1 libpng-manual.txt - A description on how to use and modify libpng
3 libpng version 1.6.25 - September 1, 2016
4 Updated and distributed by Glenn Randers-Pehrson
5 <glennrp at users.sourceforge.net>
6 Copyright (c) 1998-2016 Glenn Randers-Pehrson
8 This document is released under the libpng license.
9 For conditions of distribution and use, see the disclaimer
14 libpng versions 0.97, January 1998, through 1.6.25 - September 1, 2016
15 Updated and distributed by Glenn Randers-Pehrson
16 Copyright (c) 1998-2016 Glenn Randers-Pehrson
18 libpng 1.0 beta 6 - version 0.96 - May 28, 1997
19 Updated and distributed by Andreas Dilger
20 Copyright (c) 1996, 1997 Andreas Dilger
22 libpng 1.0 beta 2 - version 0.88 - January 26, 1996
23 For conditions of distribution and use, see copyright
24 notice in png.h. Copyright (c) 1995, 1996 Guy Eric
25 Schalnat, Group 42, Inc.
27 Updated/rewritten per request in the libpng FAQ
28 Copyright (c) 1995, 1996 Frank J. T. Wojcik
29 December 18, 1995 & January 20, 1996
38 VI. Modifying/Customizing libpng
40 VIII. Changes to Libpng from version 0.88
41 IX. Changes to Libpng from version 1.0.x to 1.2.x
42 X. Changes to Libpng from version 1.0.x/1.2.x to 1.4.x
43 XI. Changes to Libpng from version 1.4.x to 1.5.x
44 XII. Changes to Libpng from version 1.5.x to 1.6.x
45 XIII. Detecting libpng
46 XIV. Source code repository
48 XVI. Y2K Compliance in libpng
52 This file describes how to use and modify the PNG reference library
53 (known as libpng) for your own use. In addition to this
54 file, example.c is a good starting point for using the library, as
55 it is heavily commented and should include everything most people
56 will need. We assume that libpng is already installed; see the
57 INSTALL file for instructions on how to configure and install libpng.
59 For examples of libpng usage, see the files "example.c", "pngtest.c",
60 and the files in the "contrib" directory, all of which are included in
61 the libpng distribution.
63 Libpng was written as a companion to the PNG specification, as a way
64 of reducing the amount of time and effort it takes to support the PNG
65 file format in application programs.
67 The PNG specification (second edition), November 2003, is available as
68 a W3C Recommendation and as an ISO Standard (ISO/IEC 15948:2004 (E)) at
69 <http://www.w3.org/TR/2003/REC-PNG-20031110/
70 The W3C and ISO documents have identical technical content.
72 The PNG-1.2 specification is available at
73 <http://png-mng.sourceforge.net/pub/png/spec/1.2/>.
74 It is technically equivalent
75 to the PNG specification (second edition) but has some additional material.
77 The PNG-1.0 specification is available as RFC 2083
78 <http://png-mng.sourceforge.net/pub/png/spec/1.0/> and as a
79 W3C Recommendation <http://www.w3.org/TR/REC-png-961001>.
81 Some additional chunks are described in the special-purpose public chunks
82 documents at <http://www.libpng.org/pub/png/spec/register/>
85 about PNG, and the latest version of libpng, can be found at the PNG home
86 page, <http://www.libpng.org/pub/png/>.
88 Most users will not have to modify the library significantly; advanced
89 users may want to modify it more. All attempts were made to make it as
90 complete as possible, while keeping the code easy to understand.
91 Currently, this library only supports C. Support for other languages
94 Libpng has been designed to handle multiple sessions at one time,
95 to be easily modifiable, to be portable to the vast majority of
96 machines (ANSI, K&R, 16-, 32-, and 64-bit) available, and to be easy
97 to use. The ultimate goal of libpng is to promote the acceptance of
98 the PNG file format in whatever way possible. While there is still
99 work to be done (see the TODO file), libpng should cover the
100 majority of the needs of its users.
102 Libpng uses zlib for its compression and decompression of PNG files.
103 Further information about zlib, and the latest version of zlib, can
104 be found at the zlib home page, <http://zlib.net/>.
105 The zlib compression utility is a general purpose utility that is
106 useful for more than PNG files, and can be used without libpng.
107 See the documentation delivered with zlib for more details.
108 You can usually find the source files for the zlib utility wherever you
109 find the libpng source files.
111 Libpng is thread safe, provided the threads are using different
112 instances of the structures. Each thread should have its own
113 png_struct and png_info instances, and thus its own image.
114 Libpng does not protect itself against two threads using the
115 same instance of a structure.
119 There are two main structures that are important to libpng, png_struct
120 and png_info. Both are internal structures that are no longer exposed
121 in the libpng interface (as of libpng 1.5.0).
123 The png_info structure is designed to provide information about the
124 PNG file. At one time, the fields of png_info were intended to be
125 directly accessible to the user. However, this tended to cause problems
126 with applications using dynamically loaded libraries, and as a result
127 a set of interface functions for png_info (the png_get_*() and png_set_*()
128 functions) was developed, and direct access to the png_info fields was
131 The png_struct structure is the object used by the library to decode a
132 single image. As of 1.5.0 this structure is also not exposed.
134 Almost all libpng APIs require a pointer to a png_struct as the first argument.
135 Many (in particular the png_set and png_get APIs) also require a pointer
136 to png_info as the second argument. Some application visible macros
137 defined in png.h designed for basic data access (reading and writing
138 integers in the PNG format) don't take a png_info pointer, but it's almost
139 always safe to assume that a (png_struct*) has to be passed to call an API
142 You can have more than one png_info structure associated with an image,
143 as illustrated in pngtest.c, one for information valid prior to the
144 IDAT chunks and another (called "end_info" below) for things after them.
146 The png.h header file is an invaluable reference for programming with libpng.
147 And while I'm on the topic, make sure you include the libpng header file:
151 and also (as of libpng-1.5.0) the zlib header file, if you need it:
157 The png.h header file defines a number of integral types used by the
158 APIs. Most of these are fairly obvious; for example types corresponding
159 to integers of particular sizes and types for passing color values.
161 One exception is how non-integral numbers are handled. For application
162 convenience most APIs that take such numbers have C (double) arguments;
163 however, internally PNG, and libpng, use 32 bit signed integers and encode
164 the value by multiplying by 100,000. As of libpng 1.5.0 a convenience
165 macro PNG_FP_1 is defined in png.h along with a type (png_fixed_point)
166 which is simply (png_int_32).
168 All APIs that take (double) arguments also have a matching API that
169 takes the corresponding fixed point integer arguments. The fixed point
170 API has the same name as the floating point one with "_fixed" appended.
171 The actual range of values permitted in the APIs is frequently less than
172 the full range of (png_fixed_point) (-21474 to +21474). When APIs require
173 a non-negative argument the type is recorded as png_uint_32 above. Consult
174 the header file and the text below for more information.
176 Special care must be take with sCAL chunk handling because the chunk itself
177 uses non-integral values encoded as strings containing decimal floating point
178 numbers. See the comments in the header file.
182 The main header file function declarations are frequently protected by C
183 preprocessing directives of the form:
185 #ifdef PNG_feature_SUPPORTED
189 #ifdef PNG_feature_SUPPORTED
193 The library can be built without support for these APIs, although a
194 standard build will have all implemented APIs. Application programs
195 should check the feature macros before using an API for maximum
196 portability. From libpng 1.5.0 the feature macros set during the build
197 of libpng are recorded in the header file "pnglibconf.h" and this file
198 is always included by png.h.
200 If you don't need to change the library configuration from the default, skip to
201 the next section ("Reading").
203 Notice that some of the makefiles in the 'scripts' directory and (in 1.5.0) all
204 of the build project files in the 'projects' directory simply copy
205 scripts/pnglibconf.h.prebuilt to pnglibconf.h. This means that these build
206 systems do not permit easy auto-configuration of the library - they only
207 support the default configuration.
209 The easiest way to make minor changes to the libpng configuration when
210 auto-configuration is supported is to add definitions to the command line
211 using (typically) CPPFLAGS. For example:
213 CPPFLAGS=-DPNG_NO_FLOATING_ARITHMETIC
215 will change the internal libpng math implementation for gamma correction and
216 other arithmetic calculations to fixed point, avoiding the need for fast
217 floating point support. The result can be seen in the generated pnglibconf.h -
218 make sure it contains the changed feature macro setting.
220 If you need to make more extensive configuration changes - more than one or two
221 feature macro settings - you can either add -DPNG_USER_CONFIG to the build
222 command line and put a list of feature macro settings in pngusr.h or you can set
223 DFA_XTRA (a makefile variable) to a file containing the same information in the
224 form of 'option' settings.
226 A. Changing pnglibconf.h
228 A variety of methods exist to build libpng. Not all of these support
229 reconfiguration of pnglibconf.h. To reconfigure pnglibconf.h it must either be
230 rebuilt from scripts/pnglibconf.dfa using awk or it must be edited by hand.
232 Hand editing is achieved by copying scripts/pnglibconf.h.prebuilt to
233 pnglibconf.h and changing the lines defining the supported features, paying
234 very close attention to the 'option' information in scripts/pnglibconf.dfa
235 that describes those features and their requirements. This is easy to get
238 B. Configuration using DFA_XTRA
240 Rebuilding from pnglibconf.dfa is easy if a functioning 'awk', or a later
241 variant such as 'nawk' or 'gawk', is available. The configure build will
242 automatically find an appropriate awk and build pnglibconf.h.
243 The scripts/pnglibconf.mak file contains a set of make rules for doing the
244 same thing if configure is not used, and many of the makefiles in the scripts
245 directory use this approach.
247 When rebuilding simply write a new file containing changed options and set
248 DFA_XTRA to the name of this file. This causes the build to append the new file
249 to the end of scripts/pnglibconf.dfa. The pngusr.dfa file should contain lines
250 of the following forms:
254 This turns all optional features off. Include it at the start of pngusr.dfa to
255 make it easier to build a minimal configuration. You will need to turn at least
256 some features on afterward to enable either reading or writing code, or both.
261 Enable or disable a single feature. This will automatically enable other
262 features required by a feature that is turned on or disable other features that
263 require a feature which is turned off. Conflicting settings will cause an error
264 message to be emitted by awk.
266 setting feature default value
268 Changes the default value of setting 'feature' to 'value'. There are a small
269 number of settings listed at the top of pnglibconf.h, they are documented in the
270 source code. Most of these values have performance implications for the library
271 but most of them have no visible effect on the API. Some can also be overridden
274 This method of building a customized pnglibconf.h is illustrated in
275 contrib/pngminim/*. See the "$(PNGCONF):" target in the makefile and
276 pngusr.dfa in these directories.
278 C. Configuration using PNG_USER_CONFIG
280 If -DPNG_USER_CONFIG is added to the CPPFLAGS when pnglibconf.h is built,
281 the file pngusr.h will automatically be included before the options in
282 scripts/pnglibconf.dfa are processed. Your pngusr.h file should contain only
283 macro definitions turning features on or off or setting settings.
285 Apart from the global setting "everything = off" all the options listed above
286 can be set using macros in pngusr.h:
288 #define PNG_feature_SUPPORTED
294 #define PNG_NO_feature
300 #define PNG_feature value
304 setting feature default value
306 Notice that in both cases, pngusr.dfa and pngusr.h, the contents of the
307 pngusr file you supply override the contents of scripts/pnglibconf.dfa
309 If confusing or incomprehensible behavior results it is possible to
310 examine the intermediate file pnglibconf.dfn to find the full set of
311 dependency information for each setting and option. Simply locate the
312 feature in the file and read the C comments that precede it.
314 This method is also illustrated in the contrib/pngminim/* makefiles and
319 We'll now walk you through the possible functions to call when reading
320 in a PNG file sequentially, briefly explaining the syntax and purpose
321 of each one. See example.c and png.h for more detail. While
322 progressive reading is covered in the next section, you will still
323 need some of the functions discussed in this section to read a PNG
328 You will want to do the I/O initialization(*) before you get into libpng,
329 so if it doesn't work, you don't have much to undo. Of course, you
330 will also want to insure that you are, in fact, dealing with a PNG
331 file. Libpng provides a simple check to see if a file is a PNG file.
332 To use it, pass in the first 1 to 8 bytes of the file to the function
333 png_sig_cmp(), and it will return 0 (false) if the bytes match the
334 corresponding bytes of the PNG signature, or nonzero (true) otherwise.
335 Of course, the more bytes you pass in, the greater the accuracy of the
338 If you are intending to keep the file pointer open for use in libpng,
339 you must ensure you don't read more than 8 bytes from the beginning
340 of the file, and you also have to make a call to png_set_sig_bytes()
341 with the number of bytes you read from the beginning. Libpng will
342 then only check the bytes (if any) that your program didn't read.
344 (*): If you are not using the standard I/O functions, you will need
345 to replace them with custom functions. See the discussion under
348 FILE *fp = fopen(file_name, "rb");
354 if (fread(header, 1, number, fp) != number)
359 is_png = !png_sig_cmp(header, 0, number);
365 Next, png_struct and png_info need to be allocated and initialized. In
366 order to ensure that the size of these structures is correct even with a
367 dynamically linked libpng, there are functions to initialize and
368 allocate the structures. We also pass the library version, optional
369 pointers to error handling functions, and a pointer to a data struct for
370 use by the error functions, if necessary (the pointer and functions can
371 be NULL if the default error handlers are to be used). See the section
372 on Changes to Libpng below regarding the old initialization functions.
373 The structure allocation functions quietly return NULL if they fail to
374 create the structure, so your application should check for that.
376 png_structp png_ptr = png_create_read_struct
377 (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
378 user_error_fn, user_warning_fn);
383 png_infop info_ptr = png_create_info_struct(png_ptr);
387 png_destroy_read_struct(&png_ptr,
388 (png_infopp)NULL, (png_infopp)NULL);
392 If you want to use your own memory allocation routines,
393 use a libpng that was built with PNG_USER_MEM_SUPPORTED defined, and use
394 png_create_read_struct_2() instead of png_create_read_struct():
396 png_structp png_ptr = png_create_read_struct_2
397 (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
398 user_error_fn, user_warning_fn, (png_voidp)
399 user_mem_ptr, user_malloc_fn, user_free_fn);
401 The error handling routines passed to png_create_read_struct()
402 and the memory alloc/free routines passed to png_create_struct_2()
403 are only necessary if you are not using the libpng supplied error
404 handling and memory alloc/free functions.
406 When libpng encounters an error, it expects to longjmp back
407 to your routine. Therefore, you will need to call setjmp and pass
408 your png_jmpbuf(png_ptr). If you read the file from different
409 routines, you will need to update the longjmp buffer every time you enter
410 a new routine that will call a png_*() function.
412 See your documentation of setjmp/longjmp for your compiler for more
413 information on setjmp/longjmp. See the discussion on libpng error
414 handling in the Customizing Libpng section below for more information
415 on the libpng error handling. If an error occurs, and libpng longjmp's
416 back to your setjmp, you will want to call png_destroy_read_struct() to
419 if (setjmp(png_jmpbuf(png_ptr)))
421 png_destroy_read_struct(&png_ptr, &info_ptr,
427 Pass (png_infopp)NULL instead of &end_info if you didn't create
428 an end_info structure.
430 If you would rather avoid the complexity of setjmp/longjmp issues,
431 you can compile libpng with PNG_NO_SETJMP, in which case
432 errors will result in a call to PNG_ABORT() which defaults to abort().
434 You can #define PNG_ABORT() to a function that does something
435 more useful than abort(), as long as your function does not
438 Now you need to set up the input code. The default for libpng is to
439 use the C function fread(). If you use this, you will need to pass a
440 valid FILE * in the function png_init_io(). Be sure that the file is
441 opened in binary mode. If you wish to handle reading data in another
442 way, you need not call the png_init_io() function, but you must then
443 implement the libpng I/O methods discussed in the Customizing Libpng
446 png_init_io(png_ptr, fp);
448 If you had previously opened the file and read any of the signature from
449 the beginning in order to see if this was a PNG file, you need to let
450 libpng know that there are some bytes missing from the start of the file.
452 png_set_sig_bytes(png_ptr, number);
454 You can change the zlib compression buffer size to be used while
455 reading compressed data with
457 png_set_compression_buffer_size(png_ptr, buffer_size);
459 where the default size is 8192 bytes. Note that the buffer size
460 is changed immediately and the buffer is reallocated immediately,
461 instead of setting a flag to be acted upon later.
463 If you want CRC errors to be handled in a different manner than
466 png_set_crc_action(png_ptr, crit_action, ancil_action);
468 The values for png_set_crc_action() say how libpng is to handle CRC errors in
469 ancillary and critical chunks, and whether to use the data contained
470 therein. Note that it is impossible to "discard" data in a critical
473 Choices for (int) crit_action are
474 PNG_CRC_DEFAULT 0 error/quit
475 PNG_CRC_ERROR_QUIT 1 error/quit
476 PNG_CRC_WARN_USE 3 warn/use data
477 PNG_CRC_QUIET_USE 4 quiet/use data
478 PNG_CRC_NO_CHANGE 5 use the current value
480 Choices for (int) ancil_action are
481 PNG_CRC_DEFAULT 0 error/quit
482 PNG_CRC_ERROR_QUIT 1 error/quit
483 PNG_CRC_WARN_DISCARD 2 warn/discard data
484 PNG_CRC_WARN_USE 3 warn/use data
485 PNG_CRC_QUIET_USE 4 quiet/use data
486 PNG_CRC_NO_CHANGE 5 use the current value
488 Setting up callback code
490 You can set up a callback function to handle any unknown chunks in the
491 input stream. You must supply the function
493 read_chunk_callback(png_structp png_ptr,
494 png_unknown_chunkp chunk);
496 /* The unknown chunk structure contains your
497 chunk data, along with similar data for any other
504 /* Note that libpng has already taken care of
507 /* put your code here. Search for your chunk in the
508 unknown chunk structure, process it, and return one
511 return (-n); /* chunk had an error */
512 return (0); /* did not recognize */
513 return (n); /* success */
516 (You can give your function another name that you like instead of
517 "read_chunk_callback")
519 To inform libpng about your function, use
521 png_set_read_user_chunk_fn(png_ptr, user_chunk_ptr,
522 read_chunk_callback);
524 This names not only the callback function, but also a user pointer that
525 you can retrieve with
527 png_get_user_chunk_ptr(png_ptr);
529 If you call the png_set_read_user_chunk_fn() function, then all unknown
530 chunks which the callback does not handle will be saved when read. You can
531 cause them to be discarded by returning '1' ("handled") instead of '0'. This
532 behavior will change in libpng 1.7 and the default handling set by the
533 png_set_keep_unknown_chunks() function, described below, will be used when the
534 callback returns 0. If you want the existing behavior you should set the global
535 default to PNG_HANDLE_CHUNK_IF_SAFE now; this is compatible with all current
536 versions of libpng and with 1.7. Libpng 1.6 issues a warning if you keep the
537 default, or PNG_HANDLE_CHUNK_NEVER, and the callback returns 0.
539 At this point, you can set up a callback function that will be
540 called after each row has been read, which you can use to control
541 a progress meter or the like. It's demonstrated in pngtest.c.
542 You must supply a function
544 void read_row_callback(png_structp png_ptr,
545 png_uint_32 row, int pass);
547 /* put your code here */
550 (You can give it another name that you like instead of "read_row_callback")
552 To inform libpng about your function, use
554 png_set_read_status_fn(png_ptr, read_row_callback);
556 When this function is called the row has already been completely processed and
557 the 'row' and 'pass' refer to the next row to be handled. For the
558 non-interlaced case the row that was just handled is simply one less than the
559 passed in row number, and pass will always be 0. For the interlaced case the
560 same applies unless the row value is 0, in which case the row just handled was
561 the last one from one of the preceding passes. Because interlacing may skip a
562 pass you cannot be sure that the preceding pass is just 'pass-1'; if you really
563 need to know what the last pass is record (row,pass) from the callback and use
564 the last recorded value each time.
566 As with the user transform you can find the output row using the
567 PNG_ROW_FROM_PASS_ROW macro.
569 Unknown-chunk handling
571 Now you get to set the way the library processes unknown chunks in the
572 input PNG stream. Both known and unknown chunks will be read. Normal
573 behavior is that known chunks will be parsed into information in
574 various info_ptr members while unknown chunks will be discarded. This
575 behavior can be wasteful if your application will never use some known
576 chunk types. To change this, you can call:
578 png_set_keep_unknown_chunks(png_ptr, keep,
579 chunk_list, num_chunks);
581 keep - 0: default unknown chunk handling
582 1: ignore; do not keep
583 2: keep only if safe-to-copy
584 3: keep even if unsafe-to-copy
586 You can use these definitions:
587 PNG_HANDLE_CHUNK_AS_DEFAULT 0
588 PNG_HANDLE_CHUNK_NEVER 1
589 PNG_HANDLE_CHUNK_IF_SAFE 2
590 PNG_HANDLE_CHUNK_ALWAYS 3
592 chunk_list - list of chunks affected (a byte string,
593 five bytes per chunk, NULL or '\0' if
594 num_chunks is positive; ignored if
597 num_chunks - number of chunks affected; if 0, all
598 unknown chunks are affected. If positive,
599 only the chunks in the list are affected,
600 and if negative all unknown chunks and
601 all known chunks except for the IHDR,
602 PLTE, tRNS, IDAT, and IEND chunks are
605 Unknown chunks declared in this way will be saved as raw data onto a
606 list of png_unknown_chunk structures. If a chunk that is normally
607 known to libpng is named in the list, it will be handled as unknown,
608 according to the "keep" directive. If a chunk is named in successive
609 instances of png_set_keep_unknown_chunks(), the final instance will
610 take precedence. The IHDR and IEND chunks should not be named in
611 chunk_list; if they are, libpng will process them normally anyway.
612 If you know that your application will never make use of some particular
613 chunks, use PNG_HANDLE_CHUNK_NEVER (or 1) as demonstrated below.
615 Here is an example of the usage of png_set_keep_unknown_chunks(),
616 where the private "vpAg" chunk will later be processed by a user chunk
619 png_byte vpAg[5]={118, 112, 65, 103, (png_byte) '\0'};
621 #if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
622 png_byte unused_chunks[]=
624 104, 73, 83, 84, (png_byte) '\0', /* hIST */
625 105, 84, 88, 116, (png_byte) '\0', /* iTXt */
626 112, 67, 65, 76, (png_byte) '\0', /* pCAL */
627 115, 67, 65, 76, (png_byte) '\0', /* sCAL */
628 115, 80, 76, 84, (png_byte) '\0', /* sPLT */
629 116, 73, 77, 69, (png_byte) '\0', /* tIME */
635 #if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
636 /* ignore all unknown chunks
637 * (use global setting "2" for libpng16 and earlier):
639 png_set_keep_unknown_chunks(read_ptr, 2, NULL, 0);
641 /* except for vpAg: */
642 png_set_keep_unknown_chunks(read_ptr, 2, vpAg, 1);
644 /* also ignore unused known chunks: */
645 png_set_keep_unknown_chunks(read_ptr, 1, unused_chunks,
646 (int)(sizeof unused_chunks)/5);
651 The PNG specification allows the width and height of an image to be as
652 large as 2^31-1 (0x7fffffff), or about 2.147 billion rows and columns.
653 For safety, libpng imposes a default limit of 1 million rows and columns.
654 Larger images will be rejected immediately with a png_error() call. If
655 you wish to change these limits, you can use
657 png_set_user_limits(png_ptr, width_max, height_max);
659 to set your own limits (libpng may reject some very wide images
660 anyway because of potential buffer overflow conditions).
662 You should put this statement after you create the PNG structure and
663 before calling png_read_info(), png_read_png(), or png_process_data().
665 When writing a PNG datastream, put this statement before calling
666 png_write_info() or png_write_png().
668 If you need to retrieve the limits that are being applied, use
670 width_max = png_get_user_width_max(png_ptr);
671 height_max = png_get_user_height_max(png_ptr);
673 The PNG specification sets no limit on the number of ancillary chunks
674 allowed in a PNG datastream. By default, libpng imposes a limit of
675 a total of 1000 sPLT, tEXt, iTXt, zTXt, and unknown chunks to be stored.
676 If you have set up both info_ptr and end_info_ptr, the limit applies
677 separately to each. You can change the limit on the total number of such
678 chunks that will be stored, with
680 png_set_chunk_cache_max(png_ptr, user_chunk_cache_max);
682 where 0x7fffffffL means unlimited. You can retrieve this limit with
684 chunk_cache_max = png_get_chunk_cache_max(png_ptr);
686 Libpng imposes a limit of 8 Megabytes (8,000,000 bytes) on the amount of
687 memory that a compressed chunk other than IDAT can occupy, when decompressed.
688 You can change this limit with
690 png_set_chunk_malloc_max(png_ptr, user_chunk_malloc_max);
692 and you can retrieve the limit with
694 chunk_malloc_max = png_get_chunk_malloc_max(png_ptr);
696 Any chunks that would cause either of these limits to be exceeded will
699 Information about your system
701 If you intend to display the PNG or to incorporate it in other image data you
702 need to tell libpng information about your display or drawing surface so that
703 libpng can convert the values in the image to match the display.
705 From libpng-1.5.4 this information can be set before reading the PNG file
706 header. In earlier versions png_set_gamma() existed but behaved incorrectly if
707 called before the PNG file header had been read and png_set_alpha_mode() did not
710 If you need to support versions prior to libpng-1.5.4 test the version number
711 as illustrated below using "PNG_LIBPNG_VER >= 10504" and follow the procedures
712 described in the appropriate manual page.
714 You give libpng the encoding expected by your system expressed as a 'gamma'
715 value. You can also specify a default encoding for the PNG file in
716 case the required information is missing from the file. By default libpng
717 assumes that the PNG data matches your system, to keep this default call:
719 png_set_gamma(png_ptr, screen_gamma, output_gamma);
721 or you can use the fixed point equivalent:
723 png_set_gamma_fixed(png_ptr, PNG_FP_1*screen_gamma,
724 PNG_FP_1*output_gamma);
726 If you don't know the gamma for your system it is probably 2.2 - a good
727 approximation to the IEC standard for display systems (sRGB). If images are
728 too contrasty or washed out you got the value wrong - check your system
731 Many systems permit the system gamma to be changed via a lookup table in the
732 display driver, a few systems, including older Macs, change the response by
733 default. As of 1.5.4 three special values are available to handle common
736 PNG_DEFAULT_sRGB: Indicates that the system conforms to the
737 IEC 61966-2-1 standard. This matches almost
739 PNG_GAMMA_MAC_18: Indicates that the system is an older
740 (pre Mac OS 10.6) Apple Macintosh system with
741 the default settings.
