Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1 | libpng.txt - a description on how to use and modify libpng |
| 2 | |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 3 | libpng version 0.99 |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 4 | Updated and distributed by Glenn Randers-Pehrson <randeg@alumni.rpi.edu> |
| 5 | Copyright (c) 1998, Glenn Randers-Pehrson |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 6 | January 30, 1998 |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 7 | |
| 8 | based on: |
| 9 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 10 | libpng 1.0 beta 6 version 0.96 |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 11 | Updated and distributed by Andreas Dilger |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 12 | Copyright (c) 1996, 1997 Andreas Dilger |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 13 | May 28, 1997 |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 14 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 15 | libpng 1.0 beta 2 - version 0.88 |
| 16 | For conditions of distribution and use, see copyright notice in png.h |
| 17 | Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc. |
| 18 | January 26, 1996 |
| 19 | |
| 20 | Updated/rewritten per request in the libpng FAQ |
| 21 | Copyright (c) 1995 Frank J. T. Wojcik |
| 22 | December 18, 1995 && January 20, 1996 |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 23 | |
| 24 | I. Introduction |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 25 | |
| 26 | This file describes how to use and modify the PNG reference library |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 27 | (known as libpng) for your own use. There are five sections to this |
| 28 | file: introduction, structures, reading, writing, and modification and |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 29 | configuration notes for various special platforms. In addition to this |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 30 | file, example.c is a good starting point for using the library, as |
| 31 | it is heavily commented and should include everything most people |
| 32 | will need. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 33 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 34 | Libpng was written as a companion to the PNG specification, as a way |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 35 | of reducing the amount of time and effort it takes to support the PNG |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 36 | file format in application programs. The PNG specification is available |
| 37 | as RFC 2083 <ftp://ftp.uu.net/graphics/png/documents/> and as a |
| 38 | W3C Recommendation <http://www.w3.org/pub/WWW/TR/REC.png.html>. Some |
| 39 | additional chunks are described in the special-purpose public chunks |
| 40 | documents at <ftp://ftp.uu.net/graphics/png/documents/>. Other information |
| 41 | about PNG can be found at the PNG home page, <http://www.cdrom.com/pub/png/>. |
| 42 | |
| 43 | Most users will not have to modify the library significantly; advanced |
| 44 | users may want to modify it more. All attempts were made to make it as |
| 45 | complete as possible, while keeping the code easy to understand. |
| 46 | Currently, this library only supports C. Support for other languages |
| 47 | is being considered. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 48 | |
| 49 | Libpng has been designed to handle multiple sessions at one time, |
| 50 | to be easily modifiable, to be portable to the vast majority of |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 51 | machines (ANSI, K&R, 16-, 32-, and 64-bit) available, and to be easy |
| 52 | to use. The ultimate goal of libpng is to promote the acceptance of |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 53 | the PNG file format in whatever way possible. While there is still |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 54 | work to be done (see the TODO file), libpng should cover the |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 55 | majority of the needs of its users. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 56 | |
| 57 | Libpng uses zlib for its compression and decompression of PNG files. |
| 58 | The zlib compression utility is a general purpose utility that is |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 59 | useful for more than PNG files, and can be used without libpng. |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 60 | See the documentation delivered with zlib for more details. |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 61 | You can usually find the source files for the zlib utility wherever you |
| 62 | find the libpng source files. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 63 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 64 | Libpng is thread safe, provided the threads are using different |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 65 | instances of the structures. Each thread should have its own |
| 66 | png_struct and png_info instances, and thus its own image. |
| 67 | Libpng does not protect itself against two threads using the |
| 68 | same instance of a structure. |
| 69 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 70 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 71 | |
