1NOTE: This file was modified by The libjpeg-turbo Project to include only 2information relevant to libjpeg-turbo and to wordsmith certain sections. 3 4USAGE instructions for the Independent JPEG Group's JPEG software 5================================================================= 6 7This file describes usage of the JPEG conversion programs cjpeg and djpeg, 8as well as the utility programs jpegtran, rdjpgcom and wrjpgcom. (See 9the other documentation files if you wish to use the JPEG library within 10your own programs.) 11 12If you are on a Unix machine you may prefer to read the Unix-style manual 13pages in files cjpeg.1, djpeg.1, jpegtran.1, rdjpgcom.1, wrjpgcom.1. 14 15 16INTRODUCTION 17 18These programs implement JPEG image encoding, decoding, and transcoding. 19JPEG (pronounced "jay-peg") is a standardized compression method for 20full-color and grayscale images. 21 22 23GENERAL USAGE 24 25We provide two programs, cjpeg to compress an image file into JPEG format, 26and djpeg to decompress a JPEG file back into a conventional image format. 27 28On Unix-like systems, you say: 29 cjpeg [switches] [imagefile] >jpegfile 30or 31 djpeg [switches] [jpegfile] >imagefile 32The programs read the specified input file, or standard input if none is 33named. They always write to standard output (with trace/error messages to 34standard error). These conventions are handy for piping images between 35programs. 36 37On most non-Unix systems, you say: 38 cjpeg [switches] imagefile jpegfile 39or 40 djpeg [switches] jpegfile imagefile 41i.e., both the input and output files are named on the command line. This 42style is a little more foolproof, and it loses no functionality if you don't 43have pipes. (You can get this style on Unix too, if you prefer, by defining 44TWO_FILE_COMMANDLINE when you compile the programs; see install.txt.) 45 46You can also say: 47 cjpeg [switches] -outfile jpegfile imagefile 48or 49 djpeg [switches] -outfile imagefile jpegfile 50This syntax works on all systems, so it is useful for scripts. 51 52The currently supported image file formats are: PPM (PBMPLUS color format), 53PGM (PBMPLUS grayscale format), BMP, Targa, and RLE (Utah Raster Toolkit 54format). (RLE is supported only if the URT library is available.) 55cjpeg recognizes the input image format automatically, with the exception 56of some Targa-format files. You have to tell djpeg which format to generate. 57 58JPEG files are in the defacto standard JFIF file format. There are other, 59less widely used JPEG-based file formats, but we don't support them. 60 61All switch names may be abbreviated; for example, -grayscale may be written 62-gray or -gr. Most of the "basic" switches can be abbreviated to as little as 63one letter. Upper and lower case are equivalent (-BMP is the same as -bmp). 64British spellings are also accepted (e.g., -greyscale), though for brevity 65these are not mentioned below. 66 67 68CJPEG DETAILS 69 70The basic command line switches for cjpeg are: 71 72 -quality N[,...] Scale quantization tables to adjust image quality. 73 Quality is 0 (worst) to 100 (best); default is 75. 74 (See below for more info.) 75 76 -grayscale Create monochrome JPEG file from color input. 77 Be sure to use this switch when compressing a grayscale 78 BMP file, because cjpeg isn't bright enough to notice 79 whether a BMP file uses only shades of gray. By 80 saying -grayscale, you'll get a smaller JPEG file that 81 takes less time to process. 82 83 -rgb Create RGB JPEG file. 84 Using this switch suppresses the conversion from RGB 85 colorspace input to the default YCbCr JPEG colorspace. 86 87 -optimize Perform optimization of entropy encoding parameters. 88 Without this, default encoding parameters are used. 89 -optimize usually makes the JPEG file a little smaller, 90 but cjpeg runs somewhat slower and needs much more 91 memory. Image quality and speed of decompression are 92 unaffected by -optimize. 93 94 -progressive Create progressive JPEG file (see below). 95 96 -targa Input file is Targa format. Targa files that contain 97 an "identification" field will not be automatically 98 recognized by cjpeg; for such files you must specify 99 -targa to make cjpeg treat the input as Targa format. 100 For most Targa files, you won't need this switch. 101 102The -quality switch lets you trade off compressed file size against quality of 103the reconstructed image: the higher the quality setting, the larger the JPEG 104file, and the closer the output image will be to the original input. Normally 105you want to use the lowest quality setting (smallest file) that decompresses 106into something visually indistinguishable from the original image. For this 107purpose the quality setting should be between 50 and 95; the default of 75 is 108often about right. If you see defects at -quality 75, then go up 5 or 10 109counts at a time until you are happy with the output image. (The optimal 110setting will vary from one image to another.) 