742 PNG_GAMMA_LINEAR: Just the fixed point value for 1.0 - indicates
743 that the system expects data with no gamma
746 You would use the linear (unencoded) value if you need to process the pixel
747 values further because this avoids the need to decode and re-encode each
748 component value whenever arithmetic is performed. A lot of graphics software
749 uses linear values for this reason, often with higher precision component values
750 to preserve overall accuracy.
753 The output_gamma value expresses how to decode the output values, not how
754 they are encoded. The values used correspond to the normal numbers used to
755 describe the overall gamma of a computer display system; for example 2.2 for
756 an sRGB conformant system. The values are scaled by 100000 in the _fixed
757 version of the API (so 220000 for sRGB.)
759 The inverse of the value is always used to provide a default for the PNG file
760 encoding if it has no gAMA chunk and if png_set_gamma() has not been called
761 to override the PNG gamma information.
763 When the ALPHA_OPTIMIZED mode is selected the output gamma is used to encode
764 opaque pixels however pixels with lower alpha values are not encoded,
765 regardless of the output gamma setting.
767 When the standard Porter Duff handling is requested with mode 1 the output
768 encoding is set to be linear and the output_gamma value is only relevant
769 as a default for input data that has no gamma information. The linear output
770 encoding will be overridden if png_set_gamma() is called - the results may be
773 The following numbers are derived from the sRGB standard and the research
774 behind it. sRGB is defined to be approximated by a PNG gAMA chunk value of
775 0.45455 (1/2.2) for PNG. The value implicitly includes any viewing
776 correction required to take account of any differences in the color
777 environment of the original scene and the intended display environment; the
778 value expresses how to *decode* the image for display, not how the original
781 sRGB provides a peg for the PNG standard by defining a viewing environment.
782 sRGB itself, and earlier TV standards, actually use a more complex transform
783 (a linear portion then a gamma 2.4 power law) than PNG can express. (PNG is
784 limited to simple power laws.) By saying that an image for direct display on
785 an sRGB conformant system should be stored with a gAMA chunk value of 45455
786 (11.3.3.2 and 11.3.3.5 of the ISO PNG specification) the PNG specification
787 makes it possible to derive values for other display systems and
790 The Mac value is deduced from the sRGB based on an assumption that the actual
791 extra viewing correction used in early Mac display systems was implemented as
792 a power 1.45 lookup table.
794 Any system where a programmable lookup table is used or where the behavior of
795 the final display device characteristics can be changed requires system
796 specific code to obtain the current characteristic. However this can be
797 difficult and most PNG gamma correction only requires an approximate value.
799 By default, if png_set_alpha_mode() is not called, libpng assumes that all
800 values are unencoded, linear, values and that the output device also has a
801 linear characteristic. This is only very rarely correct - it is invariably
802 better to call png_set_alpha_mode() with PNG_DEFAULT_sRGB than rely on the
803 default if you don't know what the right answer is!
805 The special value PNG_GAMMA_MAC_18 indicates an older Mac system (pre Mac OS
806 10.6) which used a correction table to implement a somewhat lower gamma on an
807 otherwise sRGB system.
809 Both these values are reserved (not simple gamma values) in order to allow
810 more precise correction internally in the future.
812 NOTE: the values can be passed to either the fixed or floating
813 point APIs, but the floating point API will also accept floating point
816 The second thing you may need to tell libpng about is how your system handles
817 alpha channel information. Some, but not all, PNG files contain an alpha
818 channel. To display these files correctly you need to compose the data onto a
819 suitable background, as described in the PNG specification.
821 Libpng only supports composing onto a single color (using png_set_background;
822 see below). Otherwise you must do the composition yourself and, in this case,
823 you may need to call png_set_alpha_mode:
825 #if PNG_LIBPNG_VER >= 10504
826 png_set_alpha_mode(png_ptr, mode, screen_gamma);
828 png_set_gamma(png_ptr, screen_gamma, 1.0/screen_gamma);
831 The screen_gamma value is the same as the argument to png_set_gamma; however,
832 how it affects the output depends on the mode. png_set_alpha_mode() sets the
833 file gamma default to 1/screen_gamma, so normally you don't need to call
834 png_set_gamma. If you need different defaults call png_set_gamma() before
835 png_set_alpha_mode() - if you call it after it will override the settings made
836 by png_set_alpha_mode().
838 The mode is as follows:
840 PNG_ALPHA_PNG: The data is encoded according to the PNG
841 specification. Red, green and blue, or gray, components are
842 gamma encoded color values and are not premultiplied by the
843 alpha value. The alpha value is a linear measure of the
844 contribution of the pixel to the corresponding final output pixel.
846 You should normally use this format if you intend to perform
847 color correction on the color values; most, maybe all, color
848 correction software has no handling for the alpha channel and,
849 anyway, the math to handle pre-multiplied component values is
850 unnecessarily complex.
852 Before you do any arithmetic on the component values you need
853 to remove the gamma encoding and multiply out the alpha
854 channel. See the PNG specification for more detail. It is
855 important to note that when an image with an alpha channel is
856 scaled, linear encoded, pre-multiplied component values must
859 The remaining modes assume you don't need to do any further color correction or
860 that if you do, your color correction software knows all about alpha (it
861 probably doesn't!). They 'associate' the alpha with the color information by
862 storing color channel values that have been scaled by the alpha. The
863 advantage is that the color channels can be resampled (the image can be
864 scaled) in this form. The disadvantage is that normal practice is to store
865 linear, not (gamma) encoded, values and this requires 16-bit channels for
866 still images rather than the 8-bit channels that are just about sufficient if
867 gamma encoding is used. In addition all non-transparent pixel values,
868 including completely opaque ones, must be gamma encoded to produce the final
869 image. These are the 'STANDARD', 'ASSOCIATED' or 'PREMULTIPLIED' modes
870 described below (the latter being the two common names for associated alpha
871 color channels). Note that PNG files always contain non-associated color
872 channels; png_set_alpha_mode() with one of the modes causes the decoder to
873 convert the pixels to an associated form before returning them to your
876 Since it is not necessary to perform arithmetic on opaque color values so
877 long as they are not to be resampled and are in the final color space it is
878 possible to optimize the handling of alpha by storing the opaque pixels in
879 the PNG format (adjusted for the output color space) while storing partially
880 opaque pixels in the standard, linear, format. The accuracy required for
881 standard alpha composition is relatively low, because the pixels are
882 isolated, therefore typically the accuracy loss in storing 8-bit linear
883 values is acceptable. (This is not true if the alpha channel is used to
884 simulate transparency over large areas - use 16 bits or the PNG mode in
885 this case!) This is the 'OPTIMIZED' mode. For this mode a pixel is
886 treated as opaque only if the alpha value is equal to the maximum value.
888 PNG_ALPHA_STANDARD: The data libpng produces is encoded in the
889 standard way assumed by most correctly written graphics software.
890 The gamma encoding will be removed by libpng and the
891 linear component values will be pre-multiplied by the
894 With this format the final image must be re-encoded to
895 match the display gamma before the image is displayed.
896 If your system doesn't do that, yet still seems to
897 perform arithmetic on the pixels without decoding them,
898 it is broken - check out the modes below.
900 With PNG_ALPHA_STANDARD libpng always produces linear
901 component values, whatever screen_gamma you supply. The
902 screen_gamma value is, however, used as a default for
903 the file gamma if the PNG file has no gamma information.
905 If you call png_set_gamma() after png_set_alpha_mode() you
906 will override the linear encoding. Instead the
907 pre-multiplied pixel values will be gamma encoded but
908 the alpha channel will still be linear. This may
909 actually match the requirements of some broken software,
912 While linear 8-bit data is often used it has
913 insufficient precision for any image with a reasonable
914 dynamic range. To avoid problems, and if your software
915 supports it, use png_set_expand_16() to force all
916 components to 16 bits.
918 PNG_ALPHA_OPTIMIZED: This mode is the same as PNG_ALPHA_STANDARD
919 except that completely opaque pixels are gamma encoded according to
920 the screen_gamma value. Pixels with alpha less than 1.0
921 will still have linear components.
923 Use this format if you have control over your
924 compositing software and so don't do other arithmetic
925 (such as scaling) on the data you get from libpng. Your
926 compositing software can simply copy opaque pixels to
927 the output but still has linear values for the
930 In normal compositing, where the alpha channel encodes
931 partial pixel coverage (as opposed to broad area
932 translucency), the inaccuracies of the 8-bit
933 representation of non-opaque pixels are irrelevant.
935 You can also try this format if your software is broken;
936 it might look better.
938 PNG_ALPHA_BROKEN: This is PNG_ALPHA_STANDARD; however, all component
939 values, including the alpha channel are gamma encoded. This is
940 broken because, in practice, no implementation that uses this choice
941 correctly undoes the encoding before handling alpha composition. Use this
942 choice only if other serious errors in the software or hardware you use
943 mandate it. In most cases of broken software or hardware the bug in the
944 final display manifests as a subtle halo around composited parts of the
945 image. You may not even perceive this as a halo; the composited part of
946 the image may simply appear separate from the background, as though it had
947 been cut out of paper and pasted on afterward.
949 If you don't have to deal with bugs in software or hardware, or if you can fix
950 them, there are three recommended ways of using png_set_alpha_mode():
952 png_set_alpha_mode(png_ptr, PNG_ALPHA_PNG,
955 You can do color correction on the result (libpng does not currently
956 support color correction internally). When you handle the alpha channel
957 you need to undo the gamma encoding and multiply out the alpha.
959 png_set_alpha_mode(png_ptr, PNG_ALPHA_STANDARD,
961 png_set_expand_16(png_ptr);
963 If you are using the high level interface, don't call png_set_expand_16();
964 instead pass PNG_TRANSFORM_EXPAND_16 to the interface.
966 With this mode you can't do color correction, but you can do arithmetic,
967 including composition and scaling, on the data without further processing.
969 png_set_alpha_mode(png_ptr, PNG_ALPHA_OPTIMIZED,
972 You can avoid the expansion to 16-bit components with this mode, but you
973 lose the ability to scale the image or perform other linear arithmetic.
974 All you can do is compose the result onto a matching output. Since this
975 mode is libpng-specific you also need to write your own composition
978 The following are examples of calls to png_set_alpha_mode to achieve the
979 required overall gamma correction and, where necessary, alpha
982 png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_DEFAULT_sRGB);
984 This is the default libpng handling of the alpha channel - it is not
985 pre-multiplied into the color components. In addition the call states
986 that the output is for a sRGB system and causes all PNG files without gAMA
987 chunks to be assumed to be encoded using sRGB.
989 png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_GAMMA_MAC);
991 In this case the output is assumed to be something like an sRGB conformant
992 display preceeded by a power-law lookup table of power 1.45. This is how
993 early Mac systems behaved.
995 png_set_alpha_mode(pp, PNG_ALPHA_STANDARD, PNG_GAMMA_LINEAR);
997 This is the classic Jim Blinn approach and will work in academic
998 environments where everything is done by the book. It has the shortcoming
999 of assuming that input PNG data with no gamma information is linear - this
1000 is unlikely to be correct unless the PNG files where generated locally.
1001 Most of the time the output precision will be so low as to show
1002 significant banding in dark areas of the image.
1004 png_set_expand_16(pp);
1005 png_set_alpha_mode(pp, PNG_ALPHA_STANDARD, PNG_DEFAULT_sRGB);
1007 This is a somewhat more realistic Jim Blinn inspired approach. PNG files
1008 are assumed to have the sRGB encoding if not marked with a gamma value and
1009 the output is always 16 bits per component. This permits accurate scaling
1010 and processing of the data. If you know that your input PNG files were
1011 generated locally you might need to replace PNG_DEFAULT_sRGB with the
1012 correct value for your system.
1014 png_set_alpha_mode(pp, PNG_ALPHA_OPTIMIZED, PNG_DEFAULT_sRGB);
1016 If you just need to composite the PNG image onto an existing background
1017 and if you control the code that does this you can use the optimization
1018 setting. In this case you just copy completely opaque pixels to the
1019 output. For pixels that are not completely transparent (you just skip
1020 those) you do the composition math using png_composite or png_composite_16
1021 below then encode the resultant 8-bit or 16-bit values to match the output
1026 If neither the PNG nor the standard linear encoding work for you because
1027 of the software or hardware you use then you have a big problem. The PNG
1028 case will probably result in halos around the image. The linear encoding
1029 will probably result in a washed out, too bright, image (it's actually too
1030 contrasty.) Try the ALPHA_OPTIMIZED mode above - this will probably
1031 substantially reduce the halos. Alternatively try:
1033 png_set_alpha_mode(pp, PNG_ALPHA_BROKEN, PNG_DEFAULT_sRGB);
1035 This option will also reduce the halos, but there will be slight dark
1036 halos round the opaque parts of the image where the background is light.
1037 In the OPTIMIZED mode the halos will be light halos where the background
1038 is dark. Take your pick - the halos are unavoidable unless you can get
1039 your hardware/software fixed! (The OPTIMIZED approach is slightly
1042 When the default gamma of PNG files doesn't match the output gamma.
1043 If you have PNG files with no gamma information png_set_alpha_mode allows
1044 you to provide a default gamma, but it also sets the ouput gamma to the
1045 matching value. If you know your PNG files have a gamma that doesn't
1046 match the output you can take advantage of the fact that
1047 png_set_alpha_mode always sets the output gamma but only sets the PNG
1048 default if it is not already set:
1050 png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_DEFAULT_sRGB);
1051 png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_GAMMA_MAC);
1053 The first call sets both the default and the output gamma values, the
1054 second call overrides the output gamma without changing the default. This
1055 is easier than achieving the same effect with png_set_gamma. You must use
1056 PNG_ALPHA_PNG for the first call - internal checking in png_set_alpha will
1057 fire if more than one call to png_set_alpha_mode and png_set_background is
1058 made in the same read operation, however multiple calls with PNG_ALPHA_PNG
1061 If you don't need, or can't handle, the alpha channel you can call
1062 png_set_background() to remove it by compositing against a fixed color. Don't
1063 call png_set_strip_alpha() to do this - it will leave spurious pixel values in
1064 transparent parts of this image.
1066 png_set_background(png_ptr, &background_color,
1067 PNG_BACKGROUND_GAMMA_SCREEN, 0, 1);
1069 The background_color is an RGB or grayscale value according to the data format
1070 libpng will produce for you. Because you don't yet know the format of the PNG
1071 file, if you call png_set_background at this point you must arrange for the
1072 format produced by libpng to always have 8-bit or 16-bit components and then
1073 store the color as an 8-bit or 16-bit color as appropriate. The color contains
1074 separate gray and RGB component values, so you can let libpng produce gray or
1075 RGB output according to the input format, but low bit depth grayscale images
1076 must always be converted to at least 8-bit format. (Even though low bit depth
1077 grayscale images can't have an alpha channel they can have a transparent
1080 You set the transforms you need later, either as flags to the high level
1081 interface or libpng API calls for the low level interface. For reference the
1082 settings and API calls required are:
1085 PNG_TRANSFORM_SCALE_16 | PNG_EXPAND
1086 png_set_expand(png_ptr); png_set_scale_16(png_ptr);
1088 If you must get exactly the same inaccurate results
1089 produced by default in versions prior to libpng-1.5.4,
1090 use PNG_TRANSFORM_STRIP_16 and png_set_strip_16(png_ptr)
1094 PNG_TRANSFORM_EXPAND_16
1095 png_set_expand_16(png_ptr);
1097 In either case palette image data will be expanded to RGB. If you just want
1098 color data you can add PNG_TRANSFORM_GRAY_TO_RGB or png_set_gray_to_rgb(png_ptr)
1101 Calling png_set_background before the PNG file header is read will not work
1102 prior to libpng-1.5.4. Because the failure may result in unexpected warnings or
1103 errors it is therefore much safer to call png_set_background after the head has
1104 been read. Unfortunately this means that prior to libpng-1.5.4 it cannot be
1105 used with the high level interface.
1107 The high-level read interface
1109 At this point there are two ways to proceed; through the high-level
1110 read interface, or through a sequence of low-level read operations.
1111 You can use the high-level interface if (a) you are willing to read
1112 the entire image into memory, and (b) the input transformations
1113 you want to do are limited to the following set:
1115 PNG_TRANSFORM_IDENTITY No transformation
1116 PNG_TRANSFORM_SCALE_16 Strip 16-bit samples to
1118 PNG_TRANSFORM_STRIP_16 Chop 16-bit samples to
1119 8-bit less accurately
1120 PNG_TRANSFORM_STRIP_ALPHA Discard the alpha channel
1121 PNG_TRANSFORM_PACKING Expand 1, 2 and 4-bit
1123 PNG_TRANSFORM_PACKSWAP Change order of packed
1125 PNG_TRANSFORM_EXPAND Perform set_expand()
1126 PNG_TRANSFORM_INVERT_MONO Invert monochrome images
1127 PNG_TRANSFORM_SHIFT Normalize pixels to the
1129 PNG_TRANSFORM_BGR Flip RGB to BGR, RGBA
1131 PNG_TRANSFORM_SWAP_ALPHA Flip RGBA to ARGB or GA
1133 PNG_TRANSFORM_INVERT_ALPHA Change alpha from opacity
1135 PNG_TRANSFORM_SWAP_ENDIAN Byte-swap 16-bit samples
1136 PNG_TRANSFORM_GRAY_TO_RGB Expand grayscale samples
1137 to RGB (or GA to RGBA)
1138 PNG_TRANSFORM_EXPAND_16 Expand samples to 16 bits
1140 (This excludes setting a background color, doing gamma transformation,
1141 quantizing, and setting filler.) If this is the case, simply do this:
1143 png_read_png(png_ptr, info_ptr, png_transforms, NULL)
1145 where png_transforms is an integer containing the bitwise OR of some
1146 set of transformation flags. This call is equivalent to png_read_info(),
1147 followed the set of transformations indicated by the transform mask,
1148 then png_read_image(), and finally png_read_end().
1150 (The final parameter of this call is not yet used. Someday it might point
1151 to transformation parameters required by some future input transform.)
1153 You must use png_transforms and not call any png_set_transform() functions
1154 when you use png_read_png().
1156 After you have called png_read_png(), you can retrieve the image data
1159 row_pointers = png_get_rows(png_ptr, info_ptr);
1161 where row_pointers is an array of pointers to the pixel data for each row:
1163 png_bytep row_pointers[height];
1165 If you know your image size and pixel size ahead of time, you can allocate
1166 row_pointers prior to calling png_read_png() with
1168 if (height > PNG_UINT_32_MAX/(sizeof (png_byte)))
1170 "Image is too tall to process in memory");
1172 if (width > PNG_UINT_32_MAX/pixel_size)
1174 "Image is too wide to process in memory");
1176 row_pointers = png_malloc(png_ptr,
1177 height*(sizeof (png_bytep)));
1179 for (int i=0; i<height, i++)
1180 row_pointers[i]=NULL; /* security precaution */
1182 for (int i=0; i<height, i++)
1183 row_pointers[i]=png_malloc(png_ptr,
1186 png_set_rows(png_ptr, info_ptr, &row_pointers);
1188 Alternatively you could allocate your image in one big block and define
1189 row_pointers[i] to point into the proper places in your block.
1191 If you use png_set_rows(), the application is responsible for freeing
1192 row_pointers (and row_pointers[i], if they were separately allocated).
1194 If you don't allocate row_pointers ahead of time, png_read_png() will
1195 do it, and it'll be free'ed by libpng when you call png_destroy_*().
1197 The low-level read interface
1199 If you are going the low-level route, you are now ready to read all
1200 the file information up to the actual image data. You do this with a
1201 call to png_read_info().
1203 png_read_info(png_ptr, info_ptr);
1205 This will process all chunks up to but not including the image data.
1207 This also copies some of the data from the PNG file into the decode structure
1208 for use in later transformations. Important information copied in is:
1210 1) The PNG file gamma from the gAMA chunk. This overwrites the default value
1211 provided by an earlier call to png_set_gamma or png_set_alpha_mode.
1213 2) Prior to libpng-1.5.4 the background color from a bKGd chunk. This
1214 damages the information provided by an earlier call to png_set_background
1215 resulting in unexpected behavior. Libpng-1.5.4 no longer does this.
1217 3) The number of significant bits in each component value. Libpng uses this to
1218 optimize gamma handling by reducing the internal lookup table sizes.
1220 4) The transparent color information from a tRNS chunk. This can be modified by
1221 a later call to png_set_tRNS.
1223 Querying the info structure
1225 Functions are used to get the information from the info_ptr once it
1226 has been read. Note that these fields may not be completely filled
1227 in until png_read_end() has read the chunk data following the image.
1229 png_get_IHDR(png_ptr, info_ptr, &width, &height,
1230 &bit_depth, &color_type, &interlace_type,
1231 &compression_type, &filter_method);
1233 width - holds the width of the image
1234 in pixels (up to 2^31).
1236 height - holds the height of the image
1237 in pixels (up to 2^31).
1239 bit_depth - holds the bit depth of one of the
1240 image channels. (valid values are
1241 1, 2, 4, 8, 16 and depend also on
1242 the color_type. See also
1243 significant bits (sBIT) below).
1245 color_type - describes which color/alpha channels
1248 (bit depths 1, 2, 4, 8, 16)
1249 PNG_COLOR_TYPE_GRAY_ALPHA
1251 PNG_COLOR_TYPE_PALETTE
1252 (bit depths 1, 2, 4, 8)
1255 PNG_COLOR_TYPE_RGB_ALPHA
1258 PNG_COLOR_MASK_PALETTE
1259 PNG_COLOR_MASK_COLOR
1260 PNG_COLOR_MASK_ALPHA
1262 interlace_type - (PNG_INTERLACE_NONE or
1263 PNG_INTERLACE_ADAM7)
1265 compression_type - (must be PNG_COMPRESSION_TYPE_BASE
1268 filter_method - (must be PNG_FILTER_TYPE_BASE
1269 for PNG 1.0, and can also be
1270 PNG_INTRAPIXEL_DIFFERENCING if
1271 the PNG datastream is embedded in
1272 a MNG-1.0 datastream)
1274 Any of width, height, color_type, bit_depth,
1275 interlace_type, compression_type, or filter_method can
1276 be NULL if you are not interested in their values.
1278 Note that png_get_IHDR() returns 32-bit data into
1279 the application's width and height variables.
1280 This is an unsafe situation if these are not png_uint_32
1281 variables. In such situations, the
1282 png_get_image_width() and png_get_image_height()
1283 functions described below are safer.
1285 width = png_get_image_width(png_ptr,
1288 height = png_get_image_height(png_ptr,
1291 bit_depth = png_get_bit_depth(png_ptr,
1294 color_type = png_get_color_type(png_ptr,
1297 interlace_type = png_get_interlace_type(png_ptr,
1300 compression_type = png_get_compression_type(png_ptr,
1303 filter_method = png_get_filter_type(png_ptr,
1306 channels = png_get_channels(png_ptr, info_ptr);
1308 channels - number of channels of info for the
1309 color type (valid values are 1 (GRAY,
1310 PALETTE), 2 (GRAY_ALPHA), 3 (RGB),
1311 4 (RGB_ALPHA or RGB + filler byte))
1313 rowbytes = png_get_rowbytes(png_ptr, info_ptr);
1315 rowbytes - number of bytes needed to hold a row
1317 signature = png_get_signature(png_ptr, info_ptr);
1319 signature - holds the signature read from the
1320 file (if any). The data is kept in
1321 the same offset it would be if the
1322 whole signature were read (i.e. if an
1323 application had already read in 4
1324 bytes of signature before starting
1325 libpng, the remaining 4 bytes would
1326 be in signature[4] through signature[7]
1327 (see png_set_sig_bytes())).
1329 These are also important, but their validity depends on whether the chunk
1330 has been read. The png_get_valid(png_ptr, info_ptr, PNG_INFO_<chunk>) and
1331 png_get_<chunk>(png_ptr, info_ptr, ...) functions return non-zero if the
1332 data has been read, or zero if it is missing. The parameters to the
1333 png_get_<chunk> are set directly if they are simple data types, or a
1334 pointer into the info_ptr is returned for any complex types.
1336 The colorspace data from gAMA, cHRM, sRGB, iCCP, and sBIT chunks
1337 is simply returned to give the application information about how the
1338 image was encoded. Libpng itself only does transformations using the file
1339 gamma when combining semitransparent pixels with the background color, and,
1340 since libpng-1.6.0, when converting between 8-bit sRGB and 16-bit linear pixels
1341 within the simplified API. Libpng also uses the file gamma when converting
1342 RGB to gray, beginning with libpng-1.0.5, if the application calls
1343 png_set_rgb_to_gray()).
1345 png_get_PLTE(png_ptr, info_ptr, &palette,
1348 palette - the palette for the file
1349 (array of png_color)
1351 num_palette - number of entries in the palette
1353 png_get_gAMA(png_ptr, info_ptr, &file_gamma);
1354 png_get_gAMA_fixed(png_ptr, info_ptr, &int_file_gamma);
1356 file_gamma - the gamma at which the file is
1357 written (PNG_INFO_gAMA)
1359 int_file_gamma - 100,000 times the gamma at which the
1362 png_get_cHRM(png_ptr, info_ptr, &white_x, &white_y, &red_x,
1363 &red_y, &green_x, &green_y, &blue_x, &blue_y)
1364 png_get_cHRM_XYZ(png_ptr, info_ptr, &red_X, &red_Y, &red_Z,
1365 &green_X, &green_Y, &green_Z, &blue_X, &blue_Y,
1367 png_get_cHRM_fixed(png_ptr, info_ptr, &int_white_x,
1368 &int_white_y, &int_red_x, &int_red_y,
1369 &int_green_x, &int_green_y, &int_blue_x,
1371 png_get_cHRM_XYZ_fixed(png_ptr, info_ptr, &int_red_X, &int_red_Y,
1372 &int_red_Z, &int_green_X, &int_green_Y,
1373 &int_green_Z, &int_blue_X, &int_blue_Y,
1376 {white,red,green,blue}_{x,y}
1377 A color space encoding specified using the
1378 chromaticities of the end points and the
1379 white point. (PNG_INFO_cHRM)
1381 {red,green,blue}_{X,Y,Z}
1382 A color space encoding specified using the
1383 encoding end points - the CIE tristimulus
1384 specification of the intended color of the red,
1385 green and blue channels in the PNG RGB data.