| 72 | II. Structures |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 73 | |
| 74 | There are two main structures that are important to libpng, png_struct |
| 75 | and png_info. The first, png_struct, is an internal structure that |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 76 | will not, for the most part, be used by a user except as the first |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 77 | variable passed to every libpng function call. It is not actually |
| 78 | used in many of the functions; do not be alarmed about compiler |
| 79 | warnings that say something to the effect that "png_ptr is not used." |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 80 | |
| 81 | The png_info structure is designed to provide information about the |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 82 | PNG file. At one time, the fields of png_info were intended to be |
| 83 | directly accessible to the user. However, this tended to cause problems |
| 84 | with applications using dynamically loaded libraries, and as a result |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 85 | a set of interface functions for png_info was developed. The fields |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 86 | of png_info are still available for older applications, but it is |
| 87 | suggested that applications use the new interfaces if at all possible. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 88 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 89 | The png.h header file is an invaluable reference for programming with libpng. |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 90 | And while I'm on the topic, make sure you include the libpng header file: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 91 | |
| 92 | #include <png.h> |
| 93 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 94 | |
| 95 | |
| 96 | III. Reading |
| 97 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 98 | Reading PNG files: |
| 99 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 100 | We'll now walk you through the possible functions to call when reading |
| 101 | in a PNG file, briefly explaining the syntax and purpose of each one. |
| 102 | See example.c and png.h for more detail. While Progressive reading |
| 103 | is covered in the next section, you will still need some of the |
| 104 | functions discussed in this section to read a PNG file. |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 105 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 106 | You will want to do the I/O initialization(*) before you get into libpng, |
| 107 | so if it doesn't work, you don't have much to undo. Of course, you |
| 108 | will also want to insure that you are, in fact, dealing with a PNG |
| 109 | file. Libpng provides a simple check to see if a file is a PNG file. |
| 110 | To use it, pass in the first 1 to 8 bytes of the file, and it will |
| 111 | return true or false (1 or 0) depending on whether the bytes could be |
| 112 | part of a PNG file. Of course, the more bytes you pass in, the |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 113 | greater the accuracy of the prediction. |
| 114 | |
| 115 | If you are intending to keep the file pointer open for use in libpng, |
| 116 | you must ensure you don't read more than 8 bytes from the beginning |
| 117 | of the file, and you also have to make a call to png_set_sig_bytes_read() |
| 118 | with the number of bytes you read from the beginning. Libpng will |
| 119 | then only check the bytes (if any) that your program didn't read. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 120 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 121 | (*): If you are not using the standard I/O functions, you will need |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 122 | to replace them with custom functions. See the discussion under |
| 123 | Customizing libpng. |
| 124 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 125 | FILE *fp = fopen(file_name, "rb"); |
| 126 | if (!fp) |
| 127 | { |
| 128 | return; |
| 129 | } |
| 130 | fread(header, 1, number, fp); |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 131 | is_png = png_check_sig(header, 0, number); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 132 | if (!is_png) |
| 133 | { |
| 134 | return; |
| 135 | } |
| 136 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 137 | Next, png_struct and png_info need to be allocated and initialized. In |
| 138 | order to ensure that the size of these structures is correct even with a |
| 139 | dynamically linked libpng, there are functions to initialize and |
| 140 | allocate the structures. We also pass the library version, optional |
| 141 | pointers to error handling functions, and a pointer to a data struct for |
| 142 | use by the error functions, if necessary (the pointer and functions can |
| 143 | be NULL if the default error handlers are to be used). See the section |
| 144 | on Changes to Libpng below regarding the old initialization functions. |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 145 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 146 | png_structp png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING, |
| 147 | (void *)user_error_ptr, user_error_fn, user_warning_fn); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 148 | if (!png_ptr) |
