), preserving the legacy behavior of cjpeg
107 v6b and prior. More (or customized) quantization tables can be set
108 with the -qtables option and assigned to components with the -qslots
109 option (see the "wizard" switches below.)
110 JPEG files generated with separate luminance and chrominance quality
112 are fully compliant with standard JPEG decoders.
113 CAUTION: For this setting to be useful, be sure to pass an argument of
115 -sample 1x1 to cjpeg to disable chrominance subsampling. Otherwise,
116 the default subsampling level (2x2, AKA "4:2:0") will be used.
117 The -progressive switch creates a "progressive JPEG" file. In this
119 type of JPEG file, the data is stored in multiple scans of increasing
120 quality. If the file is being transmitted over a slow communications
121 link, the decoder can use the first scan to display a low-quality image
122 very quickly, and can then improve the display with each subsequent
123 scan. The final image is exactly equivalent to a standard JPEG file of
124 the same quality setting, and the total file size is about the same ---
125 often a little smaller.
126 Switches for advanced users:
128 -arithmetic
130 Use arithmetic coding. Caution: arithmetic coded JPEG is not
131 yet widely implemented, so many decoders will be unable to view
132 an arithmetic coded JPEG file at all.
133 -dct int
135 Use integer DCT method (default).
136 -dct fast
138 Use fast integer DCT (less accurate). In libjpeg-turbo, the
139 fast method is generally about 5-15% faster than the int method
140 when using the x86/x86-64 SIMD extensions (results may vary with
141 other SIMD implementations, or when using libjpeg-turbo without
142 SIMD extensions.) For quality levels of 90 and below, there
143 should be little or no perceptible difference between the two
144 algorithms. For quality levels above 90, however, the differ‐
145 ence between the fast and the int methods becomes more pro‐
146 nounced. With quality=97, for instance, the fast method incurs
147 generally about a 1-3 dB loss (in PSNR) relative to the int
148 method, but this can be larger for some images. Do not use the
149 fast method with quality levels above 97. The algorithm often
150 degenerates at quality=98 and above and can actually produce a
151 more lossy image than if lower quality levels had been used.
152 Also, in libjpeg-turbo, the fast method is not fully accelerated
153 for quality levels above 97, so it will be slower than the int
154 method.
155 -dct float
157 Use floating-point DCT method. The float method is mainly a
158 legacy feature. It does not produce significantly more accurate
159 results than the int method, and it is much slower. The float
160 method may also give different results on different machines due
161 to varying roundoff behavior, whereas the integer methods should
162 give the same results on all machines.
163 -restart N
165 Emit a JPEG restart marker every N MCU rows, or every N MCU
166 blocks if "B" is attached to the number. -restart 0 (the
167 default) means no restart markers.
168 -smooth N
170 Smooth the input image to eliminate dithering noise. N, ranging
171 from 1 to 100, indicates the strength of smoothing. 0 (the
172 default) means no smoothing.
173 -maxmemory N
175 Set limit for amount of memory to use in processing large
176 images. Value is in thousands of bytes, or millions of bytes if
177 "M" is attached to the number. For example, -max 4m selects
178 4000000 bytes. If more space is needed, an error will occur.
179 -outfile name
181 Send output image to the named file, not to standard output.
182 -memdst
184 Compress to memory instead of a file. This feature was imple‐
185 mented mainly as a way of testing the in-memory destination man‐
186 ager (jpeg_mem_dest()), but it is also useful for benchmarking,
187 since it reduces the I/O overhead.
188 -verbose
190 Enable debug printout. More -v's give more output. Also, ver‐
191 sion information is printed at startup.
192 -debug Same as -verbose.
194 -version
196 Print version information and exit.
197 The -restart option inserts extra markers that allow a JPEG decoder to
199 resynchronize after a transmission error. Without restart markers, any
200 damage to a compressed file will usually ruin the image from the point
201 of the error to the end of the image; with restart markers, the damage
202 is usually confined to the portion of the image up to the next restart
203 marker. Of course, the restart markers occupy extra space. We recom‐
204 mend -restart 1 for images that will be transmitted across unreliable
205 networks such as Usenet.