1386 The white point is simply the sum of the three
1387 end points. (PNG_INFO_cHRM)
1389 png_get_sRGB(png_ptr, info_ptr, &srgb_intent);
1391 srgb_intent - the rendering intent (PNG_INFO_sRGB)
1392 The presence of the sRGB chunk
1393 means that the pixel data is in the
1394 sRGB color space. This chunk also
1395 implies specific values of gAMA and
1398 png_get_iCCP(png_ptr, info_ptr, &name,
1399 &compression_type, &profile, &proflen);
1401 name - The profile name.
1403 compression_type - The compression type; always
1404 PNG_COMPRESSION_TYPE_BASE for PNG 1.0.
1405 You may give NULL to this argument to
1408 profile - International Color Consortium color
1409 profile data. May contain NULs.
1411 proflen - length of profile data in bytes.
1413 png_get_sBIT(png_ptr, info_ptr, &sig_bit);
1415 sig_bit - the number of significant bits for
1416 (PNG_INFO_sBIT) each of the gray,
1417 red, green, and blue channels,
1418 whichever are appropriate for the
1419 given color type (png_color_16)
1421 png_get_tRNS(png_ptr, info_ptr, &trans_alpha,
1422 &num_trans, &trans_color);
1424 trans_alpha - array of alpha (transparency)
1425 entries for palette (PNG_INFO_tRNS)
1427 num_trans - number of transparent entries
1430 trans_color - graylevel or color sample values of
1431 the single transparent color for
1432 non-paletted images (PNG_INFO_tRNS)
1434 png_get_hIST(png_ptr, info_ptr, &hist);
1437 hist - histogram of palette (array of
1440 png_get_tIME(png_ptr, info_ptr, &mod_time);
1442 mod_time - time image was last modified
1445 png_get_bKGD(png_ptr, info_ptr, &background);
1447 background - background color (of type
1448 png_color_16p) (PNG_VALID_bKGD)
1449 valid 16-bit red, green and blue
1450 values, regardless of color_type
1452 num_comments = png_get_text(png_ptr, info_ptr,
1453 &text_ptr, &num_text);
1455 num_comments - number of comments
1457 text_ptr - array of png_text holding image
1460 text_ptr[i].compression - type of compression used
1461 on "text" PNG_TEXT_COMPRESSION_NONE
1462 PNG_TEXT_COMPRESSION_zTXt
1463 PNG_ITXT_COMPRESSION_NONE
1464 PNG_ITXT_COMPRESSION_zTXt
1466 text_ptr[i].key - keyword for comment. Must contain
1469 text_ptr[i].text - text comments for current
1470 keyword. Can be empty.
1472 text_ptr[i].text_length - length of text string,
1473 after decompression, 0 for iTXt
1475 text_ptr[i].itxt_length - length of itxt string,
1476 after decompression, 0 for tEXt/zTXt
1478 text_ptr[i].lang - language of comment (empty
1479 string for unknown).
1481 text_ptr[i].lang_key - keyword in UTF-8
1482 (empty string for unknown).
1484 Note that the itxt_length, lang, and lang_key
1485 members of the text_ptr structure only exist when the
1486 library is built with iTXt chunk support. Prior to
1487 libpng-1.4.0 the library was built by default without
1488 iTXt support. Also note that when iTXt is supported,
1489 they contain NULL pointers when the "compression"
1490 field contains PNG_TEXT_COMPRESSION_NONE or
1491 PNG_TEXT_COMPRESSION_zTXt.
1493 num_text - number of comments (same as
1494 num_comments; you can put NULL here
1495 to avoid the duplication)
1497 Note while png_set_text() will accept text, language,
1498 and translated keywords that can be NULL pointers, the
1499 structure returned by png_get_text will always contain
1500 regular zero-terminated C strings. They might be
1501 empty strings but they will never be NULL pointers.
1503 num_spalettes = png_get_sPLT(png_ptr, info_ptr,
1506 num_spalettes - number of sPLT chunks read.
1508 palette_ptr - array of palette structures holding
1509 contents of one or more sPLT chunks
1512 png_get_oFFs(png_ptr, info_ptr, &offset_x, &offset_y,
1515 offset_x - positive offset from the left edge
1516 of the screen (can be negative)
1518 offset_y - positive offset from the top edge
1519 of the screen (can be negative)
1521 unit_type - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER
1523 png_get_pHYs(png_ptr, info_ptr, &res_x, &res_y,
1526 res_x - pixels/unit physical resolution in
1529 res_y - pixels/unit physical resolution in
1532 unit_type - PNG_RESOLUTION_UNKNOWN,
1533 PNG_RESOLUTION_METER
1535 png_get_sCAL(png_ptr, info_ptr, &unit, &width,
1538 unit - physical scale units (an integer)
1540 width - width of a pixel in physical scale units
1542 height - height of a pixel in physical scale units
1543 (width and height are doubles)
1545 png_get_sCAL_s(png_ptr, info_ptr, &unit, &width,
1548 unit - physical scale units (an integer)
1550 width - width of a pixel in physical scale units
1551 (expressed as a string)
1553 height - height of a pixel in physical scale units
1554 (width and height are strings like "2.54")
1556 num_unknown_chunks = png_get_unknown_chunks(png_ptr,
1557 info_ptr, &unknowns)
1559 unknowns - array of png_unknown_chunk
1560 structures holding unknown chunks
1562 unknowns[i].name - name of unknown chunk
1564 unknowns[i].data - data of unknown chunk
1566 unknowns[i].size - size of unknown chunk's data
1568 unknowns[i].location - position of chunk in file
1570 The value of "i" corresponds to the order in which the
1571 chunks were read from the PNG file or inserted with the
1572 png_set_unknown_chunks() function.
1574 The value of "location" is a bitwise "or" of
1576 PNG_HAVE_IHDR (0x01)
1577 PNG_HAVE_PLTE (0x02)
1578 PNG_AFTER_IDAT (0x08)
1580 The data from the pHYs chunk can be retrieved in several convenient
1583 res_x = png_get_x_pixels_per_meter(png_ptr,
1586 res_y = png_get_y_pixels_per_meter(png_ptr,
1589 res_x_and_y = png_get_pixels_per_meter(png_ptr,
1592 res_x = png_get_x_pixels_per_inch(png_ptr,
1595 res_y = png_get_y_pixels_per_inch(png_ptr,
1598 res_x_and_y = png_get_pixels_per_inch(png_ptr,
1601 aspect_ratio = png_get_pixel_aspect_ratio(png_ptr,
1604 Each of these returns 0 [signifying "unknown"] if
1605 the data is not present or if res_x is 0;
1606 res_x_and_y is 0 if res_x != res_y
1608 Note that because of the way the resolutions are
1609 stored internally, the inch conversions won't
1610 come out to exactly even number. For example,
1611 72 dpi is stored as 0.28346 pixels/meter, and
1612 when this is retrieved it is 71.9988 dpi, so
1613 be sure to round the returned value appropriately
1614 if you want to display a reasonable-looking result.
1616 The data from the oFFs chunk can be retrieved in several convenient
1619 x_offset = png_get_x_offset_microns(png_ptr, info_ptr);
1621 y_offset = png_get_y_offset_microns(png_ptr, info_ptr);
1623 x_offset = png_get_x_offset_inches(png_ptr, info_ptr);
1625 y_offset = png_get_y_offset_inches(png_ptr, info_ptr);
1627 Each of these returns 0 [signifying "unknown" if both
1628 x and y are 0] if the data is not present or if the
1629 chunk is present but the unit is the pixel. The
1630 remark about inexact inch conversions applies here
1631 as well, because a value in inches can't always be
1632 converted to microns and back without some loss
1635 For more information, see the
1636 PNG specification for chunk contents. Be careful with trusting
1637 rowbytes, as some of the transformations could increase the space
1638 needed to hold a row (expand, filler, gray_to_rgb, etc.).
1639 See png_read_update_info(), below.
1641 A quick word about text_ptr and num_text. PNG stores comments in
1642 keyword/text pairs, one pair per chunk, with no limit on the number
1643 of text chunks, and a 2^31 byte limit on their size. While there are
1644 suggested keywords, there is no requirement to restrict the use to these
1645 strings. It is strongly suggested that keywords and text be sensible
1646 to humans (that's the point), so don't use abbreviations. Non-printing
1647 symbols are not allowed. See the PNG specification for more details.
1648 There is also no requirement to have text after the keyword.
1650 Keywords should be limited to 79 Latin-1 characters without leading or
1651 trailing spaces, but non-consecutive spaces are allowed within the
1652 keyword. It is possible to have the same keyword any number of times.
1653 The text_ptr is an array of png_text structures, each holding a
1654 pointer to a language string, a pointer to a keyword and a pointer to
1655 a text string. The text string, language code, and translated
1656 keyword may be empty or NULL pointers. The keyword/text
1657 pairs are put into the array in the order that they are received.
1658 However, some or all of the text chunks may be after the image, so, to
1659 make sure you have read all the text chunks, don't mess with these
1660 until after you read the stuff after the image. This will be
1661 mentioned again below in the discussion that goes with png_read_end().
1663 Input transformations
1665 After you've read the header information, you can set up the library
1666 to handle any special transformations of the image data. The various
1667 ways to transform the data will be described in the order that they
1668 should occur. This is important, as some of these change the color
1669 type and/or bit depth of the data, and some others only work on
1670 certain color types and bit depths.
1672 Transformations you request are ignored if they don't have any meaning for a
1673 particular input data format. However some transformations can have an effect
1674 as a result of a previous transformation. If you specify a contradictory set of
1675 transformations, for example both adding and removing the alpha channel, you
1676 cannot predict the final result.
1678 The color used for the transparency values should be supplied in the same
1679 format/depth as the current image data. It is stored in the same format/depth
1680 as the image data in a tRNS chunk, so this is what libpng expects for this data.
1682 The color used for the background value depends on the need_expand argument as
1685 Data will be decoded into the supplied row buffers packed into bytes
1686 unless the library has been told to transform it into another format.
1687 For example, 4 bit/pixel paletted or grayscale data will be returned
1688 2 pixels/byte with the leftmost pixel in the high-order bits of the byte,
1689 unless png_set_packing() is called. 8-bit RGB data will be stored
1690 in RGB RGB RGB format unless png_set_filler() or png_set_add_alpha()
1691 is called to insert filler bytes, either before or after each RGB triplet.
1693 16-bit RGB data will be returned RRGGBB RRGGBB, with the most significant
1694 byte of the color value first, unless png_set_scale_16() is called to
1695 transform it to regular RGB RGB triplets, or png_set_filler() or
1696 png_set_add alpha() is called to insert two filler bytes, either before
1697 or after each RRGGBB triplet. Similarly, 8-bit or 16-bit grayscale data can
1698 be modified with png_set_filler(), png_set_add_alpha(), png_set_strip_16(),
1699 or png_set_scale_16().
1701 The following code transforms grayscale images of less than 8 to 8 bits,
1702 changes paletted images to RGB, and adds a full alpha channel if there is
1703 transparency information in a tRNS chunk. This is most useful on
1704 grayscale images with bit depths of 2 or 4 or if there is a multiple-image
1705 viewing application that wishes to treat all images in the same way.
1707 if (color_type == PNG_COLOR_TYPE_PALETTE)
1708 png_set_palette_to_rgb(png_ptr);
1710 if (png_get_valid(png_ptr, info_ptr,
1711 PNG_INFO_tRNS)) png_set_tRNS_to_alpha(png_ptr);
1713 if (color_type == PNG_COLOR_TYPE_GRAY &&
1714 bit_depth < 8) png_set_expand_gray_1_2_4_to_8(png_ptr);
1716 The first two functions are actually aliases for png_set_expand(), added
1717 in libpng version 1.0.4, with the function names expanded to improve code
1718 readability. In some future version they may actually do different
1721 As of libpng version 1.2.9, png_set_expand_gray_1_2_4_to_8() was
1722 added. It expands the sample depth without changing tRNS to alpha.
1724 As of libpng version 1.5.2, png_set_expand_16() was added. It behaves as
1725 png_set_expand(); however, the resultant channels have 16 bits rather than 8.
1726 Use this when the output color or gray channels are made linear to avoid fairly
1727 severe accuracy loss.
1730 png_set_expand_16(png_ptr);
1732 PNG can have files with 16 bits per channel. If you only can handle
1733 8 bits per channel, this will strip the pixels down to 8-bit.
1735 if (bit_depth == 16)
1736 #if PNG_LIBPNG_VER >= 10504
1737 png_set_scale_16(png_ptr);
1739 png_set_strip_16(png_ptr);
1742 (The more accurate "png_set_scale_16()" API became available in libpng version
1745 If you need to process the alpha channel on the image separately from the image
1746 data (for example if you convert it to a bitmap mask) it is possible to have
1747 libpng strip the channel leaving just RGB or gray data:
1749 if (color_type & PNG_COLOR_MASK_ALPHA)
1750 png_set_strip_alpha(png_ptr);
1752 If you strip the alpha channel you need to find some other way of dealing with
1753 the information. If, instead, you want to convert the image to an opaque
1754 version with no alpha channel use png_set_background; see below.
1756 As of libpng version 1.5.2, almost all useful expansions are supported, the
1757 major ommissions are conversion of grayscale to indexed images (which can be
1758 done trivially in the application) and conversion of indexed to grayscale (which
1759 can be done by a trivial manipulation of the palette.)
1761 In the following table, the 01 means grayscale with depth<8, 31 means
1762 indexed with depth<8, other numerals represent the color type, "T" means
1763 the tRNS chunk is present, A means an alpha channel is present, and O
1764 means tRNS or alpha is present but all pixels in the image are opaque.
1766 FROM 01 31 0 0T 0O 2 2T 2O 3 3T 3O 4A 4O 6A 6O
1768 01 - [G] - - - - - - - - - - - - -
1769 31 [Q] Q [Q] [Q] [Q] Q Q Q Q Q Q [Q] [Q] Q Q
1770 0 1 G + . . G G G G G G B B GB GB
1771 0T lt Gt t + . Gt G G Gt G G Bt Bt GBt GBt
1772 0O lt Gt t . + Gt Gt G Gt Gt G Bt Bt GBt GBt
1773 2 C P C C C + . . C - - CB CB B B
1774 2T Ct - Ct C C t + t - - - CBt CBt Bt Bt
1775 2O Ct - Ct C C t t + - - - CBt CBt Bt Bt
1776 3 [Q] p [Q] [Q] [Q] Q Q Q + . . [Q] [Q] Q Q
1777 3T [Qt] p [Qt][Q] [Q] Qt Qt Qt t + t [Qt][Qt] Qt Qt
1778 3O [Qt] p [Qt][Q] [Q] Qt Qt Qt t t + [Qt][Qt] Qt Qt
1779 4A lA G A T T GA GT GT GA GT GT + BA G GBA
1780 4O lA GBA A T T GA GT GT GA GT GT BA + GBA G
1781 6A CA PA CA C C A T tT PA P P C CBA + BA
1782 6O CA PBA CA C C A tT T PA P P CBA C BA +
1785 "+" identifies entries where 'from' and 'to' are the same.
1786 "-" means the transformation is not supported.
1787 "." means nothing is necessary (a tRNS chunk can just be ignored).
1788 "t" means the transformation is obtained by png_set_tRNS.
1789 "A" means the transformation is obtained by png_set_add_alpha().
1790 "X" means the transformation is obtained by png_set_expand().
1791 "1" means the transformation is obtained by
1792 png_set_expand_gray_1_2_4_to_8() (and by png_set_expand()
1793 if there is no transparency in the original or the final
1795 "C" means the transformation is obtained by png_set_gray_to_rgb().
1796 "G" means the transformation is obtained by png_set_rgb_to_gray().
1797 "P" means the transformation is obtained by
1798 png_set_expand_palette_to_rgb().
1799 "p" means the transformation is obtained by png_set_packing().
1800 "Q" means the transformation is obtained by png_set_quantize().
1801 "T" means the transformation is obtained by
1802 png_set_tRNS_to_alpha().
1803 "B" means the transformation is obtained by
1804 png_set_background(), or png_strip_alpha().
1806 When an entry has multiple transforms listed all are required to cause the
1807 right overall transformation. When two transforms are separated by a comma
1808 either will do the job. When transforms are enclosed in [] the transform should
1809 do the job but this is currently unimplemented - a different format will result
1810 if the suggested transformations are used.
1812 In PNG files, the alpha channel in an image
1813 is the level of opacity. If you need the alpha channel in an image to
1814 be the level of transparency instead of opacity, you can invert the
1815 alpha channel (or the tRNS chunk data) after it's read, so that 0 is
1816 fully opaque and 255 (in 8-bit or paletted images) or 65535 (in 16-bit
1817 images) is fully transparent, with
1819 png_set_invert_alpha(png_ptr);
1821 PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as
1822 they can, resulting in, for example, 8 pixels per byte for 1 bit
1823 files. This code expands to 1 pixel per byte without changing the
1824 values of the pixels:
1827 png_set_packing(png_ptr);
1829 PNG files have possible bit depths of 1, 2, 4, 8, and 16. All pixels
1830 stored in a PNG image have been "scaled" or "shifted" up to the next
1831 higher possible bit depth (e.g. from 5 bits/sample in the range [0,31]
1832 to 8 bits/sample in the range [0, 255]). However, it is also possible
1833 to convert the PNG pixel data back to the original bit depth of the
1834 image. This call reduces the pixels back down to the original bit depth:
1836 png_color_8p sig_bit;
1838 if (png_get_sBIT(png_ptr, info_ptr, &sig_bit))
1839 png_set_shift(png_ptr, sig_bit);
1841 PNG files store 3-color pixels in red, green, blue order. This code
1842 changes the storage of the pixels to blue, green, red:
1844 if (color_type == PNG_COLOR_TYPE_RGB ||
1845 color_type == PNG_COLOR_TYPE_RGB_ALPHA)
1846 png_set_bgr(png_ptr);
1848 PNG files store RGB pixels packed into 3 or 6 bytes. This code expands them
1849 into 4 or 8 bytes for windowing systems that need them in this format:
1851 if (color_type == PNG_COLOR_TYPE_RGB)
1852 png_set_filler(png_ptr, filler, PNG_FILLER_BEFORE);
1854 where "filler" is the 8-bit or 16-bit number to fill with, and the location
1855 is either PNG_FILLER_BEFORE or PNG_FILLER_AFTER, depending upon whether
1856 you want the filler before the RGB or after. When filling an 8-bit pixel,
1857 the least significant 8 bits of the number are used, if a 16-bit number is
1858 supplied. This transformation does not affect images that already have full
1859 alpha channels. To add an opaque alpha channel, use filler=0xffff and
1860 PNG_FILLER_AFTER which will generate RGBA pixels.
1862 Note that png_set_filler() does not change the color type. If you want
1863 to do that, you can add a true alpha channel with
1865 if (color_type == PNG_COLOR_TYPE_RGB ||
1866 color_type == PNG_COLOR_TYPE_GRAY)
1867 png_set_add_alpha(png_ptr, filler, PNG_FILLER_AFTER);
1869 where "filler" contains the alpha value to assign to each pixel.
1870 The png_set_add_alpha() function was added in libpng-1.2.7.
1872 If you are reading an image with an alpha channel, and you need the
1873 data as ARGB instead of the normal PNG format RGBA:
1875 if (color_type == PNG_COLOR_TYPE_RGB_ALPHA)
1876 png_set_swap_alpha(png_ptr);
1878 For some uses, you may want a grayscale image to be represented as
1879 RGB. This code will do that conversion:
1881 if (color_type == PNG_COLOR_TYPE_GRAY ||
1882 color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
1883 png_set_gray_to_rgb(png_ptr);
1885 Conversely, you can convert an RGB or RGBA image to grayscale or grayscale
1888 if (color_type == PNG_COLOR_TYPE_RGB ||
1889 color_type == PNG_COLOR_TYPE_RGB_ALPHA)
1890 png_set_rgb_to_gray(png_ptr, error_action,
1891 double red_weight, double green_weight);
1893 error_action = 1: silently do the conversion
1895 error_action = 2: issue a warning if the original
1896 image has any pixel where
1897 red != green or red != blue
1899 error_action = 3: issue an error and abort the
1900 conversion if the original
1901 image has any pixel where
1902 red != green or red != blue
1904 red_weight: weight of red component
1906 green_weight: weight of green component
1907 If either weight is negative, default
1910 In the corresponding fixed point API the red_weight and green_weight values are
1911 simply scaled by 100,000:
1913 png_set_rgb_to_gray(png_ptr, error_action,
1914 png_fixed_point red_weight,
1915 png_fixed_point green_weight);
1917 If you have set error_action = 1 or 2, you can
1918 later check whether the image really was gray, after processing
1919 the image rows, with the png_get_rgb_to_gray_status(png_ptr) function.
1920 It will return a png_byte that is zero if the image was gray or
1921 1 if there were any non-gray pixels. Background and sBIT data
1922 will be silently converted to grayscale, using the green channel
1923 data for sBIT, regardless of the error_action setting.
1925 The default values come from the PNG file cHRM chunk if present; otherwise, the
1926 defaults correspond to the ITU-R recommendation 709, and also the sRGB color
1927 space, as recommended in the Charles Poynton's Colour FAQ,
1928 Copyright (c) 2006-11-28 Charles Poynton, in section 9:
1930 <http://www.poynton.com/notes/colour_and_gamma/ColorFAQ.html#RTFToC9>
1932 Y = 0.2126 * R + 0.7152 * G + 0.0722 * B
1934 Previous versions of this document, 1998 through 2002, recommended a slightly
1937 Y = 0.212671 * R + 0.715160 * G + 0.072169 * B
1939 Libpng uses an integer approximation:
1941 Y = (6968 * R + 23434 * G + 2366 * B)/32768
1943 The calculation is done in a linear colorspace, if the image gamma
1946 The png_set_background() function has been described already; it tells libpng to
1947 composite images with alpha or simple transparency against the supplied
1948 background color. For compatibility with versions of libpng earlier than
1949 libpng-1.5.4 it is recommended that you call the function after reading the file
1950 header, even if you don't want to use the color in a bKGD chunk, if one exists.
1952 If the PNG file contains a bKGD chunk (PNG_INFO_bKGD valid),
1953 you may use this color, or supply another color more suitable for
1954 the current display (e.g., the background color from a web page). You
1955 need to tell libpng how the color is represented, both the format of the
1956 component values in the color (the number of bits) and the gamma encoding of the
1957 color. The function takes two arguments, background_gamma_mode and need_expand
1958 to convey this information; however, only two combinations are likely to be
1961 png_color_16 my_background;
1962 png_color_16p image_background;
1964 if (png_get_bKGD(png_ptr, info_ptr, &image_background))
1965 png_set_background(png_ptr, image_background,
1966 PNG_BACKGROUND_GAMMA_FILE, 1/*needs to be expanded*/, 1);
1968 png_set_background(png_ptr, &my_background,
1969 PNG_BACKGROUND_GAMMA_SCREEN, 0/*do not expand*/, 1);
1971 The second call was described above - my_background is in the format of the
1972 final, display, output produced by libpng. Because you now know the format of
1973 the PNG it is possible to avoid the need to choose either 8-bit or 16-bit
1974 output and to retain palette images (the palette colors will be modified
1975 appropriately and the tRNS chunk removed.) However, if you are doing this,
1976 take great care not to ask for transformations without checking first that
1979 In the first call the background color has the original bit depth and color type
1980 of the PNG file. So, for palette images the color is supplied as a palette
1981 index and for low bit greyscale images the color is a reduced bit value in
1982 image_background->gray.
1984 If you didn't call png_set_gamma() before reading the file header, for example
1985 if you need your code to remain compatible with older versions of libpng prior
1986 to libpng-1.5.4, this is the place to call it.
1988 Do not call it if you called png_set_alpha_mode(); doing so will damage the
1989 settings put in place by png_set_alpha_mode(). (If png_set_alpha_mode() is
1990 supported then you can certainly do png_set_gamma() before reading the PNG
1993 This API unconditionally sets the screen and file gamma values, so it will
1994 override the value in the PNG file unless it is called before the PNG file
1995 reading starts. For this reason you must always call it with the PNG file
1996 value when you call it in this position:
1998 if (png_get_gAMA(png_ptr, info_ptr, &file_gamma))
1999 png_set_gamma(png_ptr, screen_gamma, file_gamma);
2002 png_set_gamma(png_ptr, screen_gamma, 0.45455);
2004 If you need to reduce an RGB file to a paletted file, or if a paletted
2005 file has more entries than will fit on your screen, png_set_quantize()
2006 will do that. Note that this is a simple match quantization that merely
2007 finds the closest color available. This should work fairly well with
2008 optimized palettes, but fairly badly with linear color cubes. If you
2009 pass a palette that is larger than maximum_colors, the file will
2010 reduce the number of colors in the palette so it will fit into
2011 maximum_colors. If there is a histogram, libpng will use it to make
2012 more intelligent choices when reducing the palette. If there is no
2013 histogram, it may not do as good a job.
2015 if (color_type & PNG_COLOR_MASK_COLOR)
2017 if (png_get_valid(png_ptr, info_ptr,
2020 png_uint_16p histogram = NULL;
2022 png_get_hIST(png_ptr, info_ptr,
2024 png_set_quantize(png_ptr, palette, num_palette,
2025 max_screen_colors, histogram, 1);
2030 png_color std_color_cube[MAX_SCREEN_COLORS] =
2033 png_set_quantize(png_ptr, std_color_cube,
2034 MAX_SCREEN_COLORS, MAX_SCREEN_COLORS,
2039 PNG files describe monochrome as black being zero and white being one.
2040 The following code will reverse this (make black be one and white be
2043 if (bit_depth == 1 && color_type == PNG_COLOR_TYPE_GRAY)
2044 png_set_invert_mono(png_ptr);
2046 This function can also be used to invert grayscale and gray-alpha images:
2048 if (color_type == PNG_COLOR_TYPE_GRAY ||
2049 color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
2050 png_set_invert_mono(png_ptr);
2052 PNG files store 16-bit pixels in network byte order (big-endian,
2053 ie. most significant bits first). This code changes the storage to the
2054 other way (little-endian, i.e. least significant bits first, the
2055 way PCs store them):
2057 if (bit_depth == 16)
2058 png_set_swap(png_ptr);
2060 If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you
2061 need to change the order the pixels are packed into bytes, you can use:
2064 png_set_packswap(png_ptr);
2066 Finally, you can write your own transformation function if none of
2067 the existing ones meets your needs. This is done by setting a callback
2070 png_set_read_user_transform_fn(png_ptr,
2073 You must supply the function
2075 void read_transform_fn(png_structp png_ptr, png_row_infop
2076 row_info, png_bytep data)
2078 See pngtest.c for a working example. Your function will be called
2079 after all of the other transformations have been processed. Take care with
2080 interlaced images if you do the interlace yourself - the width of the row is the
2081 width in 'row_info', not the overall image width.