111 112-quality 100 will generate a quantization table of all 1's, minimizing loss 113in the quantization step (but there is still information loss in subsampling, 114as well as roundoff error). This setting is mainly of interest for 115experimental purposes. Quality values above about 95 are NOT recommended for 116normal use; the compressed file size goes up dramatically for hardly any gain 117in output image quality. 118 119In the other direction, quality values below 50 will produce very small files 120of low image quality. Settings around 5 to 10 might be useful in preparing an 121index of a large image library, for example. Try -quality 2 (or so) for some 122amusing Cubist effects. (Note: quality values below about 25 generate 2-byte 123quantization tables, which are considered optional in the JPEG standard. 124cjpeg emits a warning message when you give such a quality value, because some 125other JPEG programs may be unable to decode the resulting file. Use -baseline 126if you need to ensure compatibility at low quality values.) 127 128The -quality option has been extended in this version of cjpeg to support 129separate quality settings for luminance and chrominance (or, in general, 130separate settings for every quantization table slot.) The principle is the 131same as chrominance subsampling: since the human eye is more sensitive to 132spatial changes in brightness than spatial changes in color, the chrominance 133components can be quantized more than the luminance components without 134incurring any visible image quality loss. However, unlike subsampling, this 135feature reduces data in the frequency domain instead of the spatial domain, 136which allows for more fine-grained control. This option is useful in 137quality-sensitive applications, for which the artifacts generated by 138subsampling may be unacceptable. 139 140The -quality option accepts a comma-separated list of parameters, which 141respectively refer to the quality levels that should be assigned to the 142quantization table slots. If there are more q-table slots than parameters, 143then the last parameter is replicated. Thus, if only one quality parameter is 144given, this is used for both luminance and chrominance (slots 0 and 1, 145respectively), preserving the legacy behavior of cjpeg v6b and prior. More (or 146customized) quantization tables can be set with the -qtables option and 147assigned to components with the -qslots option (see the "wizard" switches 148below.) 149 150JPEG files generated with separate luminance and chrominance quality are 151fully compliant with standard JPEG decoders. 152 153CAUTION: For this setting to be useful, be sure to pass an argument of 154-sample 1x1 to cjpeg to disable chrominance subsampling. Otherwise, the 155default subsampling level (2x2, AKA "4:2:0") will be used. 156 157The -progressive switch creates a "progressive JPEG" file. In this type of 158JPEG file, the data is stored in multiple scans of increasing quality. If the 159file is being transmitted over a slow communications link, the decoder can use 160the first scan to display a low-quality image very quickly, and can then 161improve the display with each subsequent scan. The final image is exactly 162equivalent to a standard JPEG file of the same quality setting, and the total 163file size is about the same --- often a little smaller. 164 165Switches for advanced users: 166 167 -arithmetic Use arithmetic coding. CAUTION: arithmetic coded JPEG 168 is not yet widely implemented, so many decoders will 169 be unable to view an arithmetic coded JPEG file at 170 all. 171 172 -dct int Use integer DCT method (default). 173 -dct fast Use fast integer DCT (less accurate). 174 In libjpeg-turbo, the fast method is generally about 175 5-15% faster than the int method when using the 176 x86/x86-64 SIMD extensions (results may vary with other 177 SIMD implementations, or when using libjpeg-turbo 178 without SIMD extensions.) For quality levels of 90 and 179 below, there should be little or no perceptible 180 difference between the two algorithms. For quality 181 levels above 90, however, the difference between 182 the fast and the int methods becomes more pronounced. 183 With quality=97, for instance, the fast method incurs 184 generally about a 1-3 dB loss (in PSNR) relative to 185 the int method, but this can be larger for some images. 