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