2083 If supported, libpng provides two information routines that you can use to find
2084 where you are in processing the image:
2086 png_get_current_pass_number(png_structp png_ptr);
2087 png_get_current_row_number(png_structp png_ptr);
2089 Don't try using these outside a transform callback - firstly they are only
2090 supported if user transforms are supported, secondly they may well return
2091 unexpected results unless the row is actually being processed at the moment they
2095 images the value returned is the row in the input sub-image image. Use
2096 PNG_ROW_FROM_PASS_ROW(row, pass) and PNG_COL_FROM_PASS_COL(col, pass) to
2097 find the output pixel (x,y) given an interlaced sub-image pixel (row,col,pass).
2099 The discussion of interlace handling above contains more information on how to
2102 You can also set up a pointer to a user structure for use by your
2103 callback function, and you can inform libpng that your transform
2104 function will change the number of channels or bit depth with the
2107 png_set_user_transform_info(png_ptr, user_ptr,
2108 user_depth, user_channels);
2110 The user's application, not libpng, is responsible for allocating and
2111 freeing any memory required for the user structure.
2113 You can retrieve the pointer via the function
2114 png_get_user_transform_ptr(). For example:
2116 voidp read_user_transform_ptr =
2117 png_get_user_transform_ptr(png_ptr);
2119 The last thing to handle is interlacing; this is covered in detail below,
2120 but you must call the function here if you want libpng to handle expansion
2121 of the interlaced image.
2123 number_of_passes = png_set_interlace_handling(png_ptr);
2125 After setting the transformations, libpng can update your png_info
2126 structure to reflect any transformations you've requested with this
2129 png_read_update_info(png_ptr, info_ptr);
2131 This is most useful to update the info structure's rowbytes
2132 field so you can use it to allocate your image memory. This function
2133 will also update your palette with the correct screen_gamma and
2134 background if these have been given with the calls above. You may
2135 only call png_read_update_info() once with a particular info_ptr.
2137 After you call png_read_update_info(), you can allocate any
2138 memory you need to hold the image. The row data is simply
2139 raw byte data for all forms of images. As the actual allocation
2140 varies among applications, no example will be given. If you
2141 are allocating one large chunk, you will need to build an
2142 array of pointers to each row, as it will be needed for some
2143 of the functions below.
2145 Remember: Before you call png_read_update_info(), the png_get_*()
2146 functions return the values corresponding to the original PNG image.
2147 After you call png_read_update_info the values refer to the image
2148 that libpng will output. Consequently you must call all the png_set_
2149 functions before you call png_read_update_info(). This is particularly
2150 important for png_set_interlace_handling() - if you are going to call
2151 png_read_update_info() you must call png_set_interlace_handling() before
2152 it unless you want to receive interlaced output.
2156 After you've allocated memory, you can read the image data.
2157 The simplest way to do this is in one function call. If you are
2158 allocating enough memory to hold the whole image, you can just
2159 call png_read_image() and libpng will read in all the image data
2160 and put it in the memory area supplied. You will need to pass in
2161 an array of pointers to each row.
2163 This function automatically handles interlacing, so you don't
2164 need to call png_set_interlace_handling() (unless you call
2165 png_read_update_info()) or call this function multiple times, or any
2166 of that other stuff necessary with png_read_rows().
2168 png_read_image(png_ptr, row_pointers);
2170 where row_pointers is:
2172 png_bytep row_pointers[height];
2174 You can point to void or char or whatever you use for pixels.
2176 If you don't want to read in the whole image at once, you can
2177 use png_read_rows() instead. If there is no interlacing (check
2178 interlace_type == PNG_INTERLACE_NONE), this is simple:
2180 png_read_rows(png_ptr, row_pointers, NULL,
2183 where row_pointers is the same as in the png_read_image() call.
2185 If you are doing this just one row at a time, you can do this with
2186 a single row_pointer instead of an array of row_pointers:
2188 png_bytep row_pointer = row;
2189 png_read_row(png_ptr, row_pointer, NULL);
2191 If the file is interlaced (interlace_type != 0 in the IHDR chunk), things
2192 get somewhat harder. The only current (PNG Specification version 1.2)
2193 interlacing type for PNG is (interlace_type == PNG_INTERLACE_ADAM7);
2194 a somewhat complicated 2D interlace scheme, known as Adam7, that
2195 breaks down an image into seven smaller images of varying size, based
2196 on an 8x8 grid. This number is defined (from libpng 1.5) as
2197 PNG_INTERLACE_ADAM7_PASSES in png.h
2199 libpng can fill out those images or it can give them to you "as is".
2200 It is almost always better to have libpng handle the interlacing for you.
2201 If you want the images filled out, there are two ways to do that. The one
2202 mentioned in the PNG specification is to expand each pixel to cover
2203 those pixels that have not been read yet (the "rectangle" method).
2204 This results in a blocky image for the first pass, which gradually
2205 smooths out as more pixels are read. The other method is the "sparkle"
2206 method, where pixels are drawn only in their final locations, with the
2207 rest of the image remaining whatever colors they were initialized to
2208 before the start of the read. The first method usually looks better,
2209 but tends to be slower, as there are more pixels to put in the rows.
2211 If, as is likely, you want libpng to expand the images, call this before
2212 calling png_start_read_image() or png_read_update_info():
2214 if (interlace_type == PNG_INTERLACE_ADAM7)
2216 = png_set_interlace_handling(png_ptr);
2218 This will return the number of passes needed. Currently, this is seven,
2219 but may change if another interlace type is added. This function can be
2220 called even if the file is not interlaced, where it will return one pass.
2221 You then need to read the whole image 'number_of_passes' times. Each time
2222 will distribute the pixels from the current pass to the correct place in
2223 the output image, so you need to supply the same rows to png_read_rows in
2226 If you are not going to display the image after each pass, but are
2227 going to wait until the entire image is read in, use the sparkle
2228 effect. This effect is faster and the end result of either method
2229 is exactly the same. If you are planning on displaying the image
2230 after each pass, the "rectangle" effect is generally considered the
2233 If you only want the "sparkle" effect, just call png_read_row() or
2235 normal, with the third parameter NULL. Make sure you make pass over
2236 the image number_of_passes times, and you don't change the data in the
2237 rows between calls. You can change the locations of the data, just
2238 not the data. Each pass only writes the pixels appropriate for that
2239 pass, and assumes the data from previous passes is still valid.
2241 png_read_rows(png_ptr, row_pointers, NULL,
2244 png_read_row(png_ptr, row_pointers, NULL);
2246 If you only want the first effect (the rectangles), do the same as
2247 before except pass the row buffer in the third parameter, and leave
2248 the second parameter NULL.
2250 png_read_rows(png_ptr, NULL, row_pointers,
2253 png_read_row(png_ptr, NULL, row_pointers);
2255 If you don't want libpng to handle the interlacing details, just call
2256 png_read_rows() PNG_INTERLACE_ADAM7_PASSES times to read in all the images.
2257 Each of the images is a valid image by itself; however, you will almost
2258 certainly need to distribute the pixels from each sub-image to the
2259 correct place. This is where everything gets very tricky.
2261 If you want to retrieve the separate images you must pass the correct
2262 number of rows to each successive call of png_read_rows(). The calculation
2263 gets pretty complicated for small images, where some sub-images may
2264 not even exist because either their width or height ends up zero.
2265 libpng provides two macros to help you in 1.5 and later versions:
2267 png_uint_32 width = PNG_PASS_COLS(image_width, pass_number);
2268 png_uint_32 height = PNG_PASS_ROWS(image_height, pass_number);
2270 Respectively these tell you the width and height of the sub-image
2271 corresponding to the numbered pass. 'pass' is in in the range 0 to 6 -
2272 this can be confusing because the specification refers to the same passes
2273 as 1 to 7! Be careful, you must check both the width and height before
2274 calling png_read_rows() and not call it for that pass if either is zero.
2276 You can, of course, read each sub-image row by row. If you want to
2277 produce optimal code to make a pixel-by-pixel transformation of an
2278 interlaced image this is the best approach; read each row of each pass,
2279 transform it, and write it out to a new interlaced image.
2281 If you want to de-interlace the image yourself libpng provides further
2282 macros to help that tell you where to place the pixels in the output image.
2283 Because the interlacing scheme is rectangular - sub-image pixels are always
2284 arranged on a rectangular grid - all you need to know for each pass is the
2285 starting column and row in the output image of the first pixel plus the
2286 spacing between each pixel. As of libpng 1.5 there are four macros to
2287 retrieve this information:
2289 png_uint_32 x = PNG_PASS_START_COL(pass);
2290 png_uint_32 y = PNG_PASS_START_ROW(pass);
2291 png_uint_32 xStep = 1U << PNG_PASS_COL_SHIFT(pass);
2292 png_uint_32 yStep = 1U << PNG_PASS_ROW_SHIFT(pass);
2294 These allow you to write the obvious loop:
2296 png_uint_32 input_y = 0;
2297 png_uint_32 output_y = PNG_PASS_START_ROW(pass);
2299 while (output_y < output_image_height)
2301 png_uint_32 input_x = 0;
2302 png_uint_32 output_x = PNG_PASS_START_COL(pass);
2304 while (output_x < output_image_width)
2306 image[output_y][output_x] =
2307 subimage[pass][input_y][input_x++];
2316 Notice that the steps between successive output rows and columns are
2317 returned as shifts. This is possible because the pixels in the subimages
2318 are always a power of 2 apart - 1, 2, 4 or 8 pixels - in the original
2319 image. In practice you may need to directly calculate the output coordinate
2320 given an input coordinate. libpng provides two further macros for this
2323 png_uint_32 output_x = PNG_COL_FROM_PASS_COL(input_x, pass);
2324 png_uint_32 output_y = PNG_ROW_FROM_PASS_ROW(input_y, pass);
2326 Finally a pair of macros are provided to tell you if a particular image
2327 row or column appears in a given pass:
2329 int col_in_pass = PNG_COL_IN_INTERLACE_PASS(output_x, pass);
2330 int row_in_pass = PNG_ROW_IN_INTERLACE_PASS(output_y, pass);
2332 Bear in mind that you will probably also need to check the width and height
2333 of the pass in addition to the above to be sure the pass even exists!
2335 With any luck you are convinced by now that you don't want to do your own
2336 interlace handling. In reality normally the only good reason for doing this
2337 is if you are processing PNG files on a pixel-by-pixel basis and don't want
2338 to load the whole file into memory when it is interlaced.
2340 libpng includes a test program, pngvalid, that illustrates reading and
2341 writing of interlaced images. If you can't get interlacing to work in your
2342 code and don't want to leave it to libpng (the recommended approach), see
2343 how pngvalid.c does it.
2345 Finishing a sequential read
2347 After you are finished reading the image through the
2348 low-level interface, you can finish reading the file.
2350 If you want to use a different crc action for handling CRC errors in
2351 chunks after the image data, you can call png_set_crc_action()
2352 again at this point.
2354 If you are interested in comments or time, which may be stored either
2355 before or after the image data, you should pass the separate png_info
2356 struct if you want to keep the comments from before and after the image
2359 png_infop end_info = png_create_info_struct(png_ptr);
2363 png_destroy_read_struct(&png_ptr, &info_ptr,
2368 png_read_end(png_ptr, end_info);
2370 If you are not interested, you should still call png_read_end()
2371 but you can pass NULL, avoiding the need to create an end_info structure.
2372 If you do this, libpng will not process any chunks after IDAT other than
2373 skipping over them and perhaps (depending on whether you have called
2374 png_set_crc_action) checking their CRCs while looking for the IEND chunk.
2376 png_read_end(png_ptr, (png_infop)NULL);
2378 If you don't call png_read_end(), then your file pointer will be
2379 left pointing to the first chunk after the last IDAT, which is probably
2380 not what you want if you expect to read something beyond the end of
2383 When you are done, you can free all memory allocated by libpng like this:
2385 png_destroy_read_struct(&png_ptr, &info_ptr,
2388 or, if you didn't create an end_info structure,
2390 png_destroy_read_struct(&png_ptr, &info_ptr,
2393 It is also possible to individually free the info_ptr members that
2394 point to libpng-allocated storage with the following function:
2396 png_free_data(png_ptr, info_ptr, mask, seq)
2398 mask - identifies data to be freed, a mask
2399 containing the bitwise OR of one or
2401 PNG_FREE_PLTE, PNG_FREE_TRNS,
2402 PNG_FREE_HIST, PNG_FREE_ICCP,
2403 PNG_FREE_PCAL, PNG_FREE_ROWS,
2404 PNG_FREE_SCAL, PNG_FREE_SPLT,
2405 PNG_FREE_TEXT, PNG_FREE_UNKN,
2406 or simply PNG_FREE_ALL
2408 seq - sequence number of item to be freed
2411 This function may be safely called when the relevant storage has
2412 already been freed, or has not yet been allocated, or was allocated
2413 by the user and not by libpng, and will in those cases do nothing.
2414 The "seq" parameter is ignored if only one item of the selected data
2415 type, such as PLTE, is allowed. If "seq" is not -1, and multiple items
2416 are allowed for the data type identified in the mask, such as text or
2417 sPLT, only the n'th item in the structure is freed, where n is "seq".
2419 The default behavior is only to free data that was allocated internally
2420 by libpng. This can be changed, so that libpng will not free the data,
2421 or so that it will free data that was allocated by the user with png_malloc()
2422 or png_calloc() and passed in via a png_set_*() function, with
2424 png_data_freer(png_ptr, info_ptr, freer, mask)
2427 PNG_DESTROY_WILL_FREE_DATA
2428 PNG_SET_WILL_FREE_DATA
2429 PNG_USER_WILL_FREE_DATA
2431 mask - which data elements are affected
2432 same choices as in png_free_data()
2434 This function only affects data that has already been allocated.
2435 You can call this function after reading the PNG data but before calling
2436 any png_set_*() functions, to control whether the user or the png_set_*()
2437 function is responsible for freeing any existing data that might be present,
2438 and again after the png_set_*() functions to control whether the user
2439 or png_destroy_*() is supposed to free the data. When the user assumes
2440 responsibility for libpng-allocated data, the application must use
2441 png_free() to free it, and when the user transfers responsibility to libpng
2442 for data that the user has allocated, the user must have used png_malloc()
2443 or png_calloc() to allocate it.
2445 If you allocated your row_pointers in a single block, as suggested above in
2446 the description of the high level read interface, you must not transfer
2447 responsibility for freeing it to the png_set_rows or png_read_destroy function,
2448 because they would also try to free the individual row_pointers[i].
2450 If you allocated text_ptr.text, text_ptr.lang, and text_ptr.translated_keyword
2451 separately, do not transfer responsibility for freeing text_ptr to libpng,
2452 because when libpng fills a png_text structure it combines these members with
2453 the key member, and png_free_data() will free only text_ptr.key. Similarly,
2454 if you transfer responsibility for free'ing text_ptr from libpng to your
2455 application, your application must not separately free those members.
2457 The png_free_data() function will turn off the "valid" flag for anything
2458 it frees. If you need to turn the flag off for a chunk that was freed by
2459 your application instead of by libpng, you can use
2461 png_set_invalid(png_ptr, info_ptr, mask);
2463 mask - identifies the chunks to be made invalid,
2464 containing the bitwise OR of one or
2466 PNG_INFO_gAMA, PNG_INFO_sBIT,
2467 PNG_INFO_cHRM, PNG_INFO_PLTE,
2468 PNG_INFO_tRNS, PNG_INFO_bKGD,
2469 PNG_INFO_hIST, PNG_INFO_pHYs,
2470 PNG_INFO_oFFs, PNG_INFO_tIME,
2471 PNG_INFO_pCAL, PNG_INFO_sRGB,
2472 PNG_INFO_iCCP, PNG_INFO_sPLT,
2473 PNG_INFO_sCAL, PNG_INFO_IDAT
2475 For a more compact example of reading a PNG image, see the file example.c.
2477 Reading PNG files progressively
2479 The progressive reader is slightly different from the non-progressive
2480 reader. Instead of calling png_read_info(), png_read_rows(), and
2481 png_read_end(), you make one call to png_process_data(), which calls
2482 callbacks when it has the info, a row, or the end of the image. You
2483 set up these callbacks with png_set_progressive_read_fn(). You don't
2484 have to worry about the input/output functions of libpng, as you are
2485 giving the library the data directly in png_process_data(). I will
2486 assume that you have read the section on reading PNG files above,
2487 so I will only highlight the differences (although I will show
2490 png_structp png_ptr;
2493 /* An example code fragment of how you would
2494 initialize the progressive reader in your
2497 initialize_png_reader()
2499 png_ptr = png_create_read_struct
2500 (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
2501 user_error_fn, user_warning_fn);
2506 info_ptr = png_create_info_struct(png_ptr);
2510 png_destroy_read_struct(&png_ptr,
2511 (png_infopp)NULL, (png_infopp)NULL);
2515 if (setjmp(png_jmpbuf(png_ptr)))
2517 png_destroy_read_struct(&png_ptr, &info_ptr,
2522 /* This one's new. You can provide functions
2523 to be called when the header info is valid,
2524 when each row is completed, and when the image
2525 is finished. If you aren't using all functions,
2526 you can specify NULL parameters. Even when all
2527 three functions are NULL, you need to call
2528 png_set_progressive_read_fn(). You can use
2529 any struct as the user_ptr (cast to a void pointer
2530 for the function call), and retrieve the pointer
2531 from inside the callbacks using the function
2533 png_get_progressive_ptr(png_ptr);
2535 which will return a void pointer, which you have
2536 to cast appropriately.
2538 png_set_progressive_read_fn(png_ptr, (void *)user_ptr,
2539 info_callback, row_callback, end_callback);
2544 /* A code fragment that you call as you receive blocks
2547 process_data(png_bytep buffer, png_uint_32 length)
2549 if (setjmp(png_jmpbuf(png_ptr)))
2551 png_destroy_read_struct(&png_ptr, &info_ptr,
2556 /* This one's new also. Simply give it a chunk
2557 of data from the file stream (in order, of
2558 course). On machines with segmented memory
2559 models machines, don't give it any more than
2560 64K. The library seems to run fine with sizes
2561 of 4K. Although you can give it much less if
2562 necessary (I assume you can give it chunks of
2563 1 byte, I haven't tried less than 256 bytes
2564 yet). When this function returns, you may
2565 want to display any rows that were generated
2566 in the row callback if you don't already do
2569 png_process_data(png_ptr, info_ptr, buffer, length);
2571 /* At this point you can call png_process_data_skip if
2572 you want to handle data the library will skip yourself;
2573 it simply returns the number of bytes to skip (and stops
2574 libpng skipping that number of bytes on the next
2575 png_process_data call).
2579 /* This function is called (as set by
2580 png_set_progressive_read_fn() above) when enough data
2581 has been supplied so all of the header has been
2585 info_callback(png_structp png_ptr, png_infop info)
2587 /* Do any setup here, including setting any of
2588 the transformations mentioned in the Reading
2589 PNG files section. For now, you _must_ call
2590 either png_start_read_image() or
2591 png_read_update_info() after all the
2592 transformations are set (even if you don't set
2593 any). You may start getting rows before
2594 png_process_data() returns, so this is your
2595 last chance to prepare for that.
2597 This is where you turn on interlace handling,
2598 assuming you don't want to do it yourself.
2600 If you need to you can stop the processing of
2601 your original input data at this point by calling
2602 png_process_data_pause. This returns the number
2603 of unprocessed bytes from the last png_process_data
2604 call - it is up to you to ensure that the next call
2605 sees these bytes again. If you don't want to bother
2606 with this you can get libpng to cache the unread
2607 bytes by setting the 'save' parameter (see png.h) but
2608 then libpng will have to copy the data internally.
2612 /* This function is called when each row of image
2615 row_callback(png_structp png_ptr, png_bytep new_row,
2616 png_uint_32 row_num, int pass)
2618 /* If the image is interlaced, and you turned
2619 on the interlace handler, this function will
2620 be called for every row in every pass. Some
2621 of these rows will not be changed from the
2622 previous pass. When the row is not changed,
2623 the new_row variable will be NULL. The rows
2624 and passes are called in order, so you don't
2625 really need the row_num and pass, but I'm
2626 supplying them because it may make your life
2629 If you did not turn on interlace handling then
2630 the callback is called for each row of each
2631 sub-image when the image is interlaced. In this
2632 case 'row_num' is the row in the sub-image, not
2633 the row in the output image as it is in all other
2636 For the non-NULL rows of interlaced images when
2637 you have switched on libpng interlace handling,
2638 you must call png_progressive_combine_row()
2639 passing in the row and the old row. You can
2640 call this function for NULL rows (it will just
2641 return) and for non-interlaced images (it just
2642 does the memcpy for you) if it will make the
2643 code easier. Thus, you can just do this for
2644 all cases if you switch on interlace handling;
2647 png_progressive_combine_row(png_ptr, old_row,
2650 /* where old_row is what was displayed
2651 previously for the row. Note that the first
2652 pass (pass == 0, really) will completely cover
2653 the old row, so the rows do not have to be
2654 initialized. After the first pass (and only
2655 for interlaced images), you will have to pass
2656 the current row, and the function will combine
2657 the old row and the new row.
2659 You can also call png_process_data_pause in this
2660 callback - see above.
2665 end_callback(png_structp png_ptr, png_infop info)
2667 /* This function is called after the whole image
2668 has been read, including any chunks after the
2669 image (up to and including the IEND). You
2670 will usually have the same info chunk as you
2671 had in the header, although some data may have
2672 been added to the comments and time fields.
2674 Most people won't do much here, perhaps setting
2675 a flag that marks the image as finished.
2683 Much of this is very similar to reading. However, everything of
2684 importance is repeated here, so you won't have to constantly look
2685 back up in the reading section to understand writing.
2689 You will want to do the I/O initialization before you get into libpng,
2690 so if it doesn't work, you don't have anything to undo. If you are not
2691 using the standard I/O functions, you will need to replace them with
2692 custom writing functions. See the discussion under Customizing libpng.
2694 FILE *fp = fopen(file_name, "wb");
2699 Next, png_struct and png_info need to be allocated and initialized.
2700 As these can be both relatively large, you may not want to store these
2701 on the stack, unless you have stack space to spare. Of course, you
2702 will want to check if they return NULL. If you are also reading,
2703 you won't want to name your read structure and your write structure
2704 both "png_ptr"; you can call them anything you like, such as
2705 "read_ptr" and "write_ptr". Look at pngtest.c, for example.
2707 png_structp png_ptr = png_create_write_struct
2708 (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
2709 user_error_fn, user_warning_fn);
2714 png_infop info_ptr = png_create_info_struct(png_ptr);
2717 png_destroy_write_struct(&png_ptr,
2722 If you want to use your own memory allocation routines,
2723 define PNG_USER_MEM_SUPPORTED and use
2724 png_create_write_struct_2() instead of png_create_write_struct():
2726 png_structp png_ptr = png_create_write_struct_2
2727 (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
2728 user_error_fn, user_warning_fn, (png_voidp)
2729 user_mem_ptr, user_malloc_fn, user_free_fn);
2731 After you have these structures, you will need to set up the
2732 error handling. When libpng encounters an error, it expects to
2733 longjmp() back to your routine. Therefore, you will need to call
2734 setjmp() and pass the png_jmpbuf(png_ptr). If you
2735 write the file from different routines, you will need to update
2736 the png_jmpbuf(png_ptr) every time you enter a new routine that will
2737 call a png_*() function. See your documentation of setjmp/longjmp
2738 for your compiler for more information on setjmp/longjmp. See
2739 the discussion on libpng error handling in the Customizing Libpng
2740 section below for more information on the libpng error handling.
2742 if (setjmp(png_jmpbuf(png_ptr)))
2744 png_destroy_write_struct(&png_ptr, &info_ptr);
2751 If you would rather avoid the complexity of setjmp/longjmp issues,
2752 you can compile libpng with PNG_NO_SETJMP, in which case
2753 errors will result in a call to PNG_ABORT() which defaults to abort().
2755 You can #define PNG_ABORT() to a function that does something
2756 more useful than abort(), as long as your function does not
2759 Checking for invalid palette index on write was added at libpng
2760 1.5.10. If a pixel contains an invalid (out-of-range) index libpng issues
2761 a benign error. This is enabled by default because this condition is an
2762 error according to the PNG specification, Clause 11.3.2, but the error can
2763 be ignored in each png_ptr with
2765 png_set_check_for_invalid_index(png_ptr, 0);
2767 If the error is ignored, or if png_benign_error() treats it as a warning,
2768 any invalid pixels are written as-is by the encoder, resulting in an
2769 invalid PNG datastream as output. In this case the application is
2770 responsible for ensuring that the pixel indexes are in range when it writes
2771 a PLTE chunk with fewer entries than the bit depth would allow.
2773 Now you need to set up the output code. The default for libpng is to
2774 use the C function fwrite(). If you use this, you will need to pass a
2775 valid FILE * in the function png_init_io(). Be sure that the file is
2776 opened in binary mode. Again, if you wish to handle writing data in
2777 another way, see the discussion on libpng I/O handling in the Customizing
2778 Libpng section below.
2780 png_init_io(png_ptr, fp);
2782 If you are embedding your PNG into a datastream such as MNG, and don't
2783 want libpng to write the 8-byte signature, or if you have already
2784 written the signature in your application, use
2786 png_set_sig_bytes(png_ptr, 8);
2788 to inform libpng that it should not write a signature.
2792 At this point, you can set up a callback function that will be
2793 called after each row has been written, which you can use to control
2794 a progress meter or the like. It's demonstrated in pngtest.c.