| 149 | return; |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 150 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 151 | png_infop info_ptr = png_create_info_struct(png_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 152 | if (!info_ptr) |
| 153 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 154 | png_destroy_read_struct(&png_ptr, (png_infopp)NULL, (png_infopp)NULL); |
| 155 | return; |
| 156 | } |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 157 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 158 | png_infop end_info = png_create_info_struct(png_ptr); |
| 159 | if (!end_info) |
| 160 | { |
| 161 | png_destroy_read_struct(&png_ptr, &info_ptr, (png_infopp)NULL); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 162 | return; |
| 163 | } |
| 164 | |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 165 | The error handling routines passed to png_create_read_struct() are only |
| 166 | necessary if you are not using the libpng supplied error handling |
| 167 | functions. When libpng encounters an error, it expects to longjmp back |
| 168 | to your routine. Therefore, you will need to call setjmp and pass the |
| 169 | jmpbuf field of your png_struct. If you read the file from different |
| 170 | routines, you will need to update the jmpbuf field every time you enter |
| 171 | a new routine that will call a png_ function. |
| 172 | |
| 173 | See your documentation of setjmp/longjmp for your compiler for more |
| 174 | information on setjmp/longjmp. See the discussion on libpng error |
| 175 | handling in the Customizing Libpng section below for more information on |
| 176 | the libpng error handling. If an error occurs, and libpng longjmp's |
| 177 | back to your setjmp, you will want to call png_destroy_read_struct() to |
| 178 | free any memory. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 179 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 180 | if (setjmp(png_ptr->jmpbuf)) |
| 181 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 182 | png_destroy_read_struct(&png_ptr, &info_ptr, &end_info); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 183 | fclose(fp); |
| 184 | return; |
| 185 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 186 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 187 | Now you need to set up the input code. The default for libpng is to |
| 188 | use the C function fread(). If you use this, you will need to pass a |
| 189 | valid FILE * in the function png_init_io(). Be sure that the file is |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 190 | opened in binary mode. If you wish to handle reading data in another |
| 191 | way, you need not call the png_init_io() function, but you must then |
| 192 | implement the libpng I/O methods discussed in the Customizing Libpng |
| 193 | section below. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 194 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 195 | png_init_io(png_ptr, fp); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 196 | |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 197 | If you had previously opened the file and read any of the signature from |
| 198 | the beginning in order to see if this was a PNG file, you need to let |
| 199 | libpng know that there are some bytes missing from the start of the file. |
| 200 | |
| 201 | png_set_sig_bytes(png_ptr, number); |
| 202 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 203 | In PNG files, the alpha channel in an image is the level of opacity. |
| 204 | If you need the alpha channel in an image to be the level of transparency |
| 205 | instead of opacity, you can invert the alpha channel (or the tRNS chunk |
| 206 | data) after it's read, so that 0 is fully opaque and 255 (in 8-bit or |
| 207 | paletted images) or 65535 (in 16-bit images) is fully transparent, with |
| 208 | |
| 209 | png_set_invert_alpha(png_ptr); |
| 210 | |
| 211 | This has to appear here rather than later with the other transformations |
| 212 | because the tRNS chunk data must be modified in the case of paletted images. |
| 213 | If your image is not a paletted image, the tRNS data (which in such cases |
| 214 | represents a single color to be rendered as transparent) won't be changed. |
| 215 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 216 | You are now ready to read all the file information up to the actual |
| 217 | image data. You do this with a call to png_read_info(). |
| 218 | |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 219 | png_read_info(png_ptr, info_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 220 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 221 | Functions are used to get the information from the info_ptr: |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 222 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 223 | png_get_IHDR(png_ptr, info_ptr, &width, &height, &bit_depth, &color_type, |
| 224 | &interlace_type, &compression_type, &filter_type); |
| 225 | |
| 226 | width - holds the width of the image in pixels (up to 2^31). |
| 227 | height - holds the height of the image in pixels (up to 2^31). |
| 228 | bit_depth - holds the bit depth of one of the image channels. |
| 229 | (valid values are 1, 2, 4, 8, 16 and depend also on the |