206 The -smooth option filters the input to eliminate fine-scale noise.
208 This is often useful when converting dithered images to JPEG: a moder‐
209 ate smoothing factor of 10 to 50 gets rid of dithering patterns in the
210 input file, resulting in a smaller JPEG file and a better-looking
211 image. Too large a smoothing factor will visibly blur the image, how‐
212 ever.
213 Switches for wizards:
215 -baseline
217 Force baseline-compatible quantization tables to be generated.
218 This clamps quantization values to 8 bits even at low quality
219 settings. (This switch is poorly named, since it does not
220 ensure that the output is actually baseline JPEG. For example,
221 you can use -baseline and -progressive together.)
222 -qtables file
224 Use the quantization tables given in the specified text file.
225 -qslots N[,...]
227 Select which quantization table to use for each color component.
228 -sample HxV[,...]
230 Set JPEG sampling factors for each color component.
231 -scans file
233 Use the scan script given in the specified text file.
234 The "wizard" switches are intended for experimentation with JPEG. If
236 you don't know what you are doing, don't use them. These switches are
237 documented further in the file wizard.txt.
238
240 This example compresses the PPM file foo.ppm with a quality factor of
241 60 and saves the output as foo.jpg:
242 cjpeg -quality 60 foo.ppm > foo.jpg
244
246 Color GIF files are not the ideal input for JPEG; JPEG is really
247 intended for compressing full-color (24-bit) images. In particular,
248 don't try to convert cartoons, line drawings, and other images that
249 have only a few distinct colors. GIF works great on these, JPEG does
250 not. If you want to convert a GIF to JPEG, you should experiment with
251 cjpeg's -quality and -smooth options to get a satisfactory conversion.
252 -smooth 10 or so is often helpful.
253 Avoid running an image through a series of JPEG compression/decompres‐
255 sion cycles. Image quality loss will accumulate; after ten or so
256 cycles the image may be noticeably worse than it was after one cycle.
257 It's best to use a lossless format while manipulating an image, then
258 convert to JPEG format when you are ready to file the image away.
259 The -optimize option to cjpeg is worth using when you are making a
261 "final" version for posting or archiving. It's also a win when you are
262 using low quality settings to make very small JPEG files; the percent‐
263 age improvement is often a lot more than it is on larger files. (At
264 present, -optimize mode is always selected when generating progressive
265 JPEG files.)
266
268 JPEGMEM
269 If this environment variable is set, its value is the default
270 memory limit. The value is specified as described for the
271 -maxmemory switch. JPEGMEM overrides the default value speci‐
272 fied when the program was compiled, and itself is overridden by
273 an explicit -maxmemory.
274
276 djpeg(1), jpegtran(1), rdjpgcom(1), wrjpgcom(1)
277 ppm(5), pgm(5)
278 Wallace, Gregory K. "The JPEG Still Picture Compression Standard",
279 Communications of the ACM, April 1991 (vol. 34, no. 4), pp. 30-44.
280
282 Independent JPEG Group
283 This file was modified by The libjpeg-turbo Project to include only
285 information relevant to libjpeg-turbo, to wordsmith certain sections,
286 and to describe features not present in libjpeg.
287
289 Support for GIF input files was removed in cjpeg v6b due to concerns
290 over the Unisys LZW patent. Although this patent expired in 2006,
291 cjpeg still lacks GIF support, for these historical reasons. (Conver‐
292 sion of GIF files to JPEG is usually a bad idea anyway, since GIF is a
293 256-color format.)
294 Not all variants of BMP and Targa file formats are supported.
296 The -targa switch is not a bug, it's a feature. (It would be a bug if
298 the Targa format designers had not been clueless.)
299 18 March 2017 CJPEG(1)