2795 You must supply a function
2797 void write_row_callback(png_structp png_ptr, png_uint_32 row,
2800 /* put your code here */
2803 (You can give it another name that you like instead of "write_row_callback")
2805 To inform libpng about your function, use
2807 png_set_write_status_fn(png_ptr, write_row_callback);
2809 When this function is called the row has already been completely processed and
2810 it has also been written out. The 'row' and 'pass' refer to the next row to be
2812 non-interlaced case the row that was just handled is simply one less than the
2813 passed in row number, and pass will always be 0. For the interlaced case the
2814 same applies unless the row value is 0, in which case the row just handled was
2815 the last one from one of the preceding passes. Because interlacing may skip a
2816 pass you cannot be sure that the preceding pass is just 'pass-1', if you really
2817 need to know what the last pass is record (row,pass) from the callback and use
2818 the last recorded value each time.
2820 As with the user transform you can find the output row using the
2821 PNG_ROW_FROM_PASS_ROW macro.
2823 You now have the option of modifying how the compression library will
2824 run. The following functions are mainly for testing, but may be useful
2825 in some cases, like if you need to write PNG files extremely fast and
2826 are willing to give up some compression, or if you want to get the
2827 maximum possible compression at the expense of slower writing. If you
2828 have no special needs in this area, let the library do what it wants by
2829 not calling this function at all, as it has been tuned to deliver a good
2830 speed/compression ratio. The second parameter to png_set_filter() is
2831 the filter method, for which the only valid values are 0 (as of the
2832 July 1999 PNG specification, version 1.2) or 64 (if you are writing
2833 a PNG datastream that is to be embedded in a MNG datastream). The third
2834 parameter is a flag that indicates which filter type(s) are to be tested
2835 for each scanline. See the PNG specification for details on the specific
2839 /* turn on or off filtering, and/or choose
2840 specific filters. You can use either a single
2841 PNG_FILTER_VALUE_NAME or the bitwise OR of one
2842 or more PNG_FILTER_NAME masks.
2844 png_set_filter(png_ptr, 0,
2845 PNG_FILTER_NONE | PNG_FILTER_VALUE_NONE |
2846 PNG_FILTER_SUB | PNG_FILTER_VALUE_SUB |
2847 PNG_FILTER_UP | PNG_FILTER_VALUE_UP |
2848 PNG_FILTER_AVG | PNG_FILTER_VALUE_AVG |
2849 PNG_FILTER_PAETH | PNG_FILTER_VALUE_PAETH|
2850 PNG_ALL_FILTERS | PNG_FAST_FILTERS);
2852 If an application wants to start and stop using particular filters during
2853 compression, it should start out with all of the filters (to ensure that
2854 the previous row of pixels will be stored in case it's needed later),
2855 and then add and remove them after the start of compression.
2857 If you are writing a PNG datastream that is to be embedded in a MNG
2858 datastream, the second parameter can be either 0 or 64.
2860 The png_set_compression_*() functions interface to the zlib compression
2861 library, and should mostly be ignored unless you really know what you are
2862 doing. The only generally useful call is png_set_compression_level()
2863 which changes how much time zlib spends on trying to compress the image
2864 data. See the Compression Library (zlib.h and algorithm.txt, distributed
2865 with zlib) for details on the compression levels.
2869 /* Set the zlib compression level */
2870 png_set_compression_level(png_ptr,
2871 Z_BEST_COMPRESSION);
2873 /* Set other zlib parameters for compressing IDAT */
2874 png_set_compression_mem_level(png_ptr, 8);
2875 png_set_compression_strategy(png_ptr,
2876 Z_DEFAULT_STRATEGY);
2877 png_set_compression_window_bits(png_ptr, 15);
2878 png_set_compression_method(png_ptr, 8);
2879 png_set_compression_buffer_size(png_ptr, 8192)
2881 /* Set zlib parameters for text compression
2882 * If you don't call these, the parameters
2883 * fall back on those defined for IDAT chunks
2885 png_set_text_compression_mem_level(png_ptr, 8);
2886 png_set_text_compression_strategy(png_ptr,
2887 Z_DEFAULT_STRATEGY);
2888 png_set_text_compression_window_bits(png_ptr, 15);
2889 png_set_text_compression_method(png_ptr, 8);
2891 Setting the contents of info for output
2893 You now need to fill in the png_info structure with all the data you
2894 wish to write before the actual image. Note that the only thing you
2895 are allowed to write after the image is the text chunks and the time
2896 chunk (as of PNG Specification 1.2, anyway). See png_write_end() and
2897 the latest PNG specification for more information on that. If you
2898 wish to write them before the image, fill them in now, and flag that
2899 data as being valid. If you want to wait until after the data, don't
2900 fill them until png_write_end(). For all the fields in png_info and
2901 their data types, see png.h. For explanations of what the fields
2902 contain, see the PNG specification.
2904 Some of the more important parts of the png_info are:
2906 png_set_IHDR(png_ptr, info_ptr, width, height,
2907 bit_depth, color_type, interlace_type,
2908 compression_type, filter_method)
2910 width - holds the width of the image
2911 in pixels (up to 2^31).
2913 height - holds the height of the image
2914 in pixels (up to 2^31).
2916 bit_depth - holds the bit depth of one of the
2918 (valid values are 1, 2, 4, 8, 16
2919 and depend also on the
2920 color_type. See also significant
2923 color_type - describes which color/alpha
2924 channels are present.
2926 (bit depths 1, 2, 4, 8, 16)
2927 PNG_COLOR_TYPE_GRAY_ALPHA
2929 PNG_COLOR_TYPE_PALETTE
2930 (bit depths 1, 2, 4, 8)
2933 PNG_COLOR_TYPE_RGB_ALPHA
2936 PNG_COLOR_MASK_PALETTE
2937 PNG_COLOR_MASK_COLOR
2938 PNG_COLOR_MASK_ALPHA
2940 interlace_type - PNG_INTERLACE_NONE or
2943 compression_type - (must be
2944 PNG_COMPRESSION_TYPE_DEFAULT)
2946 filter_method - (must be PNG_FILTER_TYPE_DEFAULT
2947 or, if you are writing a PNG to
2948 be embedded in a MNG datastream,
2950 PNG_INTRAPIXEL_DIFFERENCING)
2952 If you call png_set_IHDR(), the call must appear before any of the
2953 other png_set_*() functions, because they might require access to some of
2954 the IHDR settings. The remaining png_set_*() functions can be called
2957 If you wish, you can reset the compression_type, interlace_type, or
2958 filter_method later by calling png_set_IHDR() again; if you do this, the
2959 width, height, bit_depth, and color_type must be the same in each call.
2961 png_set_PLTE(png_ptr, info_ptr, palette,
2964 palette - the palette for the file
2965 (array of png_color)
2966 num_palette - number of entries in the palette
2969 png_set_gAMA(png_ptr, info_ptr, file_gamma);
2970 png_set_gAMA_fixed(png_ptr, info_ptr, int_file_gamma);
2972 file_gamma - the gamma at which the image was
2973 created (PNG_INFO_gAMA)
2975 int_file_gamma - 100,000 times the gamma at which
2976 the image was created
2978 png_set_cHRM(png_ptr, info_ptr, white_x, white_y, red_x, red_y,
2979 green_x, green_y, blue_x, blue_y)
2980 png_set_cHRM_XYZ(png_ptr, info_ptr, red_X, red_Y, red_Z, green_X,
2981 green_Y, green_Z, blue_X, blue_Y, blue_Z)
2982 png_set_cHRM_fixed(png_ptr, info_ptr, int_white_x, int_white_y,
2983 int_red_x, int_red_y, int_green_x, int_green_y,
2984 int_blue_x, int_blue_y)
2985 png_set_cHRM_XYZ_fixed(png_ptr, info_ptr, int_red_X, int_red_Y,
2986 int_red_Z, int_green_X, int_green_Y, int_green_Z,
2987 int_blue_X, int_blue_Y, int_blue_Z)
2989 {white,red,green,blue}_{x,y}
2990 A color space encoding specified using the chromaticities
2991 of the end points and the white point.
2993 {red,green,blue}_{X,Y,Z}
2994 A color space encoding specified using the encoding end
2995 points - the CIE tristimulus specification of the intended
2996 color of the red, green and blue channels in the PNG RGB
2997 data. The white point is simply the sum of the three end
3000 png_set_sRGB(png_ptr, info_ptr, srgb_intent);
3002 srgb_intent - the rendering intent
3003 (PNG_INFO_sRGB) The presence of
3004 the sRGB chunk means that the pixel
3005 data is in the sRGB color space.
3006 This chunk also implies specific
3007 values of gAMA and cHRM. Rendering
3008 intent is the CSS-1 property that
3009 has been defined by the International
3011 (http://www.color.org).
3013 PNG_sRGB_INTENT_SATURATION,
3014 PNG_sRGB_INTENT_PERCEPTUAL,
3015 PNG_sRGB_INTENT_ABSOLUTE, or
3016 PNG_sRGB_INTENT_RELATIVE.
3019 png_set_sRGB_gAMA_and_cHRM(png_ptr, info_ptr,
3022 srgb_intent - the rendering intent
3023 (PNG_INFO_sRGB) The presence of the
3024 sRGB chunk means that the pixel
3025 data is in the sRGB color space.
3026 This function also causes gAMA and
3027 cHRM chunks with the specific values
3028 that are consistent with sRGB to be
3031 png_set_iCCP(png_ptr, info_ptr, name, compression_type,
3034 name - The profile name.
3036 compression_type - The compression type; always
3037 PNG_COMPRESSION_TYPE_BASE for PNG 1.0.
3038 You may give NULL to this argument to
3041 profile - International Color Consortium color
3042 profile data. May contain NULs.
3044 proflen - length of profile data in bytes.
3046 png_set_sBIT(png_ptr, info_ptr, sig_bit);
3048 sig_bit - the number of significant bits for
3049 (PNG_INFO_sBIT) each of the gray, red,
3050 green, and blue channels, whichever are
3051 appropriate for the given color type
3054 png_set_tRNS(png_ptr, info_ptr, trans_alpha,
3055 num_trans, trans_color);
3057 trans_alpha - array of alpha (transparency)
3058 entries for palette (PNG_INFO_tRNS)
3060 num_trans - number of transparent entries
3063 trans_color - graylevel or color sample values
3064 (in order red, green, blue) of the
3065 single transparent color for
3066 non-paletted images (PNG_INFO_tRNS)
3068 png_set_hIST(png_ptr, info_ptr, hist);
3070 hist - histogram of palette (array of
3071 png_uint_16) (PNG_INFO_hIST)
3073 png_set_tIME(png_ptr, info_ptr, mod_time);
3075 mod_time - time image was last modified
3078 png_set_bKGD(png_ptr, info_ptr, background);
3080 background - background color (of type
3081 png_color_16p) (PNG_VALID_bKGD)
3083 png_set_text(png_ptr, info_ptr, text_ptr, num_text);
3085 text_ptr - array of png_text holding image
3088 text_ptr[i].compression - type of compression used
3089 on "text" PNG_TEXT_COMPRESSION_NONE
3090 PNG_TEXT_COMPRESSION_zTXt
3091 PNG_ITXT_COMPRESSION_NONE
3092 PNG_ITXT_COMPRESSION_zTXt
3093 text_ptr[i].key - keyword for comment. Must contain
3095 text_ptr[i].text - text comments for current
3096 keyword. Can be NULL or empty.
3097 text_ptr[i].text_length - length of text string,
3098 after decompression, 0 for iTXt
3099 text_ptr[i].itxt_length - length of itxt string,
3100 after decompression, 0 for tEXt/zTXt
3101 text_ptr[i].lang - language of comment (NULL or
3103 text_ptr[i].translated_keyword - keyword in UTF-8 (NULL
3104 or empty for unknown).
3106 Note that the itxt_length, lang, and lang_key
3107 members of the text_ptr structure only exist when the
3108 library is built with iTXt chunk support. Prior to
3109 libpng-1.4.0 the library was built by default without
3110 iTXt support. Also note that when iTXt is supported,
3111 they contain NULL pointers when the "compression"
3112 field contains PNG_TEXT_COMPRESSION_NONE or
3113 PNG_TEXT_COMPRESSION_zTXt.
3115 num_text - number of comments
3117 png_set_sPLT(png_ptr, info_ptr, &palette_ptr,
3120 palette_ptr - array of png_sPLT_struct structures
3121 to be added to the list of palettes
3122 in the info structure.
3123 num_spalettes - number of palette structures to be
3126 png_set_oFFs(png_ptr, info_ptr, offset_x, offset_y,
3129 offset_x - positive offset from the left
3132 offset_y - positive offset from the top
3135 unit_type - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER
3137 png_set_pHYs(png_ptr, info_ptr, res_x, res_y,
3140 res_x - pixels/unit physical resolution
3143 res_y - pixels/unit physical resolution
3146 unit_type - PNG_RESOLUTION_UNKNOWN,
3147 PNG_RESOLUTION_METER
3149 png_set_sCAL(png_ptr, info_ptr, unit, width, height)
3151 unit - physical scale units (an integer)
3153 width - width of a pixel in physical scale units
3155 height - height of a pixel in physical scale units
3156 (width and height are doubles)
3158 png_set_sCAL_s(png_ptr, info_ptr, unit, width, height)
3160 unit - physical scale units (an integer)
3162 width - width of a pixel in physical scale units
3163 expressed as a string
3165 height - height of a pixel in physical scale units
3166 (width and height are strings like "2.54")
3168 png_set_unknown_chunks(png_ptr, info_ptr, &unknowns,
3171 unknowns - array of png_unknown_chunk
3172 structures holding unknown chunks
3173 unknowns[i].name - name of unknown chunk
3174 unknowns[i].data - data of unknown chunk
3175 unknowns[i].size - size of unknown chunk's data
3176 unknowns[i].location - position to write chunk in file
3177 0: do not write chunk
3178 PNG_HAVE_IHDR: before PLTE
3179 PNG_HAVE_PLTE: before IDAT
3180 PNG_AFTER_IDAT: after IDAT
3182 The "location" member is set automatically according to
3183 what part of the output file has already been written.
3184 You can change its value after calling png_set_unknown_chunks()
3185 as demonstrated in pngtest.c. Within each of the "locations",
3186 the chunks are sequenced according to their position in the
3187 structure (that is, the value of "i", which is the order in which
3188 the chunk was either read from the input file or defined with
3189 png_set_unknown_chunks).
3191 A quick word about text and num_text. text is an array of png_text
3192 structures. num_text is the number of valid structures in the array.
3193 Each png_text structure holds a language code, a keyword, a text value,
3194 and a compression type.
3196 The compression types have the same valid numbers as the compression
3197 types of the image data. Currently, the only valid number is zero.
3198 However, you can store text either compressed or uncompressed, unlike
3199 images, which always have to be compressed. So if you don't want the
3200 text compressed, set the compression type to PNG_TEXT_COMPRESSION_NONE.
3201 Because tEXt and zTXt chunks don't have a language field, if you
3202 specify PNG_TEXT_COMPRESSION_NONE or PNG_TEXT_COMPRESSION_zTXt
3203 any language code or translated keyword will not be written out.
3205 Until text gets around a few hundred bytes, it is not worth compressing it.
3206 After the text has been written out to the file, the compression type
3207 is set to PNG_TEXT_COMPRESSION_NONE_WR or PNG_TEXT_COMPRESSION_zTXt_WR,
3208 so that it isn't written out again at the end (in case you are calling
3209 png_write_end() with the same struct).
3211 The keywords that are given in the PNG Specification are:
3213 Title Short (one line) title or
3216 Author Name of image's creator
3218 Description Description of image (possibly long)
3220 Copyright Copyright notice
3222 Creation Time Time of original image creation
3223 (usually RFC 1123 format, see below)
3225 Software Software used to create the image
3227 Disclaimer Legal disclaimer
3229 Warning Warning of nature of content
3231 Source Device used to create the image
3233 Comment Miscellaneous comment; conversion
3234 from other image format
3236 The keyword-text pairs work like this. Keywords should be short
3237 simple descriptions of what the comment is about. Some typical
3238 keywords are found in the PNG specification, as is some recommendations
3239 on keywords. You can repeat keywords in a file. You can even write
3240 some text before the image and some after. For example, you may want
3241 to put a description of the image before the image, but leave the
3242 disclaimer until after, so viewers working over modem connections
3243 don't have to wait for the disclaimer to go over the modem before
3244 they start seeing the image. Finally, keywords should be full
3245 words, not abbreviations. Keywords and text are in the ISO 8859-1
3246 (Latin-1) character set (a superset of regular ASCII) and can not
3247 contain NUL characters, and should not contain control or other
3248 unprintable characters. To make the comments widely readable, stick
3249 with basic ASCII, and avoid machine specific character set extensions
3250 like the IBM-PC character set. The keyword must be present, but
3251 you can leave off the text string on non-compressed pairs.
3252 Compressed pairs must have a text string, as only the text string
3253 is compressed anyway, so the compression would be meaningless.
3255 PNG supports modification time via the png_time structure. Two
3256 conversion routines are provided, png_convert_from_time_t() for
3257 time_t and png_convert_from_struct_tm() for struct tm. The
3258 time_t routine uses gmtime(). You don't have to use either of
3259 these, but if you wish to fill in the png_time structure directly,
3260 you should provide the time in universal time (GMT) if possible
3261 instead of your local time. Note that the year number is the full
3262 year (e.g. 1998, rather than 98 - PNG is year 2000 compliant!), and
3263 that months start with 1.
3265 If you want to store the time of the original image creation, you should
3266 use a plain tEXt chunk with the "Creation Time" keyword. This is
3267 necessary because the "creation time" of a PNG image is somewhat vague,
3268 depending on whether you mean the PNG file, the time the image was
3269 created in a non-PNG format, a still photo from which the image was
3270 scanned, or possibly the subject matter itself. In order to facilitate
3271 machine-readable dates, it is recommended that the "Creation Time"
3272 tEXt chunk use RFC 1123 format dates (e.g. "22 May 1997 18:07:10 GMT"),
3273 although this isn't a requirement. Unlike the tIME chunk, the
3274 "Creation Time" tEXt chunk is not expected to be automatically changed
3275 by the software. To facilitate the use of RFC 1123 dates, a function
3276 png_convert_to_rfc1123_buffer(buffer, png_timep) is provided to
3277 convert from PNG time to an RFC 1123 format string. The caller must provide
3278 a writeable buffer of at least 29 bytes.
3280 Writing unknown chunks
3282 You can use the png_set_unknown_chunks function to queue up private chunks
3283 for writing. You give it a chunk name, location, raw data, and a size. You
3284 also must use png_set_keep_unknown_chunks() to ensure that libpng will
3285 handle them. That's all there is to it. The chunks will be written by the
3286 next following png_write_info_before_PLTE, png_write_info, or png_write_end
3287 function, depending upon the specified location. Any chunks previously
3288 read into the info structure's unknown-chunk list will also be written out
3289 in a sequence that satisfies the PNG specification's ordering rules.
3291 Here is an example of writing two private chunks, prVt and miNE:
3293 #ifdef PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED
3294 /* Set unknown chunk data */
3295 png_unknown_chunk unk_chunk[2];
3296 strcpy((char *) unk_chunk[0].name, "prVt";
3297 unk_chunk[0].data = (unsigned char *) "PRIVATE DATA";
3298 unk_chunk[0].size = strlen(unk_chunk[0].data)+1;
3299 unk_chunk[0].location = PNG_HAVE_IHDR;
3300 strcpy((char *) unk_chunk[1].name, "miNE";
3301 unk_chunk[1].data = (unsigned char *) "MY CHUNK DATA";
3302 unk_chunk[1].size = strlen(unk_chunk[0].data)+1;
3303 unk_chunk[1].location = PNG_AFTER_IDAT;
3304 png_set_unknown_chunks(write_ptr, write_info_ptr,
3306 /* Needed because miNE is not safe-to-copy */
3307 png_set_keep_unknown_chunks(png, PNG_HANDLE_CHUNK_ALWAYS,
3308 (png_bytep) "miNE", 1);
3309 # if PNG_LIBPNG_VER < 10600
3310 /* Deal with unknown chunk location bug in 1.5.x and earlier */
3311 png_set_unknown_chunk_location(png, info, 0, PNG_HAVE_IHDR);
3312 png_set_unknown_chunk_location(png, info, 1, PNG_AFTER_IDAT);
3314 # if PNG_LIBPNG_VER < 10500
3315 /* PNG_AFTER_IDAT writes two copies of the chunk prior to libpng-1.5.0,
3316 * one before IDAT and another after IDAT, so don't use it; only use
3317 * PNG_HAVE_IHDR location. This call resets the location previously
3318 * set by assignment and png_set_unknown_chunk_location() for chunk 1.
3320 png_set_unknown_chunk_location(png, info, 1, PNG_HAVE_IHDR);
3324 The high-level write interface
3326 At this point there are two ways to proceed; through the high-level
3327 write interface, or through a sequence of low-level write operations.
3328 You can use the high-level interface if your image data is present
3329 in the info structure. All defined output
3330 transformations are permitted, enabled by the following masks.
3332 PNG_TRANSFORM_IDENTITY No transformation
3333 PNG_TRANSFORM_PACKING Pack 1, 2 and 4-bit samples
3334 PNG_TRANSFORM_PACKSWAP Change order of packed
3336 PNG_TRANSFORM_INVERT_MONO Invert monochrome images
3337 PNG_TRANSFORM_SHIFT Normalize pixels to the
3339 PNG_TRANSFORM_BGR Flip RGB to BGR, RGBA
3341 PNG_TRANSFORM_SWAP_ALPHA Flip RGBA to ARGB or GA
3343 PNG_TRANSFORM_INVERT_ALPHA Change alpha from opacity
3345 PNG_TRANSFORM_SWAP_ENDIAN Byte-swap 16-bit samples
3346 PNG_TRANSFORM_STRIP_FILLER Strip out filler
3348 PNG_TRANSFORM_STRIP_FILLER_BEFORE Strip out leading
3350 PNG_TRANSFORM_STRIP_FILLER_AFTER Strip out trailing
3353 If you have valid image data in the info structure (you can use
3354 png_set_rows() to put image data in the info structure), simply do this:
3356 png_write_png(png_ptr, info_ptr, png_transforms, NULL)
3358 where png_transforms is an integer containing the bitwise OR of some set of
3359 transformation flags. This call is equivalent to png_write_info(),
3360 followed the set of transformations indicated by the transform mask,
3361 then png_write_image(), and finally png_write_end().
3363 (The final parameter of this call is not yet used. Someday it might point
3364 to transformation parameters required by some future output transform.)
3366 You must use png_transforms and not call any png_set_transform() functions
3367 when you use png_write_png().
3369 The low-level write interface
3371 If you are going the low-level route instead, you are now ready to
3372 write all the file information up to the actual image data. You do
3373 this with a call to png_write_info().
3375 png_write_info(png_ptr, info_ptr);
3377 Note that there is one transformation you may need to do before
3378 png_write_info(). In PNG files, the alpha channel in an image is the
3379 level of opacity. If your data is supplied as a level of transparency,
3380 you can invert the alpha channel before you write it, so that 0 is
3381 fully transparent and 255 (in 8-bit or paletted images) or 65535
3382 (in 16-bit images) is fully opaque, with
3384 png_set_invert_alpha(png_ptr);
3386 This must appear before png_write_info() instead of later with the
3387 other transformations because in the case of paletted images the tRNS
3388 chunk data has to be inverted before the tRNS chunk is written. If
3389 your image is not a paletted image, the tRNS data (which in such cases
3390 represents a single color to be rendered as transparent) won't need to
3391 be changed, and you can safely do this transformation after your
3392 png_write_info() call.
3394 If you need to write a private chunk that you want to appear before
3395 the PLTE chunk when PLTE is present, you can write the PNG info in
3396 two steps, and insert code to write your own chunk between them:
3398 png_write_info_before_PLTE(png_ptr, info_ptr);
3399 png_set_unknown_chunks(png_ptr, info_ptr, ...);
3400 png_write_info(png_ptr, info_ptr);
3402 After you've written the file information, you can set up the library
3403 to handle any special transformations of the image data. The various
3404 ways to transform the data will be described in the order that they
3405 should occur. This is important, as some of these change the color
3406 type and/or bit depth of the data, and some others only work on
3407 certain color types and bit depths. Even though each transformation
3408 checks to see if it has data that it can do something with, you should
3409 make sure to only enable a transformation if it will be valid for the
3410 data. For example, don't swap red and blue on grayscale data.
3412 PNG files store RGB pixels packed into 3 or 6 bytes. This code tells
3413 the library to strip input data that has 4 or 8 bytes per pixel down
3414 to 3 or 6 bytes (or strip 2 or 4-byte grayscale+filler data to 1 or 2
3417 png_set_filler(png_ptr, 0, PNG_FILLER_BEFORE);
3419 where the 0 is unused, and the location is either PNG_FILLER_BEFORE or
3420 PNG_FILLER_AFTER, depending upon whether the filler byte in the pixel
3421 is stored XRGB or RGBX.
3423 PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as
3424 they can, resulting in, for example, 8 pixels per byte for 1 bit files.
3425 If the data is supplied at 1 pixel per byte, use this code, which will
3426 correctly pack the pixels into a single byte:
3428 png_set_packing(png_ptr);
3430 PNG files reduce possible bit depths to 1, 2, 4, 8, and 16. If your
3431 data is of another bit depth, you can write an sBIT chunk into the
3432 file so that decoders can recover the original data if desired.