| 230 | color_type. See also significant bits (sBIT) below). |
| 231 | color_type - describes which color/alpha channels are present. |
| 232 | PNG_COLOR_TYPE_GRAY (bit depths 1, 2, 4, 8, 16) |
| 233 | PNG_COLOR_TYPE_GRAY_ALPHA (bit depths 8, 16) |
| 234 | PNG_COLOR_TYPE_PALETTE (bit depths 1, 2, 4, 8) |
| 235 | PNG_COLOR_TYPE_RGB (bit_depths 8, 16) |
| 236 | PNG_COLOR_TYPE_RGB_ALPHA (bit_depths 8, 16) |
| 237 | |
| 238 | PNG_COLOR_MASK_PALETTE |
| 239 | PNG_COLOR_MASK_COLOR |
| 240 | PNG_COLOR_MASK_ALPHA |
| 241 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 242 | filter_type - (must be PNG_FILTER_TYPE_BASE for PNG 1.0) |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 243 | compression_type - (must be PNG_COMPRESSION_TYPE_BASE for PNG 1.0) |
| 244 | interlace_type - (PNG_INTERLACE_TYPE_NONE or PNG_INTERLACE_TYPE_ADAM7) |
| 245 | Any or all of interlace_type, compression_type, of filter_type can be |
| 246 | NULL if you are not interested in their values. |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 247 | |
| 248 | channels = png_get_channels(png_ptr, info_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 249 | channels - number of channels of info for the color type |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 250 | (valid values are 1 (GRAY, PALETTE), 2 (GRAY_ALPHA), |
| 251 | 3 (RGB), 4 (RGB_ALPHA or RGB + filler byte)) |
| 252 | rowbytes = png_get_rowbytes(png_ptr, info_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 253 | rowbytes - number of bytes needed to hold a row |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 254 | |
| 255 | signature = png_get_signature(png_ptr, info_ptr); |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 256 | signature - holds the signature read from the file (if any). The |
| 257 | data is kept in the same offset it would be if the |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 258 | whole signature were read (ie if an application had |
| 259 | already read in 4 bytes of signature before staring |
| 260 | libpng, the remaining 4 bytes would be in signature[4] |
| 261 | through signature[7] (see png_set_sig_bytes())). |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 262 | |
| 263 | Information from each if the IHDR fields can be retrieve separately as well: |
| 264 | |
| 265 | width = png_get_image_width(png_ptr, info_ptr); |
| 266 | height = png_get_image_height(png_ptr, info_ptr); |
| 267 | bit_depth = png_get_bit_depth(png_ptr, info_ptr); |
| 268 | color_type = png_get_color_type(png_ptr, info_ptr); |
| 269 | filter_type = png_get_filter_type(png_ptr, info_ptr); |
| 270 | compression_type = png_get_compression_type(png_ptr, info_ptr); |
| 271 | interlace_type = png_get_interlace_type(png_ptr, info_ptr); |
| 272 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 273 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 274 | These are also important, but their validity depends on whether the chunk |
| 275 | has been read. The png_get_valid(png_ptr, info_ptr, PNG_INFO_<chunk>) and |
| 276 | png_get_<chunk>(png_ptr, info_ptr, ...) functions return non-zero if the |
| 277 | data has been read, or zero if it is missing. The parameters to the |
| 278 | png_get_<chunk> are set directly if they are simple data types, or a pointer |
| 279 | into the info_ptr is returned for any complex types. |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 280 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 281 | png_get_PLTE(png_ptr, info_ptr, &palette, &num_palette); |
| 282 | palette - the palette for the file (array of png_color) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 283 | num_palette - number of entries in the palette |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 284 | |
| 285 | png_get_gAMA(png_ptr, info_ptr, &gamma); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 286 | gamma - the gamma the file is written at (PNG_INFO_gAMA) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 287 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 288 | png_get_sRGB(png_ptr, info_ptr, &srgb_intent); |
| 289 | srgb_intent - the rendering intent (PNG_INFO_sRGB) |
| 290 | The presence of the sRGB chunk means that the pixel |
| 291 | data is in the sRGB color space. This chunk also |
| 292 | implies specific values of gAMA and cHRM. |
| 293 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 294 | png_get_sBIT(png_ptr, info_ptr, &sig_bit); |
| 295 | sig_bit - the number of significant bits for (PNG_INFO_sBIT) |
| 296 | the gray, red, green, and blue channels, whichever |
| 297 | are appropriate for the given color type (png_color_16) |
| 298 | |
| 299 | png_get_tRNS(png_ptr, info_ptr, &trans, &num_trans, &trans_values); |
| 300 | trans - array of transparent entries for palette (PNG_INFO_tRNS) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 301 | trans_values - transparent pixel for non-paletted images (PNG_INFO_tRNS) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 302 | num_trans - number of transparent entries (PNG_INFO_tRNS) |
| 303 | |
| 304 | png_get_hIST(png_ptr, info_ptr, &hist); (PNG_INFO_hIST) |
| 305 | hist - histogram of palette (array of png_color_16) |
| 306 | |
| 307 | png_get_tIME(png_ptr, info_ptr, &mod_time); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 308 | mod_time - time image was last modified (PNG_VALID_tIME) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 309 | |