186 Do not use the fast method with quality levels above 187 97. The algorithm often degenerates at quality=98 and 188 above and can actually produce a more lossy image than 189 if lower quality levels had been used. Also, in 190 libjpeg-turbo, the fast method is not fully accerated 191 for quality levels above 97, so it will be slower than 192 the int method. 193 -dct float Use floating-point DCT method. 194 The float method is mainly a legacy feature. It does 195 not produce significantly more accurate results than 196 the int method, and it is much slower. The float 197 method may also give different results on different 198 machines due to varying roundoff behavior, whereas the 199 integer methods should give the same results on all 200 machines. 201 202 -restart N Emit a JPEG restart marker every N MCU rows, or every 203 N MCU blocks if "B" is attached to the number. 204 -restart 0 (the default) means no restart markers. 205 206 -smooth N Smooth the input image to eliminate dithering noise. 207 N, ranging from 1 to 100, indicates the strength of 208 smoothing. 0 (the default) means no smoothing. 209 210 -maxmemory N Set limit for amount of memory to use in processing 211 large images. Value is in thousands of bytes, or 212 millions of bytes if "M" is attached to the number. 213 For example, -max 4m selects 4000000 bytes. If more 214 space is needed, temporary files will be used. 215 216 -verbose Enable debug printout. More -v's give more printout. 217 or -debug Also, version information is printed at startup. 218 219The -restart option inserts extra markers that allow a JPEG decoder to 220resynchronize after a transmission error. Without restart markers, any damage 221to a compressed file will usually ruin the image from the point of the error 222to the end of the image; with restart markers, the damage is usually confined 223to the portion of the image up to the next restart marker. Of course, the 224restart markers occupy extra space. We recommend -restart 1 for images that 225will be transmitted across unreliable networks such as Usenet. 226 227The -smooth option filters the input to eliminate fine-scale noise. This is 228often useful when converting dithered images to JPEG: a moderate smoothing 229factor of 10 to 50 gets rid of dithering patterns in the input file, resulting 230in a smaller JPEG file and a better-looking image. Too large a smoothing 231factor will visibly blur the image, however. 232 233Switches for wizards: 234 235 -baseline Force baseline-compatible quantization tables to be 236 generated. This clamps quantization values to 8 bits 237 even at low quality settings. (This switch is poorly 238 named, since it does not ensure that the output is 239 actually baseline JPEG. For example, you can use 240 -baseline and -progressive together.) 241 242 -qtables file Use the quantization tables given in the specified 243 text file. 244 245 -qslots N[,...] Select which quantization table to use for each color 246 component. 247 248 -sample HxV[,...] Set JPEG sampling factors for each color component. 249 250 -scans file Use the scan script given in the specified text file. 251 252The "wizard" switches are intended for experimentation with JPEG. If you 253don't know what you are doing, DON'T USE THEM. These switches are documented 254further in the file wizard.txt. 255 256 257DJPEG DETAILS 258 259The basic command line switches for djpeg are: 260 261 -colors N Reduce image to at most N colors. This reduces the 262 or -quantize N number of colors used in the output image, so that it 263 can be displayed on a colormapped display or stored in 264 a colormapped file format. For example, if you have 265 an 8-bit display, you'd need to reduce to 256 or fewer 266 colors. (-colors is the recommended name, -quantize 267 is provided only for backwards compatibility.) 268 269 -fast Select recommended processing options for fast, low 270 quality output. (The default options are chosen for 271 highest quality output.) Currently, this is equivalent 272 to "-dct fast -nosmooth -onepass -dither ordered". 273 274 -grayscale Force grayscale output even if JPEG file is color. 275 Useful for viewing on monochrome displays; also, 276 djpeg runs noticeably faster in this mode. 277 278 -scale M/N Scale the output image by a factor M/N. Currently 279 the scale factor must be M/8, where M is an integer 280 between 1 and 16 inclusive, or any reduced fraction 281 thereof (such as 1/2, 3/4, etc. Scaling is handy if 282 the image is larger than your screen; also, djpeg runs 283 much faster when scaling down the output. 284 285 -bmp Select BMP output format (Windows flavor). 8-bit 286 colormapped format is emitted if -colors or -grayscale 287 is specified, or if the JPEG file is grayscale; 288 otherwise, 24-bit full-color format is emitted. 