3434 /* Set the true bit depth of the image data */
3435 if (color_type & PNG_COLOR_MASK_COLOR)
3437 sig_bit.red = true_bit_depth;
3438 sig_bit.green = true_bit_depth;
3439 sig_bit.blue = true_bit_depth;
3444 sig_bit.gray = true_bit_depth;
3447 if (color_type & PNG_COLOR_MASK_ALPHA)
3449 sig_bit.alpha = true_bit_depth;
3452 png_set_sBIT(png_ptr, info_ptr, &sig_bit);
3454 If the data is stored in the row buffer in a bit depth other than
3455 one supported by PNG (e.g. 3 bit data in the range 0-7 for a 4-bit PNG),
3456 this will scale the values to appear to be the correct bit depth as
3459 png_set_shift(png_ptr, &sig_bit);
3461 PNG files store 16-bit pixels in network byte order (big-endian,
3462 ie. most significant bits first). This code would be used if they are
3463 supplied the other way (little-endian, i.e. least significant bits
3464 first, the way PCs store them):
3467 png_set_swap(png_ptr);
3469 If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you
3470 need to change the order the pixels are packed into bytes, you can use:
3473 png_set_packswap(png_ptr);
3475 PNG files store 3 color pixels in red, green, blue order. This code
3476 would be used if they are supplied as blue, green, red:
3478 png_set_bgr(png_ptr);
3480 PNG files describe monochrome as black being zero and white being
3481 one. This code would be used if the pixels are supplied with this reversed
3482 (black being one and white being zero):
3484 png_set_invert_mono(png_ptr);
3486 Finally, you can write your own transformation function if none of
3487 the existing ones meets your needs. This is done by setting a callback
3490 png_set_write_user_transform_fn(png_ptr,
3491 write_transform_fn);
3493 You must supply the function
3495 void write_transform_fn(png_structp png_ptr, png_row_infop
3496 row_info, png_bytep data)
3498 See pngtest.c for a working example. Your function will be called
3499 before any of the other transformations are processed. If supported
3500 libpng also supplies an information routine that may be called from
3503 png_get_current_row_number(png_ptr);
3504 png_get_current_pass_number(png_ptr);
3506 This returns the current row passed to the transform. With interlaced
3507 images the value returned is the row in the input sub-image image. Use
3508 PNG_ROW_FROM_PASS_ROW(row, pass) and PNG_COL_FROM_PASS_COL(col, pass) to
3509 find the output pixel (x,y) given an interlaced sub-image pixel (row,col,pass).
3511 The discussion of interlace handling above contains more information on how to
3514 You can also set up a pointer to a user structure for use by your
3517 png_set_user_transform_info(png_ptr, user_ptr, 0, 0);
3519 The user_channels and user_depth parameters of this function are ignored
3520 when writing; you can set them to zero as shown.
3522 You can retrieve the pointer via the function png_get_user_transform_ptr().
3525 voidp write_user_transform_ptr =
3526 png_get_user_transform_ptr(png_ptr);
3528 It is possible to have libpng flush any pending output, either manually,
3529 or automatically after a certain number of lines have been written. To
3530 flush the output stream a single time call:
3532 png_write_flush(png_ptr);
3534 and to have libpng flush the output stream periodically after a certain
3535 number of scanlines have been written, call:
3537 png_set_flush(png_ptr, nrows);
3539 Note that the distance between rows is from the last time png_write_flush()
3540 was called, or the first row of the image if it has never been called.
3541 So if you write 50 lines, and then png_set_flush 25, it will flush the
3542 output on the next scanline, and every 25 lines thereafter, unless
3543 png_write_flush() is called before 25 more lines have been written.
3544 If nrows is too small (less than about 10 lines for a 640 pixel wide
3545 RGB image) the image compression may decrease noticeably (although this
3546 may be acceptable for real-time applications). Infrequent flushing will
3547 only degrade the compression performance by a few percent over images
3548 that do not use flushing.
3550 Writing the image data
3552 That's it for the transformations. Now you can write the image data.
3553 The simplest way to do this is in one function call. If you have the
3554 whole image in memory, you can just call png_write_image() and libpng
3555 will write the image. You will need to pass in an array of pointers to
3556 each row. This function automatically handles interlacing, so you don't
3557 need to call png_set_interlace_handling() or call this function multiple
3558 times, or any of that other stuff necessary with png_write_rows().
3560 png_write_image(png_ptr, row_pointers);
3562 where row_pointers is:
3564 png_byte *row_pointers[height];
3566 You can point to void or char or whatever you use for pixels.
3568 If you don't want to write the whole image at once, you can
3569 use png_write_rows() instead. If the file is not interlaced,
3572 png_write_rows(png_ptr, row_pointers,
3575 row_pointers is the same as in the png_write_image() call.
3577 If you are just writing one row at a time, you can do this with
3578 a single row_pointer instead of an array of row_pointers:
3580 png_bytep row_pointer = row;
3582 png_write_row(png_ptr, row_pointer);
3584 When the file is interlaced, things can get a good deal more complicated.
3585 The only currently (as of the PNG Specification version 1.2, dated July
3586 1999) defined interlacing scheme for PNG files is the "Adam7" interlace
3587 scheme, that breaks down an image into seven smaller images of varying
3588 size. libpng will build these images for you, or you can do them
3589 yourself. If you want to build them yourself, see the PNG specification
3590 for details of which pixels to write when.
3592 If you don't want libpng to handle the interlacing details, just
3593 use png_set_interlace_handling() and call png_write_rows() the
3594 correct number of times to write all the sub-images
3595 (png_set_interlace_handling() returns the number of sub-images.)
3597 If you want libpng to build the sub-images, call this before you start
3600 number_of_passes = png_set_interlace_handling(png_ptr);
3602 This will return the number of passes needed. Currently, this is seven,
3603 but may change if another interlace type is added.
3605 Then write the complete image number_of_passes times.
3607 png_write_rows(png_ptr, row_pointers, number_of_rows);
3609 Think carefully before you write an interlaced image. Typically code that
3610 reads such images reads all the image data into memory, uncompressed, before
3611 doing any processing. Only code that can display an image on the fly can
3612 take advantage of the interlacing and even then the image has to be exactly
3613 the correct size for the output device, because scaling an image requires
3614 adjacent pixels and these are not available until all the passes have been
3617 If you do write an interlaced image you will hardly ever need to handle
3618 the interlacing yourself. Call png_set_interlace_handling() and use the
3619 approach described above.
3621 The only time it is conceivable that you will really need to write an
3622 interlaced image pass-by-pass is when you have read one pass by pass and
3623 made some pixel-by-pixel transformation to it, as described in the read
3624 code above. In this case use the PNG_PASS_ROWS and PNG_PASS_COLS macros
3625 to determine the size of each sub-image in turn and simply write the rows
3626 you obtained from the read code.
3628 Finishing a sequential write
3630 After you are finished writing the image, you should finish writing
3631 the file. If you are interested in writing comments or time, you should
3632 pass an appropriately filled png_info pointer. If you are not interested,
3635 png_write_end(png_ptr, info_ptr);
3637 When you are done, you can free all memory used by libpng like this:
3639 png_destroy_write_struct(&png_ptr, &info_ptr);
3641 It is also possible to individually free the info_ptr members that
3642 point to libpng-allocated storage with the following function:
3644 png_free_data(png_ptr, info_ptr, mask, seq)
3646 mask - identifies data to be freed, a mask
3647 containing the bitwise OR of one or
3649 PNG_FREE_PLTE, PNG_FREE_TRNS,
3650 PNG_FREE_HIST, PNG_FREE_ICCP,
3651 PNG_FREE_PCAL, PNG_FREE_ROWS,
3652 PNG_FREE_SCAL, PNG_FREE_SPLT,
3653 PNG_FREE_TEXT, PNG_FREE_UNKN,
3654 or simply PNG_FREE_ALL
3656 seq - sequence number of item to be freed
3659 This function may be safely called when the relevant storage has
3660 already been freed, or has not yet been allocated, or was allocated
3661 by the user and not by libpng, and will in those cases do nothing.
3662 The "seq" parameter is ignored if only one item of the selected data
3663 type, such as PLTE, is allowed. If "seq" is not -1, and multiple items
3664 are allowed for the data type identified in the mask, such as text or
3665 sPLT, only the n'th item in the structure is freed, where n is "seq".
3667 If you allocated data such as a palette that you passed in to libpng
3668 with png_set_*, you must not free it until just before the call to
3669 png_destroy_write_struct().
3671 The default behavior is only to free data that was allocated internally
3672 by libpng. This can be changed, so that libpng will not free the data,
3673 or so that it will free data that was allocated by the user with png_malloc()
3674 or png_calloc() and passed in via a png_set_*() function, with
3676 png_data_freer(png_ptr, info_ptr, freer, mask)
3679 PNG_DESTROY_WILL_FREE_DATA
3680 PNG_SET_WILL_FREE_DATA
3681 PNG_USER_WILL_FREE_DATA
3683 mask - which data elements are affected
3684 same choices as in png_free_data()
3686 For example, to transfer responsibility for some data from a read structure
3687 to a write structure, you could use
3689 png_data_freer(read_ptr, read_info_ptr,
3690 PNG_USER_WILL_FREE_DATA,
3691 PNG_FREE_PLTE|PNG_FREE_tRNS|PNG_FREE_hIST)
3693 png_data_freer(write_ptr, write_info_ptr,
3694 PNG_DESTROY_WILL_FREE_DATA,
3695 PNG_FREE_PLTE|PNG_FREE_tRNS|PNG_FREE_hIST)
3697 thereby briefly reassigning responsibility for freeing to the user but
3698 immediately afterwards reassigning it once more to the write_destroy
3699 function. Having done this, it would then be safe to destroy the read
3700 structure and continue to use the PLTE, tRNS, and hIST data in the write
3703 This function only affects data that has already been allocated.
3704 You can call this function before calling after the png_set_*() functions
3705 to control whether the user or png_destroy_*() is supposed to free the data.
3706 When the user assumes responsibility for libpng-allocated data, the
3707 application must use
3708 png_free() to free it, and when the user transfers responsibility to libpng
3709 for data that the user has allocated, the user must have used png_malloc()
3710 or png_calloc() to allocate it.
3712 If you allocated text_ptr.text, text_ptr.lang, and text_ptr.translated_keyword
3713 separately, do not transfer responsibility for freeing text_ptr to libpng,
3714 because when libpng fills a png_text structure it combines these members with
3715 the key member, and png_free_data() will free only text_ptr.key. Similarly,
3716 if you transfer responsibility for free'ing text_ptr from libpng to your
3717 application, your application must not separately free those members.
3718 For a more compact example of writing a PNG image, see the file example.c.
3722 The simplified API, which became available in libpng-1.6.0, hides the details
3723 of both libpng and the PNG file format itself.
3724 It allows PNG files to be read into a very limited number of
3725 in-memory bitmap formats or to be written from the same formats. If these
3726 formats do not accommodate your needs then you can, and should, use the more
3727 sophisticated APIs above - these support a wide variety of in-memory formats
3728 and a wide variety of sophisticated transformations to those formats as well
3729 as a wide variety of APIs to manipulate ancilliary information.
3731 To read a PNG file using the simplified API:
3733 1) Declare a 'png_image' structure (see below) on the stack, set the
3734 version field to PNG_IMAGE_VERSION and the 'opaque' pointer to NULL
3735 (this is REQUIRED, your program may crash if you don't do it.)
3737 2) Call the appropriate png_image_begin_read... function.
3739 3) Set the png_image 'format' member to the required sample format.
3741 4) Allocate a buffer for the image and, if required, the color-map.
3743 5) Call png_image_finish_read to read the image and, if required, the
3744 color-map into your buffers.
3746 There are no restrictions on the format of the PNG input itself; all valid
3747 color types, bit depths, and interlace methods are acceptable, and the
3748 input image is transformed as necessary to the requested in-memory format
3749 during the png_image_finish_read() step. The only caveat is that if you
3750 request a color-mapped image from a PNG that is full-color or makes
3751 complex use of an alpha channel the transformation is extremely lossy and the
3752 result may look terrible.
3754 To write a PNG file using the simplified API:
3756 1) Declare a 'png_image' structure on the stack and memset()
3759 2) Initialize the members of the structure that describe the
3760 image, setting the 'format' member to the format of the
3763 3) Call the appropriate png_image_write... function with a
3764 pointer to the image and, if necessary, the color-map to write
3767 png_image is a structure that describes the in-memory format of an image
3768 when it is being read or defines the in-memory format of an image that you
3769 need to write. The "png_image" structure contains the following members:
3771 png_controlp opaque Initialize to NULL, free with png_image_free
3772 png_uint_32 version Set to PNG_IMAGE_VERSION
3773 png_uint_32 width Image width in pixels (columns)
3774 png_uint_32 height Image height in pixels (rows)
3775 png_uint_32 format Image format as defined below
3776 png_uint_32 flags A bit mask containing informational flags
3777 png_uint_32 colormap_entries; Number of entries in the color-map
3778 png_uint_32 warning_or_error;
3781 In the event of an error or warning the "warning_or_error"
3782 field will be set to a non-zero value and the 'message' field will contain
3783 a '\0' terminated string with the libpng error or warning message. If both
3784 warnings and an error were encountered, only the error is recorded. If there
3785 are multiple warnings, only the first one is recorded.
3787 The upper 30 bits of the "warning_or_error" value are reserved; the low two
3788 bits contain a two bit code such that a value more than 1 indicates a failure
3789 in the API just called:
3791 0 - no warning or error
3794 3 - error preceded by warning
3796 The pixels (samples) of the image have one to four channels whose components
3797 have original values in the range 0 to 1.0:
3799 1: A single gray or luminance channel (G).
3800 2: A gray/luminance channel and an alpha channel (GA).
3801 3: Three red, green, blue color channels (RGB).
3802 4: Three color channels and an alpha channel (RGBA).
3804 The channels are encoded in one of two ways:
3806 a) As a small integer, value 0..255, contained in a single byte. For the
3807 alpha channel the original value is simply value/255. For the color or
3808 luminance channels the value is encoded according to the sRGB specification
3809 and matches the 8-bit format expected by typical display devices.
3811 The color/gray channels are not scaled (pre-multiplied) by the alpha
3812 channel and are suitable for passing to color management software.
3814 b) As a value in the range 0..65535, contained in a 2-byte integer, in
3815 the native byte order of the platform on which the application is running.
3816 All channels can be converted to the original value by dividing by 65535; all
3817 channels are linear. Color channels use the RGB encoding (RGB end-points) of
3818 the sRGB specification. This encoding is identified by the
3819 PNG_FORMAT_FLAG_LINEAR flag below.
3821 When the simplified API needs to convert between sRGB and linear colorspaces,
3822 the actual sRGB transfer curve defined in the sRGB specification (see the
3823 article at http://en.wikipedia.org/wiki/SRGB) is used, not the gamma=1/2.2
3824 approximation used elsewhere in libpng.
3826 When an alpha channel is present it is expected to denote pixel coverage
3827 of the color or luminance channels and is returned as an associated alpha
3828 channel: the color/gray channels are scaled (pre-multiplied) by the alpha
3831 The samples are either contained directly in the image data, between 1 and 8
3832 bytes per pixel according to the encoding, or are held in a color-map indexed
3833 by bytes in the image data. In the case of a color-map the color-map entries
3834 are individual samples, encoded as above, and the image data has one byte per
3835 pixel to select the relevant sample from the color-map.
3839 The #defines to be used in png_image::format. Each #define identifies a
3840 particular layout of channel data and, if present, alpha values. There are
3841 separate defines for each of the two component encodings.
3843 A format is built up using single bit flag values. All combinations are
3844 valid. Formats can be built up from the flag values or you can use one of
3845 the predefined values below. When testing formats always use the FORMAT_FLAG
3846 macros to test for individual features - future versions of the library may
3849 When reading or writing color-mapped images the format should be set to the
3850 format of the entries in the color-map then png_image_{read,write}_colormap
3851 called to read or write the color-map and set the format correctly for the
3852 image data. Do not set the PNG_FORMAT_FLAG_COLORMAP bit directly!
3854 NOTE: libpng can be built with particular features disabled. If you see
3855 compiler errors because the definition of one of the following flags has been
3856 compiled out it is because libpng does not have the required support. It is
3857 possible, however, for the libpng configuration to enable the format on just
3858 read or just write; in that case you may see an error at run time.
3859 You can guard against this by checking for the definition of the
3860 appropriate "_SUPPORTED" macro, one of:
3862 PNG_SIMPLIFIED_{READ,WRITE}_{BGR,AFIRST}_SUPPORTED
3864 PNG_FORMAT_FLAG_ALPHA format with an alpha channel
3865 PNG_FORMAT_FLAG_COLOR color format: otherwise grayscale
3866 PNG_FORMAT_FLAG_LINEAR 2-byte channels else 1-byte
3867 PNG_FORMAT_FLAG_COLORMAP image data is color-mapped
3868 PNG_FORMAT_FLAG_BGR BGR colors, else order is RGB
3869 PNG_FORMAT_FLAG_AFIRST alpha channel comes first
3871 Supported formats are as follows. Future versions of libpng may support more
3872 formats; for compatibility with older versions simply check if the format
3873 macro is defined using #ifdef. These defines describe the in-memory layout
3874 of the components of the pixels of the image.
3876 First the single byte (sRGB) formats:
3888 Then the linear 2-byte formats. When naming these "Y" is used to
3889 indicate a luminance (gray) channel. The component order within the pixel
3890 is always the same - there is no provision for swapping the order of the
3891 components in the linear format. The components are 16-bit integers in
3892 the native byte order for your platform, and there is no provision for
3893 swapping the bytes to a different endian condition.
3896 PNG_FORMAT_LINEAR_Y_ALPHA
3897 PNG_FORMAT_LINEAR_RGB
3898 PNG_FORMAT_LINEAR_RGB_ALPHA
3900 With color-mapped formats the image data is one byte for each pixel. The byte
3901 is an index into the color-map which is formatted as above. To obtain a
3902 color-mapped format it is sufficient just to add the PNG_FOMAT_FLAG_COLORMAP
3903 to one of the above definitions, or you can use one of the definitions below.
3905 PNG_FORMAT_RGB_COLORMAP
3906 PNG_FORMAT_BGR_COLORMAP
3907 PNG_FORMAT_RGBA_COLORMAP
3908 PNG_FORMAT_ARGB_COLORMAP
3909 PNG_FORMAT_BGRA_COLORMAP
3910 PNG_FORMAT_ABGR_COLORMAP
3914 These are convenience macros to derive information from a png_image
3915 structure. The PNG_IMAGE_SAMPLE_ macros return values appropriate to the
3916 actual image sample values - either the entries in the color-map or the
3917 pixels in the image. The PNG_IMAGE_PIXEL_ macros return corresponding values
3918 for the pixels and will always return 1 for color-mapped formats. The
3919 remaining macros return information about the rows in the image and the
3922 NOTE: All the macros that take a png_image::format parameter are compile time
3923 constants if the format parameter is, itself, a constant. Therefore these
3924 macros can be used in array declarations and case labels where required.
3925 Similarly the macros are also pre-processor constants (sizeof is not used) so
3926 they can be used in #if tests.
3928 PNG_IMAGE_SAMPLE_CHANNELS(fmt)
3929 Returns the total number of channels in a given format: 1..4
3931 PNG_IMAGE_SAMPLE_COMPONENT_SIZE(fmt)
3932 Returns the size in bytes of a single component of a pixel or color-map
3933 entry (as appropriate) in the image: 1 or 2.
3935 PNG_IMAGE_SAMPLE_SIZE(fmt)
3936 This is the size of the sample data for one sample. If the image is
3937 color-mapped it is the size of one color-map entry (and image pixels are
3938 one byte in size), otherwise it is the size of one image pixel.
3940 PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(fmt)
3941 The maximum size of the color-map required by the format expressed in a
3942 count of components. This can be used to compile-time allocate a
3945 png_uint_16 colormap[PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(linear_fmt)];
3947 png_byte colormap[PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(sRGB_fmt)];
3949 Alternatively use the PNG_IMAGE_COLORMAP_SIZE macro below to use the
3950 information from one of the png_image_begin_read_ APIs and dynamically
3951 allocate the required memory.
3953 PNG_IMAGE_COLORMAP_SIZE(fmt)
3954 The size of the color-map required by the format; this is the size of the
3955 color-map buffer passed to the png_image_{read,write}_colormap APIs. It is
3956 a fixed number determined by the format so can easily be allocated on the
3959 Corresponding information about the pixels
3961 PNG_IMAGE_PIXEL_CHANNELS(fmt)
3962 The number of separate channels (components) in a pixel; 1 for a
3965 PNG_IMAGE_PIXEL_COMPONENT_SIZE(fmt)\
3966 The size, in bytes, of each component in a pixel; 1 for a color-mapped
3969 PNG_IMAGE_PIXEL_SIZE(fmt)
3970 The size, in bytes, of a complete pixel; 1 for a color-mapped image.
3972 Information about the whole row, or whole image
3974 PNG_IMAGE_ROW_STRIDE(image)
3975 Returns the total number of components in a single row of the image; this
3976 is the minimum 'row stride', the minimum count of components between each
3977 row. For a color-mapped image this is the minimum number of bytes in a
3980 If you need the stride measured in bytes, row_stride_bytes is
3981 PNG_IMAGE_ROW_STRIDE(image) * PNG_IMAGE_PIXEL_COMPONENT_SIZE(fmt)
3982 plus any padding bytes that your application might need, for example
3983 to start the next row on a 4-byte boundary.
3985 PNG_IMAGE_BUFFER_SIZE(image, row_stride)
3986 Return the size, in bytes, of an image buffer given a png_image and a row
3987 stride - the number of components to leave space for in each row.
3989 PNG_IMAGE_SIZE(image)
3990 Return the size, in bytes, of the image in memory given just a png_image;
3991 the row stride is the minimum stride required for the image.
3993 PNG_IMAGE_COLORMAP_SIZE(image)
3994 Return the size, in bytes, of the color-map of this image. If the image
3995 format is not a color-map format this will return a size sufficient for
3996 256 entries in the given format; check PNG_FORMAT_FLAG_COLORMAP if
3997 you don't want to allocate a color-map in this case.
4001 Flags containing additional information about the image are held in
4002 the 'flags' field of png_image.
4004 PNG_IMAGE_FLAG_COLORSPACE_NOT_sRGB == 0x01
4005 This indicates the the RGB values of the in-memory bitmap do not
4006 correspond to the red, green and blue end-points defined by sRGB.
4008 PNG_IMAGE_FLAG_FAST == 0x02
4009 On write emphasise speed over compression; the resultant PNG file will be
4010 larger but will be produced significantly faster, particular for large
4011 images. Do not use this option for images which will be distributed, only
4012 used it when producing intermediate files that will be read back in
4013 repeatedly. For a typical 24-bit image the option will double the read
4014 speed at the cost of increasing the image size by 25%, however for many
4015 more compressible images the PNG file can be 10 times larger with only a
4018 PNG_IMAGE_FLAG_16BIT_sRGB == 0x04
4019 On read if the image is a 16-bit per component image and there is no gAMA
4020 or sRGB chunk assume that the components are sRGB encoded. Notice that
4021 images output by the simplified API always have gamma information; setting
4022 this flag only affects the interpretation of 16-bit images from an
4023 external source. It is recommended that the application expose this flag
4024 to the user; the user can normally easily recognize the difference between
4025 linear and sRGB encoding. This flag has no effect on write - the data
4026 passed to the write APIs must have the correct encoding (as defined
4029 If the flag is not set (the default) input 16-bit per component data is
4030 assumed to be linear.
4032 NOTE: the flag can only be set after the png_image_begin_read_ call,
4033 because that call initializes the 'flags' field.
4037 The png_image passed to the read APIs must have been initialized by setting
4038 the png_controlp field 'opaque' to NULL (or, better, memset the whole thing.)
4040 int png_image_begin_read_from_file( png_imagep image,
4041 const char *file_name)
4043 The named file is opened for read and the image header
4044 is filled in from the PNG header in the file.
4046 int png_image_begin_read_from_stdio (png_imagep image,
4049 The PNG header is read from the stdio FILE object.
4051 int png_image_begin_read_from_memory(png_imagep image,
4052 png_const_voidp memory, png_size_t size)
4054 The PNG header is read from the given memory buffer.
4056 int png_image_finish_read(png_imagep image,
4057 png_colorp background, void *buffer,
4058 png_int_32 row_stride, void *colormap));
4060 Finish reading the image into the supplied buffer and
4061 clean up the png_image structure.
4063 row_stride is the step, in png_byte or png_uint_16 units
4064 as appropriate, between adjacent rows. A positive stride
4065 indicates that the top-most row is first in the buffer -
4066 the normal top-down arrangement. A negative stride
4067 indicates that the bottom-most row is first in the buffer.
4069 background need only be supplied if an alpha channel must
4070 be removed from a png_byte format and the removal is to be
4071 done by compositing on a solid color; otherwise it may be
4072 NULL and any composition will be done directly onto the
4073 buffer. The value is an sRGB color to use for the
4074 background, for grayscale output the green channel is used.
4076 For linear output removing the alpha channel is always done
4077 by compositing on black.
4079 void png_image_free(png_imagep image)
4081 Free any data allocated by libpng in image->opaque,
4082 setting the pointer to NULL. May be called at any time
4083 after the structure is initialized.
4085 When the simplified API needs to convert between sRGB and linear colorspaces,
4086 the actual sRGB transfer curve defined in the sRGB specification (see the
4087 article at http://en.wikipedia.org/wiki/SRGB) is used, not the gamma=1/2.2
4088 approximation used elsewhere in libpng.
4092 For write you must initialize a png_image structure to describe the image to
4095 version: must be set to PNG_IMAGE_VERSION
4096 opaque: must be initialized to NULL
4097 width: image width in pixels
4098 height: image height in rows
4099 format: the format of the data you wish to write
4100 flags: set to 0 unless one of the defined flags applies; set
4101 PNG_IMAGE_FLAG_COLORSPACE_NOT_sRGB for color format images
4102 where the RGB values do not correspond to the colors in sRGB.
4103 colormap_entries: set to the number of entries in the color-map (0 to 256)
4105 int png_image_write_to_file, (png_imagep image,
4106 const char *file, int convert_to_8bit, const void *buffer,
4107 png_int_32 row_stride, const void *colormap));
4109 Write the image to the named file.
4111 int png_image_write_to_memory (png_imagep image, void *memory,
4112 png_alloc_size_t * PNG_RESTRICT memory_bytes,
4113 int convert_to_8_bit, const void *buffer, ptrdiff_t row_stride,
4114 const void *colormap));
4116 Write the image to memory.
4118 int png_image_write_to_stdio(png_imagep image, FILE *file,
4119 int convert_to_8_bit, const void *buffer,
4120 png_int_32 row_stride, const void *colormap)
4122 Write the image to the given (FILE*).
4124 With all write APIs if image is in one of the linear formats with
4125 (png_uint_16) data then setting convert_to_8_bit will cause the output to be
4126 a (png_byte) PNG gamma encoded according to the sRGB specification, otherwise
4127 a 16-bit linear encoded PNG file is written.
4129 With all APIs row_stride is handled as in the read APIs - it is the spacing
4130 from one row to the next in component sized units (float) and if negative
4131 indicates a bottom-up row layout in the buffer. If you pass zero, libpng will
4132 calculate the row_stride for you from the width and number of channels.
4134 Note that the write API does not support interlacing, sub-8-bit pixels,
4135 indexed (paletted) images, or most ancillary chunks.