| 310 | png_get_bKGD(png_ptr, info_ptr, &background); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 311 | background - background color (PNG_VALID_bKGD) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 312 | |
| 313 | num_text = png_get_text(png_ptr, info_ptr, &text_ptr); |
| 314 | text_ptr - array of png_text holding image comments |
| 315 | text_ptr[i]->key - keyword for comment. |
| 316 | text_ptr[i]->text - text comments for current keyword. |
| 317 | text_ptr[i]->compression - type of compression used on "text" |
| 318 | PNG_TEXT_COMPRESSION_NONE or |
| 319 | PNG_TEXT_COMPRESSION_zTXt |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 320 | num_text - number of comments |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 321 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 322 | png_get_oFFs(png_ptr, info_ptr, &offset_x, &offset_y, &unit_type); |
| 323 | offset_x - positive offset from the left edge of the screen |
| 324 | offset_y - positive offset from the top edge of the screen |
| 325 | unit_type - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER |
| 326 | |
| 327 | png_get_pHYs(png_ptr, info_ptr, &res_x, &res_y, &unit_type); |
| 328 | res_x - pixels/unit physical resolution in x direction |
| 329 | res_y - pixels/unit physical resolution in x direction |
| 330 | unit_type - PNG_RESOLUTION_UNKOWN, PNG_RESOLUTION_METER |
| 331 | |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 332 | The data from the pHYs chunk can be retrieved in several convenient |
| 333 | forms: |
| 334 | |
| 335 | res_x_and_y = png_get_pixels_per_meter(png_ptr, info_ptr) |
| 336 | aspect_ratio = png_get_pixel_aspect_ratio(png_ptr, info_ptr) |
| 337 | |
| 338 | (Each of these returns 0 [signifying "unknown"] if the data is not |
| 339 | present or if res_x is 0; res_x_and_y is 0 if res_x != res_y) |
| 340 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 341 | For more information, see the png_info definition in png.h and the |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 342 | PNG specification for chunk contents. Be careful with trusting |
| 343 | rowbytes, as some of the transformations could increase the space |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 344 | needed to hold a row (expand, filler, gray_to_rgb, etc.). |
| 345 | See png_read_update_info(), below. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 346 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 347 | A quick word about text_ptr and num_text. PNG stores comments in |
| 348 | keyword/text pairs, one pair per chunk, with no limit on the number |
| 349 | of text chunks, and a 2^31 byte limit on their size. While there are |
| 350 | suggested keywords, there is no requirement to restrict the use to these |
| 351 | strings. It is strongly suggested that keywords and text be sensible |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 352 | to humans (that's the point), so don't use abbreviations. Non-printing |
| 353 | symbols are not allowed. See the PNG specification for more details. |
| 354 | There is also no requirement to have text after the keyword. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 355 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 356 | Keywords should be limited to 79 Latin-1 characters without leading or |
| 357 | trailing spaces, but non-consecutive spaces are allowed within the |
| 358 | keyword. It is possible to have the same keyword any number of times. |
| 359 | The text_ptr is an array of png_text structures, each holding pointer |
| 360 | to a keyword and a pointer to a text string. Only the text string may |
| 361 | be null. The keyword/text pairs are put into the array in the order |
| 362 | that they are received. However, some or all of the text chunks may be |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 363 | after the image, so, to make sure you have read all the text chunks, |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 364 | don't mess with these until after you read the stuff after the image. |
| 365 | This will be mentioned again below in the discussion that goes with |
| 366 | png_read_end(). |
| 367 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 368 | After you've read the header information, you can set up the library |
| 369 | to handle any special transformations of the image data. The various |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 370 | ways to transform the data will be described in the order that they |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 371 | should occur. This is important, as some of these change the color |
| 372 | type and/or bit depth of the data, and some others only work on |
| 373 | certain color types and bit depths. Even though each transformation |
| 374 | checks to see if it has data that it can do somthing with, you should |
| 375 | make sure to only enable a transformation if it will be valid for the |
| 376 | data. For example, don't swap red and blue on grayscale data. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 377 | |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 378 | The colors used for the background and transparency values should be |
| 379 | supplied in the same format/depth as the current image data. They |