289 290 -gif Select GIF output format. Since GIF does not support 291 more than 256 colors, -colors 256 is assumed (unless 292 you specify a smaller number of colors). If you 293 specify -fast, the default number of colors is 216. 294 295 -os2 Select BMP output format (OS/2 1.x flavor). 8-bit 296 colormapped format is emitted if -colors or -grayscale 297 is specified, or if the JPEG file is grayscale; 298 otherwise, 24-bit full-color format is emitted. 299 300 -pnm Select PBMPLUS (PPM/PGM) output format (this is the 301 default format). PGM is emitted if the JPEG file is 302 grayscale or if -grayscale is specified; otherwise 303 PPM is emitted. 304 305 -rle Select RLE output format. (Requires URT library.) 306 307 -targa Select Targa output format. Grayscale format is 308 emitted if the JPEG file is grayscale or if 309 -grayscale is specified; otherwise, colormapped format 310 is emitted if -colors is specified; otherwise, 24-bit 311 full-color format is emitted. 312 313Switches for advanced users: 314 315 -dct int Use integer DCT method (default). 316 -dct fast Use fast integer DCT (less accurate). 317 In libjpeg-turbo, the fast method is generally about 318 5-15% faster than the int method when using the 319 x86/x86-64 SIMD extensions (results may vary with other 320 SIMD implementations, or when using libjpeg-turbo 321 without SIMD extensions.) If the JPEG image was 322 compressed using a quality level of 85 or below, then 323 there should be little or no perceptible difference 324 between the two algorithms. When decompressing images 325 that were compressed using quality levels above 85, 326 however, the difference between the fast and int 327 methods becomes more pronounced. With images 328 compressed using quality=97, for instance, the fast 329 method incurs generally about a 4-6 dB loss (in PSNR) 330 relative to the int method, but this can be larger for 331 some images. If you can avoid it, do not use the fast 332 method when decompressing images that were compressed 333 using quality levels above 97. The algorithm often 334 degenerates for such images and can actually produce 335 a more lossy output image than if the JPEG image had 336 been compressed using lower quality levels. 337 -dct float Use floating-point DCT method. 338 The float method is mainly a legacy feature. It does 339 not produce significantly more accurate results than 340 the int method, and it is much slower. The float 341 method may also give different results on different 342 machines due to varying roundoff behavior, whereas the 343 integer methods should give the same results on all 344 machines. 345 346 -dither fs Use Floyd-Steinberg dithering in color quantization. 347 -dither ordered Use ordered dithering in color quantization. 348 -dither none Do not use dithering in color quantization. 349 By default, Floyd-Steinberg dithering is applied when 350 quantizing colors; this is slow but usually produces 351 the best results. Ordered dither is a compromise 352 between speed and quality; no dithering is fast but 353 usually looks awful. Note that these switches have 354 no effect unless color quantization is being done. 355 Ordered dither is only available in -onepass mode. 356 357 -map FILE Quantize to the colors used in the specified image 358 file. This is useful for producing multiple files 359 with identical color maps, or for forcing a predefined 360 set of colors to be used. The FILE must be a GIF 361 or PPM file. This option overrides -colors and 362 -onepass. 363 364 -nosmooth Use a faster, lower-quality upsampling routine. 365 366 -onepass Use one-pass instead of two-pass color quantization. 367 The one-pass method is faster and needs less memory, 368 but it produces a lower-quality image. -onepass is 369 ignored unless you also say -colors N. Also, 370 the one-pass method is always used for grayscale 371 output (the two-pass method is no improvement then). 372 373 -maxmemory N Set limit for amount of memory to use in processing 374 large images. Value is in thousands of bytes, or 375 millions of bytes if "M" is attached to the number. 376 For example, -max 4m selects 4000000 bytes. If more 377 space is needed, temporary files will be used. 378 379 -verbose Enable debug printout. More -v's give more printout. 380 or -debug Also, version information is printed at startup. 381 382 383HINTS FOR CJPEG 384 385Color GIF files are not the ideal input for JPEG; JPEG is really intended for 386compressing full-color (24-bit) images. In particular, don't try to convert 387cartoons, line drawings, and other images that have only a few distinct 388colors. GIF works great on these, JPEG does not. If you want to convert a 389GIF to JPEG, you should experiment with cjpeg's -quality and -smooth options 390to get a satisfactory conversion. -smooth 10 or so is often helpful. 