4137 VI. Modifying/Customizing libpng
4139 There are two issues here. The first is changing how libpng does
4140 standard things like memory allocation, input/output, and error handling.
4141 The second deals with more complicated things like adding new chunks,
4142 adding new transformations, and generally changing how libpng works.
4143 Both of those are compile-time issues; that is, they are generally
4144 determined at the time the code is written, and there is rarely a need
4145 to provide the user with a means of changing them.
4147 Memory allocation, input/output, and error handling
4149 All of the memory allocation, input/output, and error handling in libpng
4150 goes through callbacks that are user-settable. The default routines are
4151 in pngmem.c, pngrio.c, pngwio.c, and pngerror.c, respectively. To change
4152 these functions, call the appropriate png_set_*_fn() function.
4154 Memory allocation is done through the functions png_malloc(), png_calloc(),
4155 and png_free(). The png_malloc() and png_free() functions currently just
4156 call the standard C functions and png_calloc() calls png_malloc() and then
4157 clears the newly allocated memory to zero; note that png_calloc(png_ptr, size)
4158 is not the same as the calloc(number, size) function provided by stdlib.h.
4159 There is limited support for certain systems with segmented memory
4160 architectures and the types of pointers declared by png.h match this; you
4161 will have to use appropriate pointers in your application. If you prefer
4162 to use a different method of allocating and freeing data, you can use
4163 png_create_read_struct_2() or png_create_write_struct_2() to register your
4164 own functions as described above. These functions also provide a void
4165 pointer that can be retrieved via
4167 mem_ptr=png_get_mem_ptr(png_ptr);
4169 Your replacement memory functions must have prototypes as follows:
4171 png_voidp malloc_fn(png_structp png_ptr,
4172 png_alloc_size_t size);
4174 void free_fn(png_structp png_ptr, png_voidp ptr);
4176 Your malloc_fn() must return NULL in case of failure. The png_malloc()
4177 function will normally call png_error() if it receives a NULL from the
4178 system memory allocator or from your replacement malloc_fn().
4180 Your free_fn() will never be called with a NULL ptr, since libpng's
4181 png_free() checks for NULL before calling free_fn().
4183 Input/Output in libpng is done through png_read() and png_write(),
4184 which currently just call fread() and fwrite(). The FILE * is stored in
4185 png_struct and is initialized via png_init_io(). If you wish to change
4186 the method of I/O, the library supplies callbacks that you can set
4187 through the function png_set_read_fn() and png_set_write_fn() at run
4188 time, instead of calling the png_init_io() function. These functions
4189 also provide a void pointer that can be retrieved via the function
4190 png_get_io_ptr(). For example:
4192 png_set_read_fn(png_structp read_ptr,
4193 voidp read_io_ptr, png_rw_ptr read_data_fn)
4195 png_set_write_fn(png_structp write_ptr,
4196 voidp write_io_ptr, png_rw_ptr write_data_fn,
4197 png_flush_ptr output_flush_fn);
4199 voidp read_io_ptr = png_get_io_ptr(read_ptr);
4200 voidp write_io_ptr = png_get_io_ptr(write_ptr);
4202 The replacement I/O functions must have prototypes as follows:
4204 void user_read_data(png_structp png_ptr,
4205 png_bytep data, png_size_t length);
4207 void user_write_data(png_structp png_ptr,
4208 png_bytep data, png_size_t length);
4210 void user_flush_data(png_structp png_ptr);
4212 The user_read_data() function is responsible for detecting and
4213 handling end-of-data errors.
4215 Supplying NULL for the read, write, or flush functions sets them back
4216 to using the default C stream functions, which expect the io_ptr to
4217 point to a standard *FILE structure. It is probably a mistake
4218 to use NULL for one of write_data_fn and output_flush_fn but not both
4219 of them, unless you have built libpng with PNG_NO_WRITE_FLUSH defined.
4220 It is an error to read from a write stream, and vice versa.
4222 Error handling in libpng is done through png_error() and png_warning().
4223 Errors handled through png_error() are fatal, meaning that png_error()
4224 should never return to its caller. Currently, this is handled via
4225 setjmp() and longjmp() (unless you have compiled libpng with
4226 PNG_NO_SETJMP, in which case it is handled via PNG_ABORT()),
4227 but you could change this to do things like exit() if you should wish,
4228 as long as your function does not return.
4230 On non-fatal errors, png_warning() is called
4231 to print a warning message, and then control returns to the calling code.
4232 By default png_error() and png_warning() print a message on stderr via
4233 fprintf() unless the library is compiled with PNG_NO_CONSOLE_IO defined
4234 (because you don't want the messages) or PNG_NO_STDIO defined (because
4235 fprintf() isn't available). If you wish to change the behavior of the error
4236 functions, you will need to set up your own message callbacks. These
4237 functions are normally supplied at the time that the png_struct is created.
4238 It is also possible to redirect errors and warnings to your own replacement
4239 functions after png_create_*_struct() has been called by calling:
4241 png_set_error_fn(png_structp png_ptr,
4242 png_voidp error_ptr, png_error_ptr error_fn,
4243 png_error_ptr warning_fn);
4245 png_voidp error_ptr = png_get_error_ptr(png_ptr);
4247 If NULL is supplied for either error_fn or warning_fn, then the libpng
4248 default function will be used, calling fprintf() and/or longjmp() if a
4249 problem is encountered. The replacement error functions should have
4250 parameters as follows:
4252 void user_error_fn(png_structp png_ptr,
4253 png_const_charp error_msg);
4255 void user_warning_fn(png_structp png_ptr,
4256 png_const_charp warning_msg);
4258 The motivation behind using setjmp() and longjmp() is the C++ throw and
4259 catch exception handling methods. This makes the code much easier to write,
4260 as there is no need to check every return code of every function call.
4261 However, there are some uncertainties about the status of local variables
4262 after a longjmp, so the user may want to be careful about doing anything
4263 after setjmp returns non-zero besides returning itself. Consult your
4264 compiler documentation for more details. For an alternative approach, you
4265 may wish to use the "cexcept" facility (see http://cexcept.sourceforge.net),
4266 which is illustrated in pngvalid.c and in contrib/visupng.
4268 Beginning in libpng-1.4.0, the png_set_benign_errors() API became available.
4269 You can use this to handle certain errors (normally handled as errors)
4272 png_set_benign_errors (png_ptr, int allowed);
4274 allowed: 0: treat png_benign_error() as an error.
4275 1: treat png_benign_error() as a warning.
4277 As of libpng-1.6.0, the default condition is to treat benign errors as
4278 warnings while reading and as errors while writing.
4282 If you need to read or write custom chunks, you may need to get deeper
4283 into the libpng code. The library now has mechanisms for storing
4284 and writing chunks of unknown type; you can even declare callbacks
4285 for custom chunks. However, this may not be good enough if the
4286 library code itself needs to know about interactions between your
4287 chunk and existing `intrinsic' chunks.
4289 If you need to write a new intrinsic chunk, first read the PNG
4290 specification. Acquire a first level of understanding of how it works.
4291 Pay particular attention to the sections that describe chunk names,
4292 and look at how other chunks were designed, so you can do things
4293 similarly. Second, check out the sections of libpng that read and
4294 write chunks. Try to find a chunk that is similar to yours and use
4295 it as a template. More details can be found in the comments inside
4296 the code. It is best to handle private or unknown chunks in a generic method,
4297 via callback functions, instead of by modifying libpng functions. This
4298 is illustrated in pngtest.c, which uses a callback function to handle a
4299 private "vpAg" chunk and the new "sTER" chunk, which are both unknown to
4302 If you wish to write your own transformation for the data, look through
4303 the part of the code that does the transformations, and check out some of
4304 the simpler ones to get an idea of how they work. Try to find a similar
4305 transformation to the one you want to add and copy off of it. More details
4306 can be found in the comments inside the code itself.
4308 Configuring for gui/windowing platforms:
4310 You will need to write new error and warning functions that use the GUI
4311 interface, as described previously, and set them to be the error and
4312 warning functions at the time that png_create_*_struct() is called,
4313 in order to have them available during the structure initialization.
4314 They can be changed later via png_set_error_fn(). On some compilers,
4315 you may also have to change the memory allocators (png_malloc, etc.).
4319 There are special functions to configure the compression. Perhaps the
4320 most useful one changes the compression level, which currently uses
4321 input compression values in the range 0 - 9. The library normally
4322 uses the default compression level (Z_DEFAULT_COMPRESSION = 6). Tests
4323 have shown that for a large majority of images, compression values in
4324 the range 3-6 compress nearly as well as higher levels, and do so much
4325 faster. For online applications it may be desirable to have maximum speed
4326 (Z_BEST_SPEED = 1). With versions of zlib after v0.99, you can also
4327 specify no compression (Z_NO_COMPRESSION = 0), but this would create
4328 files larger than just storing the raw bitmap. You can specify the
4329 compression level by calling:
4332 png_set_compression_level(png_ptr, level);
4334 Another useful one is to reduce the memory level used by the library.
4335 The memory level defaults to 8, but it can be lowered if you are
4336 short on memory (running DOS, for example, where you only have 640K).
4337 Note that the memory level does have an effect on compression; among
4338 other things, lower levels will result in sections of incompressible
4339 data being emitted in smaller stored blocks, with a correspondingly
4340 larger relative overhead of up to 15% in the worst case.
4343 png_set_compression_mem_level(png_ptr, level);
4345 The other functions are for configuring zlib. They are not recommended
4346 for normal use and may result in writing an invalid PNG file. See
4347 zlib.h for more information on what these mean.
4350 png_set_compression_strategy(png_ptr,
4353 png_set_compression_window_bits(png_ptr,
4356 png_set_compression_method(png_ptr, method);
4358 This controls the size of the IDAT chunks (default 8192):
4360 png_set_compression_buffer_size(png_ptr, size);
4362 As of libpng version 1.5.4, additional APIs became
4363 available to set these separately for non-IDAT
4364 compressed chunks such as zTXt, iTXt, and iCCP:
4367 #if PNG_LIBPNG_VER >= 10504
4368 png_set_text_compression_level(png_ptr, level);
4370 png_set_text_compression_mem_level(png_ptr, level);
4372 png_set_text_compression_strategy(png_ptr,
4375 png_set_text_compression_window_bits(png_ptr,
4378 png_set_text_compression_method(png_ptr, method);
4381 Controlling row filtering
4383 If you want to control whether libpng uses filtering or not, which
4384 filters are used, and how it goes about picking row filters, you
4385 can call one of these functions. The selection and configuration
4386 of row filters can have a significant impact on the size and
4387 encoding speed and a somewhat lesser impact on the decoding speed
4388 of an image. Filtering is enabled by default for RGB and grayscale
4389 images (with and without alpha), but not for paletted images nor
4390 for any images with bit depths less than 8 bits/pixel.
4392 The 'method' parameter sets the main filtering method, which is
4393 currently only '0' in the PNG 1.2 specification. The 'filters'
4394 parameter sets which filter(s), if any, should be used for each
4395 scanline. Possible values are PNG_ALL_FILTERS, PNG_NO_FILTERS,
4396 or PNG_FAST_FILTERS to turn filtering on and off, or to turn on
4397 just the fast-decoding subset of filters, respectively.
4399 Individual filter types are PNG_FILTER_NONE, PNG_FILTER_SUB,
4400 PNG_FILTER_UP, PNG_FILTER_AVG, PNG_FILTER_PAETH, which can be bitwise
4401 ORed together with '|' to specify one or more filters to use.
4402 These filters are described in more detail in the PNG specification.
4403 If you intend to change the filter type during the course of writing
4404 the image, you should start with flags set for all of the filters
4405 you intend to use so that libpng can initialize its internal
4406 structures appropriately for all of the filter types. (Note that this
4407 means the first row must always be adaptively filtered, because libpng
4408 currently does not allocate the filter buffers until png_write_row()
4409 is called for the first time.)
4411 filters = PNG_NO_FILTERS;
4412 filters = PNG_ALL_FILTERS;
4413 filters = PNG_FAST_FILTERS;
4417 filters = PNG_FILTER_NONE | PNG_FILTER_SUB |
4418 PNG_FILTER_UP | PNG_FILTER_AVG |
4421 png_set_filter(png_ptr, PNG_FILTER_TYPE_BASE,
4424 The second parameter can also be
4425 PNG_INTRAPIXEL_DIFFERENCING if you are
4426 writing a PNG to be embedded in a MNG
4427 datastream. This parameter must be the
4428 same as the value of filter_method used
4431 Requesting debug printout
4433 The macro definition PNG_DEBUG can be used to request debugging
4434 printout. Set it to an integer value in the range 0 to 3. Higher
4435 numbers result in increasing amounts of debugging information. The
4436 information is printed to the "stderr" file, unless another file
4437 name is specified in the PNG_DEBUG_FILE macro definition.
4439 When PNG_DEBUG > 0, the following functions (macros) become available:
4441 png_debug(level, message)
4442 png_debug1(level, message, p1)
4443 png_debug2(level, message, p1, p2)
4445 in which "level" is compared to PNG_DEBUG to decide whether to print
4446 the message, "message" is the formatted string to be printed,
4447 and p1 and p2 are parameters that are to be embedded in the string
4448 according to printf-style formatting directives. For example,
4450 png_debug1(2, "foo=%d", foo);
4455 fprintf(PNG_DEBUG_FILE, "foo=%d\n", foo);
4457 When PNG_DEBUG is defined but is zero, the macros aren't defined, but you
4458 can still use PNG_DEBUG to control your own debugging:
4464 When PNG_DEBUG = 1, the macros are defined, but only png_debug statements
4465 having level = 0 will be printed. There aren't any such statements in
4466 this version of libpng, but if you insert some they will be printed.
4470 The MNG specification (available at http://www.libpng.org/pub/mng) allows
4471 certain extensions to PNG for PNG images that are embedded in MNG datastreams.
4472 Libpng can support some of these extensions. To enable them, use the
4473 png_permit_mng_features() function:
4475 feature_set = png_permit_mng_features(png_ptr, mask)
4477 mask is a png_uint_32 containing the bitwise OR of the
4478 features you want to enable. These include
4479 PNG_FLAG_MNG_EMPTY_PLTE
4480 PNG_FLAG_MNG_FILTER_64
4481 PNG_ALL_MNG_FEATURES
4483 feature_set is a png_uint_32 that is the bitwise AND of
4484 your mask with the set of MNG features that is
4485 supported by the version of libpng that you are using.
4487 It is an error to use this function when reading or writing a standalone
4488 PNG file with the PNG 8-byte signature. The PNG datastream must be wrapped
4489 in a MNG datastream. As a minimum, it must have the MNG 8-byte signature
4490 and the MHDR and MEND chunks. Libpng does not provide support for these
4491 or any other MNG chunks; your application must provide its own support for
4492 them. You may wish to consider using libmng (available at
4493 http://www.libmng.com) instead.
4495 VIII. Changes to Libpng from version 0.88
4497 It should be noted that versions of libpng later than 0.96 are not
4498 distributed by the original libpng author, Guy Schalnat, nor by
4499 Andreas Dilger, who had taken over from Guy during 1996 and 1997, and
4500 distributed versions 0.89 through 0.96, but rather by another member
4501 of the original PNG Group, Glenn Randers-Pehrson. Guy and Andreas are
4502 still alive and well, but they have moved on to other things.
4504 The old libpng functions png_read_init(), png_write_init(),
4505 png_info_init(), png_read_destroy(), and png_write_destroy() have been
4506 moved to PNG_INTERNAL in version 0.95 to discourage their use. These
4507 functions will be removed from libpng version 1.4.0.
4509 The preferred method of creating and initializing the libpng structures is
4510 via the png_create_read_struct(), png_create_write_struct(), and
4511 png_create_info_struct() because they isolate the size of the structures
4512 from the application, allow version error checking, and also allow the
4513 use of custom error handling routines during the initialization, which
4514 the old functions do not. The functions png_read_destroy() and
4515 png_write_destroy() do not actually free the memory that libpng
4516 allocated for these structs, but just reset the data structures, so they
4517 can be used instead of png_destroy_read_struct() and
4518 png_destroy_write_struct() if you feel there is too much system overhead
4519 allocating and freeing the png_struct for each image read.
4521 Setting the error callbacks via png_set_message_fn() before
4522 png_read_init() as was suggested in libpng-0.88 is no longer supported
4523 because this caused applications that do not use custom error functions
4524 to fail if the png_ptr was not initialized to zero. It is still possible
4525 to set the error callbacks AFTER png_read_init(), or to change them with
4526 png_set_error_fn(), which is essentially the same function, but with a new
4527 name to force compilation errors with applications that try to use the old
4530 Support for the sCAL, iCCP, iTXt, and sPLT chunks was added at libpng-1.0.6;
4531 however, iTXt support was not enabled by default.
4533 Starting with version 1.0.7, you can find out which version of the library
4534 you are using at run-time:
4536 png_uint_32 libpng_vn = png_access_version_number();
4538 The number libpng_vn is constructed from the major version, minor
4539 version with leading zero, and release number with leading zero,
4540 (e.g., libpng_vn for version 1.0.7 is 10007).
4542 Note that this function does not take a png_ptr, so you can call it
4543 before you've created one.
4545 You can also check which version of png.h you used when compiling your
4548 png_uint_32 application_vn = PNG_LIBPNG_VER;
4550 IX. Changes to Libpng from version 1.0.x to 1.2.x
4552 Support for user memory management was enabled by default. To
4553 accomplish this, the functions png_create_read_struct_2(),
4554 png_create_write_struct_2(), png_set_mem_fn(), png_get_mem_ptr(),
4555 png_malloc_default(), and png_free_default() were added.
4557 Support for the iTXt chunk has been enabled by default as of
4560 Support for certain MNG features was enabled.
4562 Support for numbered error messages was added. However, we never got
4563 around to actually numbering the error messages. The function
4564 png_set_strip_error_numbers() was added (Note: the prototype for this
4565 function was inadvertently removed from png.h in PNG_NO_ASSEMBLER_CODE
4566 builds of libpng-1.2.15. It was restored in libpng-1.2.36).
4568 The png_malloc_warn() function was added at libpng-1.2.3. This issues
4569 a png_warning and returns NULL instead of aborting when it fails to
4570 acquire the requested memory allocation.
4572 Support for setting user limits on image width and height was enabled
4573 by default. The functions png_set_user_limits(), png_get_user_width_max(),
4574 and png_get_user_height_max() were added at libpng-1.2.6.
4576 The png_set_add_alpha() function was added at libpng-1.2.7.
4578 The function png_set_expand_gray_1_2_4_to_8() was added at libpng-1.2.9.
4579 Unlike png_set_gray_1_2_4_to_8(), the new function does not expand the
4580 tRNS chunk to alpha. The png_set_gray_1_2_4_to_8() function is
4583 A number of macro definitions in support of runtime selection of
4584 assembler code features (especially Intel MMX code support) were
4585 added at libpng-1.2.0:
4587 PNG_ASM_FLAG_MMX_SUPPORT_COMPILED
4588 PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU
4589 PNG_ASM_FLAG_MMX_READ_COMBINE_ROW
4590 PNG_ASM_FLAG_MMX_READ_INTERLACE
4591 PNG_ASM_FLAG_MMX_READ_FILTER_SUB
4592 PNG_ASM_FLAG_MMX_READ_FILTER_UP
4593 PNG_ASM_FLAG_MMX_READ_FILTER_AVG
4594 PNG_ASM_FLAG_MMX_READ_FILTER_PAETH
4595 PNG_ASM_FLAGS_INITIALIZED
4601 We added the following functions in support of runtime
4602 selection of assembler code features:
4604 png_get_mmx_flagmask()
4605 png_set_mmx_thresholds()
4607 png_get_mmx_bitdepth_threshold()
4608 png_get_mmx_rowbytes_threshold()
4611 We replaced all of these functions with simple stubs in libpng-1.2.20,
4612 when the Intel assembler code was removed due to a licensing issue.
4614 These macros are deprecated:
4616 PNG_READ_TRANSFORMS_NOT_SUPPORTED
4617 PNG_PROGRESSIVE_READ_NOT_SUPPORTED
4618 PNG_NO_SEQUENTIAL_READ_SUPPORTED
4619 PNG_WRITE_TRANSFORMS_NOT_SUPPORTED
4620 PNG_READ_ANCILLARY_CHUNKS_NOT_SUPPORTED
4621 PNG_WRITE_ANCILLARY_CHUNKS_NOT_SUPPORTED
4623 They have been replaced, respectively, by:
4625 PNG_NO_READ_TRANSFORMS
4626 PNG_NO_PROGRESSIVE_READ
4627 PNG_NO_SEQUENTIAL_READ
4628 PNG_NO_WRITE_TRANSFORMS
4629 PNG_NO_READ_ANCILLARY_CHUNKS
4630 PNG_NO_WRITE_ANCILLARY_CHUNKS
4632 PNG_MAX_UINT was replaced with PNG_UINT_31_MAX. It has been
4633 deprecated since libpng-1.0.16 and libpng-1.2.6.
4636 png_check_sig(sig, num)
4638 !png_sig_cmp(sig, 0, num)
4639 It has been deprecated since libpng-0.90.
4642 png_set_gray_1_2_4_to_8()
4643 which also expands tRNS to alpha was replaced with
4644 png_set_expand_gray_1_2_4_to_8()
4645 which does not. It has been deprecated since libpng-1.0.18 and 1.2.9.
4647 X. Changes to Libpng from version 1.0.x/1.2.x to 1.4.x
4649 Private libpng prototypes and macro definitions were moved from
4650 png.h and pngconf.h into a new pngpriv.h header file.
4652 Functions png_set_benign_errors(), png_benign_error(), and
4653 png_chunk_benign_error() were added.
4655 Support for setting the maximum amount of memory that the application
4656 will allocate for reading chunks was added, as a security measure.
4657 The functions png_set_chunk_cache_max() and png_get_chunk_cache_max()
4658 were added to the library.
4660 We implemented support for I/O states by adding png_ptr member io_state
4661 and functions png_get_io_chunk_name() and png_get_io_state() in pngget.c
4663 We added PNG_TRANSFORM_GRAY_TO_RGB to the available high-level
4666 Checking for and reporting of errors in the IHDR chunk is more thorough.
4668 Support for global arrays was removed, to improve thread safety.
4670 Some obsolete/deprecated macros and functions have been removed.
4672 Typecasted NULL definitions such as
4673 #define png_voidp_NULL (png_voidp)NULL
4674 were eliminated. If you used these in your application, just use
4677 The png_struct and info_struct members "trans" and "trans_values" were
4678 changed to "trans_alpha" and "trans_color", respectively.
4680 The obsolete, unused pnggccrd.c and pngvcrd.c files and related makefiles
4683 The PNG_1_0_X and PNG_1_2_X macros were eliminated.
4685 The PNG_LEGACY_SUPPORTED macro was eliminated.
4687 Many WIN32_WCE #ifdefs were removed.
4689 The functions png_read_init(info_ptr), png_write_init(info_ptr),
4690 png_info_init(info_ptr), png_read_destroy(), and png_write_destroy()
4691 have been removed. They have been deprecated since libpng-0.95.
4693 The png_permit_empty_plte() was removed. It has been deprecated
4694 since libpng-1.0.9. Use png_permit_mng_features() instead.
4696 We removed the obsolete stub functions png_get_mmx_flagmask(),
4697 png_set_mmx_thresholds(), png_get_asm_flags(),
4698 png_get_mmx_bitdepth_threshold(), png_get_mmx_rowbytes_threshold(),
4699 png_set_asm_flags(), and png_mmx_supported()
4701 We removed the obsolete png_check_sig(), png_memcpy_check(), and
4702 png_memset_check() functions. Instead use !png_sig_cmp(), memcpy(),
4703 and memset(), respectively.
4705 The function png_set_gray_1_2_4_to_8() was removed. It has been
4706 deprecated since libpng-1.0.18 and 1.2.9, when it was replaced with
4707 png_set_expand_gray_1_2_4_to_8() because the former function also
4708 expanded any tRNS chunk to an alpha channel.
4710 Macros for png_get_uint_16, png_get_uint_32, and png_get_int_32
4711 were added and are used by default instead of the corresponding
4712 functions. Unfortunately,
4713 from libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
4714 function) incorrectly returned a value of type png_uint_32.
4716 We changed the prototype for png_malloc() from
4717 png_malloc(png_structp png_ptr, png_uint_32 size)
4719 png_malloc(png_structp png_ptr, png_alloc_size_t size)
4721 This also applies to the prototype for the user replacement malloc_fn().
4723 The png_calloc() function was added and is used in place of
4724 of "png_malloc(); memset();" except in the case in png_read_png()
4725 where the array consists of pointers; in this case a "for" loop is used
4726 after the png_malloc() to set the pointers to NULL, to give robust.
4727 behavior in case the application runs out of memory part-way through
4730 We changed the prototypes of png_get_compression_buffer_size() and
4731 png_set_compression_buffer_size() to work with png_size_t instead of
4734 Support for numbered error messages was removed by default, since we
4735 never got around to actually numbering the error messages. The function
4736 png_set_strip_error_numbers() was removed from the library by default.
4738 The png_zalloc() and png_zfree() functions are no longer exported.
4739 The png_zalloc() function no longer zeroes out the memory that it
4740 allocates. Applications that called png_zalloc(png_ptr, number, size)
4741 can call png_calloc(png_ptr, number*size) instead, and can call
4742 png_free() instead of png_zfree().
4744 Support for dithering was disabled by default in libpng-1.4.0, because
4745 it has not been well tested and doesn't actually "dither".
4747 removed, however, and could be enabled by building libpng with
4748 PNG_READ_DITHER_SUPPORTED defined. In libpng-1.4.2, this support
4749 was re-enabled, but the function was renamed png_set_quantize() to
4750 reflect more accurately what it actually does. At the same time,
4751 the PNG_DITHER_[RED,GREEN_BLUE]_BITS macros were also renamed to
4752 PNG_QUANTIZE_[RED,GREEN,BLUE]_BITS, and PNG_READ_DITHER_SUPPORTED
4753 was renamed to PNG_READ_QUANTIZE_SUPPORTED.
4755 We removed the trailing '.' from the warning and error messages.
4757 XI. Changes to Libpng from version 1.4.x to 1.5.x
4759 From libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
4760 function) incorrectly returned a value of type png_uint_32.