| 380 | are stored in the same format/depth as the image data in a bKGD or tRNS |
| 381 | chunk, so this is what libpng expects for this data. The colors are |
| 382 | transformed to keep in sync with the image data when an application |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 383 | calls the png_read_update_info() routine (see below). |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 384 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 385 | Data will be decoded into the supplied row buffers packed into bytes |
| 386 | unless the library has been told to transform it into another format. |
| 387 | For example, 4 bit/pixel paletted or grayscale data will be returned |
| 388 | 2 pixels/byte with the leftmost pixel in the high-order bits of the |
| 389 | byte, unless png_set_packing() is called. 8-bit RGB data will be stored |
| 390 | in RGBRGBRGB format unless png_set_filler() is called to insert filler |
| 391 | bytes, either before or after each RGB triplet. 16-bit RGB data will |
| 392 | be returned RRGGBBRRGGBB, with the most significant byte of the color |
| 393 | value first, unless png_set_strip_16() is called to transform it to |
| 394 | regular RGBRGB triplets. |
| 395 | |
| 396 | The following code transforms grayscale images of less than 8 to 8 bits, |
| 397 | changes paletted images to RGB, and adds a full alpha channel if there is |
| 398 | transparency information in a tRNS chunk. This is most useful on |
| 399 | grayscale images with bit depths of 2 or 4 or if there is a multiple-image |
| 400 | viewing application that wishes to treat all images in the same way. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 401 | |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 402 | if (color_type == PNG_COLOR_TYPE_PALETTE && bit_depth <= 8) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 403 | png_set_expand(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 404 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 405 | if (color_type == PNG_COLOR_TYPE_GRAY && bit_depth < 8) |
| 406 | png_set_expand(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 407 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 408 | if (png_get_valid(png_ptr, info_ptr, PNG_INFO_tRNS)) |
| 409 | png_set_expand(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 410 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 411 | PNG can have files with 16 bits per channel. If you only can handle |
| 412 | 8 bits per channel, this will strip the pixels down to 8 bit. |
| 413 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 414 | if (bit_depth == 16) |
| 415 | png_set_strip_16(png_ptr); |
| 416 | |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 417 | The png_set_background() function tells libpng to composite images |
| 418 | with alpha or simple transparency against the supplied background |
| 419 | color. If the PNG file contains a bKGD chunk (PNG_INFO_bKGD valid), |
| 420 | you may use this color, or supply another color more suitable for |
| 421 | the current display (e.g., the background color from a web page). You |
| 422 | need to tell libpng whether the color is in the gamma space of the |
| 423 | display (PNG_BACKGROUND_GAMMA_SCREEN for colors you supply), the file |
| 424 | (PNG_BACKGROUND_GAMMA_FILE for colors from the bKGD chunk), or one |
| 425 | that is neither of these gammas (PNG_BACKGROUND_GAMMA_UNIQUE - I don't |
| 426 | know why anyone would use this, but it's here). |
| 427 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 428 | If, for some reason, you don't need the alpha channel on an image, |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 429 | and you want to remove it rather than combining it with the background |
| 430 | (but the image author certainly had in mind that you *would* combine |
| 431 | it with the background, so that's what you should probably do): |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 432 | |
| 433 | if (color_type & PNG_COLOR_MASK_ALPHA) |
| 434 | png_set_strip_alpha(png_ptr); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 435 | |
| 436 | PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as |
| 437 | they can, resulting in, for example, 8 pixels per byte for 1 bit |
| 438 | files. This code expands to 1 pixel per byte without changing the |
| 439 | values of the pixels: |
| 440 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 441 | if (bit_depth < 8) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 442 | png_set_packing(png_ptr); |
| 443 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 444 | PNG files have possible bit depths of 1, 2, 4, 8, and 16. All pixels |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 445 | stored in a PNG image have been "scaled" or "shifted" up to the next |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 446 | higher possible bit depth (eg from 5 bits/sample in the range [0,31] to |
| 447 | 8 bits/sample in the range [0, 255]). However, it is also possible to |
| 448 | convert the PNG pixel data back to the original bit depth of the image. |
| 449 | This call reduces the pixels back down to the original bit depth: |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 450 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 451 | png_color_16p sig_bit; |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 452 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 453 | if (png_get_sBIT(png_ptr, info_ptr, &sig_bit)) |