391 392Avoid running an image through a series of JPEG compression/decompression 393cycles. Image quality loss will accumulate; after ten or so cycles the image 394may be noticeably worse than it was after one cycle. It's best to use a 395lossless format while manipulating an image, then convert to JPEG format when 396you are ready to file the image away. 397 398The -optimize option to cjpeg is worth using when you are making a "final" 399version for posting or archiving. It's also a win when you are using low 400quality settings to make very small JPEG files; the percentage improvement 401is often a lot more than it is on larger files. (At present, -optimize 402mode is always selected when generating progressive JPEG files.) 403 404Support for GIF input files was removed in cjpeg v6b due to concerns over 405the Unisys LZW patent. Although this patent expired in 2006, cjpeg still 406lacks GIF support, for these historical reasons. (Conversion of GIF files to 407JPEG is usually a bad idea anyway.) 408 409 410HINTS FOR DJPEG 411 412To get a quick preview of an image, use the -grayscale and/or -scale switches. 413"-grayscale -scale 1/8" is the fastest case. 414 415Several options are available that trade off image quality to gain speed. 416"-fast" turns on the recommended settings. 417 418"-dct fast" and/or "-nosmooth" gain speed at a small sacrifice in quality. 419When producing a color-quantized image, "-onepass -dither ordered" is fast but 420much lower quality than the default behavior. "-dither none" may give 421acceptable results in two-pass mode, but is seldom tolerable in one-pass mode. 422 423Two-pass color quantization requires a good deal of memory; on MS-DOS machines 424it may run out of memory even with -maxmemory 0. In that case you can still 425decompress, with some loss of image quality, by specifying -onepass for 426one-pass quantization. 427 428To avoid the Unisys LZW patent, djpeg produces uncompressed GIF files. These 429are larger than they should be, but are readable by standard GIF decoders. 430 431 432HINTS FOR BOTH PROGRAMS 433 434If more space is needed than will fit in the available main memory (as 435determined by -maxmemory), temporary files will be used. (MS-DOS versions 436will try to get extended or expanded memory first.) The temporary files are 437often rather large: in typical cases they occupy three bytes per pixel, for 438example 3*800*600 = 1.44Mb for an 800x600 image. If you don't have enough 439free disk space, leave out -progressive and -optimize (for cjpeg) or specify 440-onepass (for djpeg). 441 442On MS-DOS, the temporary files are created in the directory named by the TMP 443or TEMP environment variable, or in the current directory if neither of those 444exist. Amiga implementations put the temp files in the directory named by 445JPEGTMP:, so be sure to assign JPEGTMP: to a disk partition with adequate free 446space. 447 448The default memory usage limit (-maxmemory) is set when the software is 449compiled. If you get an "insufficient memory" error, try specifying a smaller 450-maxmemory value, even -maxmemory 0 to use the absolute minimum space. You 451may want to recompile with a smaller default value if this happens often. 452 453On machines that have "environment" variables, you can define the environment 454variable JPEGMEM to set the default memory limit. The value is specified as 455described for the -maxmemory switch. JPEGMEM overrides the default value 456specified when the program was compiled, and itself is overridden by an 457explicit -maxmemory switch. 458 459On MS-DOS machines, -maxmemory is the amount of main (conventional) memory to 460use. (Extended or expanded memory is also used if available.) Most 461DOS-specific versions of this software do their own memory space estimation 462and do not need you to specify -maxmemory. 463 464 465JPEGTRAN 466 467jpegtran performs various useful transformations of JPEG files. 468It can translate the coded representation from one variant of JPEG to another, 469for example from baseline JPEG to progressive JPEG or vice versa. It can also 470perform some rearrangements of the image data, for example turning an image 471from landscape to portrait format by rotation. 472 473jpegtran works by rearranging the compressed data (DCT coefficients), without 474ever fully decoding the image. Therefore, its transformations are lossless: 475there is no image degradation at all, which would not be true if you used 476djpeg followed by cjpeg to accomplish the same conversion. But by the same 477token, jpegtran cannot perform lossy operations such as changing the image 478quality. 