4761 The incorrect macro was removed from libpng-1.4.5.
4763 Checking for invalid palette index on write was added at libpng
4764 1.5.10. If a pixel contains an invalid (out-of-range) index libpng issues
4765 a benign error. This is enabled by default because this condition is an
4766 error according to the PNG specification, Clause 11.3.2, but the error can
4767 be ignored in each png_ptr with
4769 png_set_check_for_invalid_index(png_ptr, allowed);
4772 0: disable benign error (accept the
4773 invalid data without warning).
4774 1: enable benign error (treat the
4775 invalid data as an error or a
4778 If the error is ignored, or if png_benign_error() treats it as a warning,
4779 any invalid pixels are decoded as opaque black by the decoder and written
4780 as-is by the encoder.
4782 Retrieving the maximum palette index found was added at libpng-1.5.15.
4783 This statement must appear after png_read_png() or png_read_image() while
4784 reading, and after png_write_png() or png_write_image() while writing.
4786 int max_palette = png_get_palette_max(png_ptr, info_ptr);
4788 This will return the maximum palette index found in the image, or "-1" if
4789 the palette was not checked, or "0" if no palette was found. Note that this
4790 does not account for any palette index used by ancillary chunks such as the
4791 bKGD chunk; you must check those separately to determine the maximum
4792 palette index actually used.
4794 There are no substantial API changes between the non-deprecated parts of
4795 the 1.4.5 API and the 1.5.0 API; however, the ability to directly access
4796 members of the main libpng control structures, png_struct and png_info,
4797 deprecated in earlier versions of libpng, has been completely removed from
4798 libpng 1.5, and new private "pngstruct.h", "pnginfo.h", and "pngdebug.h"
4799 header files were created.
4801 We no longer include zlib.h in png.h. The include statement has been moved
4802 to pngstruct.h, where it is not accessible by applications. Applications that
4803 need access to information in zlib.h will need to add the '#include "zlib.h"'
4804 directive. It does not matter whether this is placed prior to or after
4805 the '"#include png.h"' directive.
4807 The png_sprintf(), png_strcpy(), and png_strncpy() macros are no longer used
4810 We moved the png_strlen(), png_memcpy(), png_memset(), and png_memcmp()
4811 macros into a private header file (pngpriv.h) that is not accessible to
4814 In png_get_iCCP, the type of "profile" was changed from png_charpp
4815 to png_bytepp, and in png_set_iCCP, from png_charp to png_const_bytep.
4817 There are changes of form in png.h, including new and changed macros to
4818 declare parts of the API. Some API functions with arguments that are
4819 pointers to data not modified within the function have been corrected to
4820 declare these arguments with PNG_CONST.
4822 Much of the internal use of C macros to control the library build has also
4823 changed and some of this is visible in the exported header files, in
4824 particular the use of macros to control data and API elements visible
4825 during application compilation may require significant revision to
4826 application code. (It is extremely rare for an application to do this.)
4828 Any program that compiled against libpng 1.4 and did not use deprecated
4829 features or access internal library structures should compile and work
4830 against libpng 1.5, except for the change in the prototype for
4831 png_get_iCCP() and png_set_iCCP() API functions mentioned above.
4833 libpng 1.5.0 adds PNG_ PASS macros to help in the reading and writing of
4834 interlaced images. The macros return the number of rows and columns in
4835 each pass and information that can be used to de-interlace and (if
4836 absolutely necessary) interlace an image.
4838 libpng 1.5.0 adds an API png_longjmp(png_ptr, value). This API calls
4839 the application-provided png_longjmp_ptr on the internal, but application
4840 initialized, longjmp buffer. It is provided as a convenience to avoid
4841 the need to use the png_jmpbuf macro, which had the unnecessary side
4842 effect of resetting the internal png_longjmp_ptr value.
4844 libpng 1.5.0 includes a complete fixed point API. By default this is
4845 present along with the corresponding floating point API. In general the
4846 fixed point API is faster and smaller than the floating point one because
4847 the PNG file format used fixed point, not floating point. This applies
4848 even if the library uses floating point in internal calculations. A new
4849 macro, PNG_FLOATING_ARITHMETIC_SUPPORTED, reveals whether the library
4850 uses floating point arithmetic (the default) or fixed point arithmetic
4851 internally for performance critical calculations such as gamma correction.
4852 In some cases, the gamma calculations may produce slightly different
4853 results. This has changed the results in png_rgb_to_gray and in alpha
4854 composition (png_set_background for example). This applies even if the
4855 original image was already linear (gamma == 1.0) and, therefore, it is
4856 not necessary to linearize the image. This is because libpng has *not*
4857 been changed to optimize that case correctly, yet.
4859 Fixed point support for the sCAL chunk comes with an important caveat;
4860 the sCAL specification uses a decimal encoding of floating point values
4861 and the accuracy of PNG fixed point values is insufficient for
4862 representation of these values. Consequently a "string" API
4863 (png_get_sCAL_s and png_set_sCAL_s) is the only reliable way of reading
4864 arbitrary sCAL chunks in the absence of either the floating point API or
4865 internal floating point calculations. Starting with libpng-1.5.0, both
4866 of these functions are present when PNG_sCAL_SUPPORTED is defined. Prior
4867 to libpng-1.5.0, their presence also depended upon PNG_FIXED_POINT_SUPPORTED
4868 being defined and PNG_FLOATING_POINT_SUPPORTED not being defined.
4870 Applications no longer need to include the optional distribution header
4871 file pngusr.h or define the corresponding macros during application
4872 build in order to see the correct variant of the libpng API. From 1.5.0
4873 application code can check for the corresponding _SUPPORTED macro:
4875 #ifdef PNG_INCH_CONVERSIONS_SUPPORTED
4876 /* code that uses the inch conversion APIs. */
4879 This macro will only be defined if the inch conversion functions have been
4880 compiled into libpng. The full set of macros, and whether or not support
4881 has been compiled in, are available in the header file pnglibconf.h.
4882 This header file is specific to the libpng build. Notice that prior to
4883 1.5.0 the _SUPPORTED macros would always have the default definition unless
4884 reset by pngusr.h or by explicit settings on the compiler command line.
4885 These settings may produce compiler warnings or errors in 1.5.0 because
4886 of macro redefinition.
4888 Applications can now choose whether to use these macros or to call the
4889 corresponding function by defining PNG_USE_READ_MACROS or
4890 PNG_NO_USE_READ_MACROS before including png.h. Notice that this is
4891 only supported from 1.5.0; defining PNG_NO_USE_READ_MACROS prior to 1.5.0
4892 will lead to a link failure.
4894 Prior to libpng-1.5.4, the zlib compressor used the same set of parameters
4895 when compressing the IDAT data and textual data such as zTXt and iCCP.
4896 In libpng-1.5.4 we reinitialized the zlib stream for each type of data.
4897 We added five png_set_text_*() functions for setting the parameters to
4898 use with textual data.
4900 Prior to libpng-1.5.4, the PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
4901 option was off by default, and slightly inaccurate scaling occurred.
4902 This option can no longer be turned off, and the choice of accurate
4903 or inaccurate 16-to-8 scaling is by using the new png_set_scale_16_to_8()
4904 API for accurate scaling or the old png_set_strip_16_to_8() API for simple
4905 chopping. In libpng-1.5.4, the PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
4906 macro became PNG_READ_SCALE_16_TO_8_SUPPORTED, and the PNG_READ_16_TO_8
4907 macro became PNG_READ_STRIP_16_TO_8_SUPPORTED, to enable the two
4908 png_set_*_16_to_8() functions separately.
4910 Prior to libpng-1.5.4, the png_set_user_limits() function could only be
4911 used to reduce the width and height limits from the value of
4912 PNG_USER_WIDTH_MAX and PNG_USER_HEIGHT_MAX, although this document said
4913 that it could be used to override them. Now this function will reduce or
4914 increase the limits.
4916 Starting in libpng-1.5.10, the user limits can be set en masse with the
4917 configuration option PNG_SAFE_LIMITS_SUPPORTED. If this option is enabled,
4918 a set of "safe" limits is applied in pngpriv.h. These can be overridden by
4919 application calls to png_set_user_limits(), png_set_user_chunk_cache_max(),
4920 and/or png_set_user_malloc_max() that increase or decrease the limits. Also,
4921 in libpng-1.5.10 the default width and height limits were increased
4922 from 1,000,000 to 0x7fffffff (i.e., made unlimited). Therefore, the
4925 png_user_width_max 0x7fffffff 1,000,000
4926 png_user_height_max 0x7fffffff 1,000,000
4927 png_user_chunk_cache_max 0 (unlimited) 128
4928 png_user_chunk_malloc_max 0 (unlimited) 8,000,000
4930 The png_set_option() function (and the "options" member of the png struct) was
4931 added to libpng-1.5.15, with option PNG_ARM_NEON.
4933 The library now supports a complete fixed point implementation and can
4934 thus be used on systems that have no floating point support or very
4935 limited or slow support. Previously gamma correction, an essential part
4936 of complete PNG support, required reasonably fast floating point.
4938 As part of this the choice of internal implementation has been made
4939 independent of the choice of fixed versus floating point APIs and all the
4940 missing fixed point APIs have been implemented.
4942 The exact mechanism used to control attributes of API functions has
4943 changed, as described in the INSTALL file.
4945 A new test program, pngvalid, is provided in addition to pngtest.
4946 pngvalid validates the arithmetic accuracy of the gamma correction
4947 calculations and includes a number of validations of the file format.
4948 A subset of the full range of tests is run when "make check" is done
4949 (in the 'configure' build.) pngvalid also allows total allocated memory
4950 usage to be evaluated and performs additional memory overwrite validation.
4952 Many changes to individual feature macros have been made. The following
4953 are the changes most likely to be noticed by library builders who
4956 1) All feature macros now have consistent naming:
4958 #define PNG_NO_feature turns the feature off
4959 #define PNG_feature_SUPPORTED turns the feature on
4961 pnglibconf.h contains one line for each feature macro which is either:
4963 #define PNG_feature_SUPPORTED
4965 if the feature is supported or:
4967 /*#undef PNG_feature_SUPPORTED*/
4969 if it is not. Library code consistently checks for the 'SUPPORTED' macro.
4970 It does not, and libpng applications should not, check for the 'NO' macro
4971 which will not normally be defined even if the feature is not supported.
4972 The 'NO' macros are only used internally for setting or not setting the
4973 corresponding 'SUPPORTED' macros.
4975 Compatibility with the old names is provided as follows:
4977 PNG_INCH_CONVERSIONS turns on PNG_INCH_CONVERSIONS_SUPPORTED
4979 And the following definitions disable the corresponding feature:
4981 PNG_SETJMP_NOT_SUPPORTED disables SETJMP
4982 PNG_READ_TRANSFORMS_NOT_SUPPORTED disables READ_TRANSFORMS
4983 PNG_NO_READ_COMPOSITED_NODIV disables READ_COMPOSITE_NODIV
4984 PNG_WRITE_TRANSFORMS_NOT_SUPPORTED disables WRITE_TRANSFORMS
4985 PNG_READ_ANCILLARY_CHUNKS_NOT_SUPPORTED disables READ_ANCILLARY_CHUNKS
4986 PNG_WRITE_ANCILLARY_CHUNKS_NOT_SUPPORTED disables WRITE_ANCILLARY_CHUNKS
4988 Library builders should remove use of the above, inconsistent, names.
4990 2) Warning and error message formatting was previously conditional on
4991 the STDIO feature. The library has been changed to use the
4992 CONSOLE_IO feature instead. This means that if CONSOLE_IO is disabled
4993 the library no longer uses the printf(3) functions, even though the
4994 default read/write implementations use (FILE) style stdio.h functions.
4996 3) Three feature macros now control the fixed/floating point decisions:
4998 PNG_FLOATING_POINT_SUPPORTED enables the floating point APIs
5000 PNG_FIXED_POINT_SUPPORTED enables the fixed point APIs; however, in
5001 practice these are normally required internally anyway (because the PNG
5002 file format is fixed point), therefore in most cases PNG_NO_FIXED_POINT
5003 merely stops the function from being exported.
5005 PNG_FLOATING_ARITHMETIC_SUPPORTED chooses between the internal floating
5006 point implementation or the fixed point one. Typically the fixed point
5007 implementation is larger and slower than the floating point implementation
5008 on a system that supports floating point; however, it may be faster on a
5009 system which lacks floating point hardware and therefore uses a software
5012 4) Added PNG_{READ,WRITE}_INT_FUNCTIONS_SUPPORTED. This allows the
5013 functions to read and write ints to be disabled independently of
5014 PNG_USE_READ_MACROS, which allows libpng to be built with the functions
5015 even though the default is to use the macros - this allows applications
5016 to choose at app buildtime whether or not to use macros (previously
5017 impossible because the functions weren't in the default build.)
5019 XII. Changes to Libpng from version 1.5.x to 1.6.x
5021 A "simplified API" has been added (see documentation in png.h and a simple
5022 example in contrib/examples/pngtopng.c). The new publicly visible API
5023 includes the following:
5032 png_image_begin_read_from_file()
5033 png_image_begin_read_from_stdio()
5034 png_image_begin_read_from_memory()
5035 png_image_finish_read()
5038 png_image_write_to_file()
5039 png_image_write_to_memory()
5040 png_image_write_to_stdio()
5042 Starting with libpng-1.6.0, you can configure libpng to prefix all exported
5043 symbols, using the PNG_PREFIX macro.
5045 We no longer include string.h in png.h. The include statement has been moved
5046 to pngpriv.h, where it is not accessible by applications. Applications that
5047 need access to information in string.h must add an '#include <string.h>'
5048 directive. It does not matter whether this is placed prior to or after
5049 the '#include "png.h"' directive.
5051 The following API are now DEPRECATED:
5053 png_convert_to_rfc1123() which has been replaced
5054 with png_convert_to_rfc1123_buffer()
5055 png_malloc_default()
5059 The following have been removed:
5060 png_get_io_chunk_name(), which has been replaced
5061 with png_get_io_chunk_type(). The new
5062 function returns a 32-bit integer instead of
5064 The png_sizeof(), png_strlen(), png_memcpy(), png_memcmp(), and
5065 png_memset() macros are no longer used in the libpng sources and
5066 have been removed. These had already been made invisible to applications
5067 (i.e., defined in the private pngpriv.h header file) since libpng-1.5.0.
5069 The signatures of many exported functions were changed, such that
5070 png_structp became png_structrp or png_const_structrp
5071 png_infop became png_inforp or png_const_inforp
5072 where "rp" indicates a "restricted pointer".
5074 Dropped support for 16-bit platforms. The support for FAR/far types has
5075 been eliminated and the definition of png_alloc_size_t is now controlled
5076 by a flag so that 'small size_t' systems can select it if necessary.
5078 Error detection in some chunks has improved; in particular the iCCP chunk
5079 reader now does pretty complete validation of the basic format. Some bad
5080 profiles that were previously accepted are now accepted with a warning or
5081 rejected, depending upon the png_set_benign_errors() setting, in particular
5082 the very old broken Microsoft/HP 3144-byte sRGB profile. Starting with
5083 libpng-1.6.11, recognizing and checking sRGB profiles can be avoided by
5086 #if defined(PNG_SKIP_sRGB_CHECK_PROFILE) && \
5087 defined(PNG_SET_OPTION_SUPPORTED)
5088 png_set_option(png_ptr, PNG_SKIP_sRGB_CHECK_PROFILE,
5092 It's not a good idea to do this if you are using the "simplified API",
5093 which needs to be able to recognize sRGB profiles conveyed via the iCCP
5096 The PNG spec requirement that only grayscale profiles may appear in images
5097 with color type 0 or 4 and that even if the image only contains gray pixels,
5098 only RGB profiles may appear in images with color type 2, 3, or 6, is now
5099 enforced. The sRGB chunk is allowed to appear in images with any color type
5100 and is interpreted by libpng to convey a one-tracer-curve gray profile or a
5101 three-tracer-curve RGB profile as appropriate.
5103 Libpng 1.5.x erroneously used /MD for Debug DLL builds; if you used the debug
5104 builds in your app and you changed your app to use /MD you will need to
5105 change it back to /MDd for libpng 1.6.x.
5107 Prior to libpng-1.6.0 a warning would be issued if the iTXt chunk contained
5108 an empty language field or an empty translated keyword. Both of these
5109 are allowed by the PNG specification, so these warnings are no longer issued.
5111 The library now issues an error if the application attempts to set a
5112 transform after it calls png_read_update_info() or if it attempts to call
5113 both png_read_update_info() and png_start_read_image() or to call either
5114 of them more than once.
5116 The default condition for benign_errors is now to treat benign errors as
5117 warnings while reading and as errors while writing.
5119 The library now issues a warning if both background processing and RGB to
5120 gray are used when gamma correction happens. As with previous versions of
5121 the library the results are numerically very incorrect in this case.
5123 There are some minor arithmetic changes in some transforms such as
5124 png_set_background(), that might be detected by certain regression tests.
5126 Unknown chunk handling has been improved internally, without any API change.
5127 This adds more correct option control of the unknown handling, corrects
5128 a pre-existing bug where the per-chunk 'keep' setting is ignored, and makes
5129 it possible to skip IDAT chunks in the sequential reader.
5131 The machine-generated configure files are no longer included in branches
5132 libpng16 and later of the GIT repository. They continue to be included
5133 in the tarball releases, however.
5135 Libpng-1.6.0 through 1.6.2 used the CMF bytes at the beginning of the IDAT
5136 stream to set the size of the sliding window for reading instead of using the
5137 default 32-kbyte sliding window size. It was discovered that there are
5138 hundreds of PNG files in the wild that have incorrect CMF bytes that caused
5139 zlib to issue the "invalid distance too far back" error and reject the file.
5140 Libpng-1.6.3 and later calculate their own safe CMF from the image dimensions,
5141 provide a way to revert to the libpng-1.5.x behavior (ignoring the CMF bytes
5142 and using a 32-kbyte sliding window), by using
5144 png_set_option(png_ptr, PNG_MAXIMUM_INFLATE_WINDOW,
5147 and provide a tool (contrib/tools/pngfix) for rewriting a PNG file while
5148 optimizing the CMF bytes in its IDAT chunk correctly.
5150 Libpng-1.6.0 and libpng-1.6.1 wrote uncompressed iTXt chunks with the wrong
5151 length, which resulted in PNG files that cannot be read beyond the bad iTXt
5152 chunk. This error was fixed in libpng-1.6.3, and a tool (called
5153 contrib/tools/png-fix-itxt) has been added to the libpng distribution.
5155 Starting with libpng-1.6.17, the PNG_SAFE_LIMITS macro was eliminated
5156 and safe limits are used by default (users who need larger limits
5157 can still override them at compile time or run time, as described above).
5161 png_user_width_max 1,000,000 2,147,483,647
5162 png_user_height_max 1,000,000 2,147,483,647
5163 png_user_chunk_cache_max 128 unlimited
5164 png_user_chunk_malloc_max 8,000,000 unlimited
5166 Starting with libpng-1.6.18, a PNG_RELEASE_BUILD macro was added, which allows
5167 library builders to control compilation for an installed system (a release build).
5168 It can be set for testing debug or beta builds to ensure that they will compile
5169 when the build type is switched to RC or STABLE. In essence this overrides the
5170 PNG_LIBPNG_BUILD_BASE_TYPE definition which is not directly user controllable.
5172 Starting with libpng-1.6.19, attempting to set an over-length PLTE chunk
5173 is an error. Previously this requirement of the PNG specification was not
5174 enforced, and the palette was always limited to 256 entries. An over-length
5175 PLTE chunk found in an input PNG is silently truncated.
5177 XIII. Detecting libpng
5179 The png_get_io_ptr() function has been present since libpng-0.88, has never
5180 changed, and is unaffected by conditional compilation macros. It is the
5181 best choice for use in configure scripts for detecting the presence of any
5182 libpng version since 0.88. In an autoconf "configure.in" you could use
5184 AC_CHECK_LIB(png, png_get_io_ptr, ...
5186 XV. Source code repository
5188 Since about February 2009, version 1.2.34, libpng has been under "git" source
5189 control. The git repository was built from old libpng-x.y.z.tar.gz files
5190 going back to version 0.70. You can access the git repository (read only)
5193 git://git.code.sf.net/p/libpng/code
5195 or you can browse it with a web browser by selecting the "code" button at
5197 https://sourceforge.net/projects/libpng
5199 Patches can be sent to glennrp at users.sourceforge.net or to
5200 png-mng-implement at lists.sourceforge.net or you can upload them to
5201 the libpng bug tracker at
5203 http://libpng.sourceforge.net
5205 We also accept patches built from the tar or zip distributions, and
5206 simple verbal discriptions of bug fixes, reported either to the
5207 SourceForge bug tracker, to the png-mng-implement at lists.sf.net
5208 mailing list, or directly to glennrp.
5212 Our coding style is similar to the "Allman" style
5213 (See http://en.wikipedia.org/wiki/Indent_style#Allman_style), with curly
5214 braces on separate lines:
5221 else if (another condition)
5226 The braces can be omitted from simple one-line actions:
5231 We use 3-space indentation, except for continued statements which
5232 are usually indented the same as the first line of the statement
5233 plus four more spaces.
5235 For macro definitions we use 2-space indentation, always leaving the "#"
5236 in the first column.
5238 #ifndef PNG_NO_FEATURE
5239 # ifndef PNG_FEATURE_SUPPORTED
5240 # define PNG_FEATURE_SUPPORTED
5244 Comments appear with the leading "/*" at the same indentation as
5245 the statement that follows the comment:
5247 /* Single-line comment */
5250 /* This is a multiple-line
5255 Very short comments can be placed after the end of the statement
5256 to which they pertain:
5258 statement; /* comment */
5260 We don't use C++ style ("//") comments. We have, however,
5261 used them in the past in some now-abandoned MMX assembler
5264 Functions and their curly braces are not indented, and
5265 exported functions are marked with PNGAPI:
5267 /* This is a public function that is visible to
5268 * application programmers. It does thus-and-so.
5271 png_exported_function(png_ptr, png_info, foo)
5276 The return type and decorations are placed on a separate line
5277 ahead of the function name, as illustrated above.
5279 The prototypes for all exported functions appear in png.h,
5280 above the comment that says
5282 /* Maintainer: Put new public prototypes here ... */
5284 We mark all non-exported functions with "/* PRIVATE */"":
5287 png_non_exported_function(png_ptr, png_info, foo)
5292 The prototypes for non-exported functions (except for those in
5293 pngtest) appear in pngpriv.h above the comment that says
5295 /* Maintainer: Put new private prototypes here ^ */
5297 To avoid polluting the global namespace, the names of all exported
5298 functions and variables begin with "png_", and all publicly visible C
5299 preprocessor macros begin with "PNG". We request that applications that
5300 use libpng *not* begin any of their own symbols with either of these strings.
5302 We put a space after the "sizeof" operator and we omit the
5303 optional parentheses around its argument when the argument
5304 is an expression, not a type name, and we always enclose the
5305 sizeof operator, with its argument, in parentheses:
5307 (sizeof (png_uint_32))
5310 Prior to libpng-1.6.0 we used a "png_sizeof()" macro, formatted as
5311 though it were a function.
5313 Control keywords if, for, while, and switch are always followed by a space
5314 to distinguish them from function calls, which have no trailing space.
5316 We put a space after each comma and after each semicolon
5317 in "for" statements, and we put spaces before and after each
5318 C binary operator and after "for" or "while", and before
5319 "?". We don't put a space between a typecast and the expression
5320 being cast, nor do we put one between a function name and the
5321 left parenthesis that follows it:
5323 for (i = 2; i > 0; --i)
5324 y[i] = a(x) + (int)b;
5326 We prefer #ifdef and #ifndef to #if defined() and #if !defined()
5327 when there is only one macro being tested. We always use parentheses
5330 We express integer constants that are used as bit masks in hex format,
5331 with an even number of lower-case hex digits, and to make them unsigned
5332 (e.g., 0x00U, 0xffU, 0x0100U) and long if they are greater than 0x7fff
5335 We prefer to use underscores rather than camelCase in names, except
5336 for a few type names that we inherit from zlib.h.
5338 We prefer "if (something != 0)" and "if (something == 0)"
5339 over "if (something)" and if "(!something)", respectively.
5341 We do not use the TAB character for indentation in the C sources.
5343 Lines do not exceed 80 characters.
5345 Other rules can be inferred by inspecting the libpng source.
5347 XVI. Y2K Compliance in libpng
5349 Since the PNG Development group is an ad-hoc body, we can't make
5350 an official declaration.
5352 This is your unofficial assurance that libpng from version 0.71 and
5353 upward through 1.6.25 are Y2K compliant. It is my belief that earlier
5354 versions were also Y2K compliant.
5356 Libpng only has two year fields. One is a 2-byte unsigned integer
5357 that will hold years up to 65535. The other, which is deprecated,
5358 holds the date in text format, and will hold years up to 9999.
5361 "png_uint_16 year" in png_time_struct.
5364 "char time_buffer[29]" in png_struct. This is no longer used
5365 in libpng-1.6.x and will be removed from libpng-1.7.0.
5367 There are seven time-related functions:
5369 png_convert_to_rfc_1123_buffer() in png.c
5370 (formerly png_convert_to_rfc_1152() in error, and
5371 also formerly png_convert_to_rfc_1123())
5372 png_convert_from_struct_tm() in pngwrite.c, called
5374 png_convert_from_time_t() in pngwrite.c
5375 png_get_tIME() in pngget.c
5376 png_handle_tIME() in pngrutil.c, called in pngread.c
5377 png_set_tIME() in pngset.c
5378 png_write_tIME() in pngwutil.c, called in pngwrite.c
5380 All appear to handle dates properly in a Y2K environment. The
5381 png_convert_from_time_t() function calls gmtime() to convert from system
5382 clock time, which returns (year - 1900), which we properly convert to
5383 the full 4-digit year. There is a possibility that applications using
5384 libpng are not passing 4-digit years into the png_convert_to_rfc_1123()
5385 function, or that they are incorrectly passing only a 2-digit year
5386 instead of "year - 1900" into the png_convert_from_struct_tm() function,
5387 but this is not under our control. The libpng documentation has always
5388 stated that it works with 4-digit years, and the APIs have been
5391 The tIME chunk itself is also Y2K compliant. It uses a 2-byte unsigned
5392 integer to hold the year, and can hold years as large as 65535.
5394 zlib, upon which libpng depends, is also Y2K compliant. It contains
5395 no date-related code.
5398 Glenn Randers-Pehrson
5400 PNG Development Group