479 480jpegtran uses a command line syntax similar to cjpeg or djpeg. 481On Unix-like systems, you say: 482 jpegtran [switches] [inputfile] >outputfile 483On most non-Unix systems, you say: 484 jpegtran [switches] inputfile outputfile 485where both the input and output files are JPEG files. 486 487To specify the coded JPEG representation used in the output file, 488jpegtran accepts a subset of the switches recognized by cjpeg: 489 -optimize Perform optimization of entropy encoding parameters. 490 -progressive Create progressive JPEG file. 491 -arithmetic Use arithmetic coding. 492 -restart N Emit a JPEG restart marker every N MCU rows, or every 493 N MCU blocks if "B" is attached to the number. 494 -scans file Use the scan script given in the specified text file. 495See the previous discussion of cjpeg for more details about these switches. 496If you specify none of these switches, you get a plain baseline-JPEG output 497file. The quality setting and so forth are determined by the input file. 498 499The image can be losslessly transformed by giving one of these switches: 500 -flip horizontal Mirror image horizontally (left-right). 501 -flip vertical Mirror image vertically (top-bottom). 502 -rotate 90 Rotate image 90 degrees clockwise. 503 -rotate 180 Rotate image 180 degrees. 504 -rotate 270 Rotate image 270 degrees clockwise (or 90 ccw). 505 -transpose Transpose image (across UL-to-LR axis). 506 -transverse Transverse transpose (across UR-to-LL axis). 507 508The transpose transformation has no restrictions regarding image dimensions. 509The other transformations operate rather oddly if the image dimensions are not 510a multiple of the iMCU size (usually 8 or 16 pixels), because they can only 511transform complete blocks of DCT coefficient data in the desired way. 512 513jpegtran's default behavior when transforming an odd-size image is designed 514to preserve exact reversibility and mathematical consistency of the 515transformation set. As stated, transpose is able to flip the entire image 516area. Horizontal mirroring leaves any partial iMCU column at the right edge 517untouched, but is able to flip all rows of the image. Similarly, vertical 518mirroring leaves any partial iMCU row at the bottom edge untouched, but is 519able to flip all columns. The other transforms can be built up as sequences 520of transpose and flip operations; for consistency, their actions on edge 521pixels are defined to be the same as the end result of the corresponding 522transpose-and-flip sequence. 523 524For practical use, you may prefer to discard any untransformable edge pixels 525rather than having a strange-looking strip along the right and/or bottom edges 526of a transformed image. To do this, add the -trim switch: 527 -trim Drop non-transformable edge blocks. 528Obviously, a transformation with -trim is not reversible, so strictly speaking 529jpegtran with this switch is not lossless. Also, the expected mathematical 530equivalences between the transformations no longer hold. For example, 531"-rot 270 -trim" trims only the bottom edge, but "-rot 90 -trim" followed by 532"-rot 180 -trim" trims both edges. 533 534If you are only interested in perfect transformations, add the -perfect switch: 535 -perfect Fail with an error if the transformation is not 536 perfect. 537For example, you may want to do 538 jpegtran -rot 90 -perfect foo.jpg || djpeg foo.jpg | pnmflip -r90 | cjpeg 539to do a perfect rotation, if available, or an approximated one if not. 540 541This version of jpegtran also offers a lossless crop option, which discards 542data outside of a given image region but losslessly preserves what is inside. 543Like the rotate and flip transforms, lossless crop is restricted by the current 544JPEG format; the upper left corner of the selected region must fall on an iMCU 545boundary. If it doesn't, then it is silently moved up and/or left to the 546nearest iMCU boundary (the lower right corner is unchanged.) 547 548The image can be losslessly cropped by giving the switch: 549 -crop WxH+X+Y Crop to a rectangular region of width W and height H, 550 starting at point X,Y. 551 552Other not-strictly-lossless transformation switches are: 553 554 -grayscale Force grayscale output. 555This option discards the chrominance channels if the input image is YCbCr 556(ie, a standard color JPEG), resulting in a grayscale JPEG file. The 557luminance channel is preserved exactly, so this is a better method of reducing 558to grayscale than decompression, conversion, and recompression. This switch 559is particularly handy for fixing a monochrome picture that was mistakenly 560encoded as a color JPEG. (In such a case, the space savings from getting rid 561of the near-empty chroma channels won't be large; but the decoding time for 562a grayscale JPEG is substantially less than that for a color JPEG.) 563 564jpegtran also recognizes these switches that control what to do with "extra" 565markers, such as comment blocks: 566 -copy none Copy no extra markers from source file. This setting 567 suppresses all comments and other excess baggage 568 present in the source file. 569 -copy comments Copy only comment markers. This setting copies 570 comments from the source file but discards 571 any other data that is inessential for image display. 572 -copy all Copy all extra markers. This setting preserves 573 miscellaneous markers found in the source file, such 574 as JFIF thumbnails, Exif data, and Photoshop settings. 575 In some files, these extra markers can be sizable. 576The default behavior is -copy comments. (Note: in IJG releases v6 and v6a, 577jpegtran always did the equivalent of -copy none.) 578 579Additional switches recognized by jpegtran are: 580 -outfile filename 581 -maxmemory N 582 -verbose 583 -debug 584These work the same as in cjpeg or djpeg. 585 586 587THE COMMENT UTILITIES 588 589The JPEG standard allows "comment" (COM) blocks to occur within a JPEG file. 590Although the standard doesn't actually define what COM blocks are for, they 591are widely used to hold user-supplied text strings. This lets you add 592annotations, titles, index terms, etc to your JPEG files, and later retrieve 593them as text. COM blocks do not interfere with the image stored in the JPEG 594file. The maximum size of a COM block is 64K, but you can have as many of 595them as you like in one JPEG file. 596 597We provide two utility programs to display COM block contents and add COM 598blocks to a JPEG file. 599 600rdjpgcom searches a JPEG file and prints the contents of any COM blocks on 601standard output. The command line syntax is 602 rdjpgcom [-raw] [-verbose] [inputfilename] 603The switch "-raw" (or just "-r") causes rdjpgcom to output non-printable 604characters in JPEG comments. These characters are normally escaped for 605security reasons. 606The switch "-verbose" (or just "-v") causes rdjpgcom to also display the JPEG 607image dimensions. If you omit the input file name from the command line, 608the JPEG file is read from standard input. (This may not work on some 609operating systems, if binary data can't be read from stdin.) 610 611wrjpgcom adds a COM block, containing text you provide, to a JPEG file. 612Ordinarily, the COM block is added after any existing COM blocks, but you 613can delete the old COM blocks if you wish. wrjpgcom produces a new JPEG 614file; it does not modify the input file. DO NOT try to overwrite the input 615file by directing wrjpgcom's output back into it; on most systems this will 616just destroy your file. 617 618The command line syntax for wrjpgcom is similar to cjpeg's. On Unix-like 619systems, it is 620 wrjpgcom [switches] [inputfilename] 621The output file is written to standard output. The input file comes from 622the named file, or from standard input if no input file is named. 623 624On most non-Unix systems, the syntax is 625 wrjpgcom [switches] inputfilename outputfilename 626where both input and output file names must be given explicitly. 627 628wrjpgcom understands three switches: 629 -replace Delete any existing COM blocks from the file. 630 -comment "Comment text" Supply new COM text on command line. 631 -cfile name Read text for new COM block from named file. 632(Switch names can be abbreviated.) If you have only one line of comment text 633to add, you can provide it on the command line with -comment. The comment 634text must be surrounded with quotes so that it is treated as a single 635argument. Longer comments can be read from a text file. 636 637If you give neither -comment nor -cfile, then wrjpgcom will read the comment 638text from standard input. (In this case an input image file name MUST be 639supplied, so that the source JPEG file comes from somewhere else.) You can 640enter multiple lines, up to 64KB worth. Type an end-of-file indicator 641(usually control-D or control-Z) to terminate the comment text entry. 642 643wrjpgcom will not add a COM block if the provided comment string is empty. 644Therefore -replace -comment "" can be used to delete all COM blocks from a 645file. 646 647These utility programs do not depend on the IJG JPEG library. In 648particular, the source code for rdjpgcom is intended as an illustration of 649the minimum amount of code required to parse a JPEG file header correctly. 650