| 454 | png_set_shift(png_ptr, sig_bit); |
| 455 | |
| 456 | PNG files store 3-color pixels in red, green, blue order. This code |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 457 | changes the storage of the pixels to blue, green, red: |
| 458 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 459 | if (color_type == PNG_COLOR_TYPE_RGB || |
| 460 | color_type == PNG_COLOR_TYPE_RGB_ALPHA) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 461 | png_set_bgr(png_ptr); |
| 462 | |
| 463 | PNG files store RGB pixels packed into 3 bytes. This code expands them |
| 464 | into 4 bytes for windowing systems that need them in this format: |
| 465 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 466 | if (bit_depth == 8 && color_type == PNG_COLOR_TYPE_RGB) |
| 467 | png_set_filler(png_ptr, filler, PNG_FILLER_BEFORE); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 468 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 469 | where "filler" is the number to fill with, and the location is |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 470 | either PNG_FILLER_BEFORE or PNG_FILLER_AFTER, depending upon whether |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 471 | you want the filler before the RGB or after. This transformation |
| 472 | does not affect images that already have full alpha channels. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 473 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 474 | If you are reading an image with an alpha channel, and you need the |
| 475 | data as ARGB instead of the normal PNG format RGBA: |
| 476 | |
| 477 | if (color_type == PNG_COLOR_TYPE_RGB_ALPHA) |
| 478 | png_set_swap_alpha(png_ptr); |
| 479 | |
| 480 | For some uses, you may want a grayscale image to be represented as |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 481 | RGB. This code will do that conversion: |
| 482 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 483 | if (color_type == PNG_COLOR_TYPE_GRAY || |
| 484 | color_type == PNG_COLOR_TYPE_GRAY_ALPHA) |
| 485 | png_set_gray_to_rgb(png_ptr); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 486 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 487 | If you have a grayscale and you are using png_set_expand() to change to |
| 488 | a higher bit-depth you must indicate if the supplied background gray |
| 489 | is supplied in the original file bit depth (need_expand = 1) or in the |
| 490 | expanded bit depth (need_expand = 0). Similarly, if you are reading |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 491 | a paletted image, you must indicate if you have supplied the background |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 492 | as a palette index that needs to be expanded (need_expand = 1). You can |
| 493 | also specify an RGB triplet that isn't in the palette when setting your |
| 494 | background for a paletted image. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 495 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 496 | png_color_16 my_background; |
| 497 | png_color_16p image_background; |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 498 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 499 | if (png_get_bKGD(png_ptr, info_ptr, &image_background)) |
| 500 | png_set_background(png_ptr, image_background), |
| 501 | PNG_BACKGROUND_GAMMA_FILE, 1, 1.0); |
| 502 | else |
| 503 | png_set_background(png_ptr, &my_background, |
| 504 | PNG_BACKGROUND_GAMMA_SCREEN, 0, 1.0); |
| 505 | |
| 506 | To properly display PNG images on any kind of system, the application needs |
| 507 | to know what the display gamma is. Ideally, the user will know this, and |
| 508 | the application will allow them to set it. One method of allowing the user |
| 509 | to set the display gamma separately for each system is to check for the |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 510 | DISPLAY_GAMMA and VIEWING_GAMMA environment variables or for a SCREEN_GAMMA |
| 511 | environment variable, which will hopefully be correctly set. |
| 512 | |
| 513 | Note that display_gamma is the gamma of your display, while screen_gamma is |
| 514 | the overall gamma correction required to produce pleasing results, |
| 515 | which depends on the lighting conditions in the surrounding environment. |
| 516 | Screen_gamma is display_gamma/viewing_gamma, where viewing_gamma is |
| 517 | the amount of additional gamma correction needed to compensate for |
| 518 | a dim (viewing_gamma=1.125) or dark (viewing_gamma=1.25) environment. |
| 519 | In a brightly lit room, no compensation other than the display_gamma |
| 520 | is needed (viewing_gamma=1.0). |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 521 | |
| 522 | if (/* We have a user-defined screen gamma value */) |
| 523 | { |
| 524 | screen_gamma = user_defined_screen_gamma; |
| 525 | } |
| 526 | /* One way that applications can share the same screen gamma value */ |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 527 | else if ((gamma_str = getenv("SCREEN_GAMMA")) != NULL) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 528 | { |
| 529 | screen_gamma = atof(gamma_str); |
| 530 | } |
| 531 | /* If we don't have another value */ |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 532 | else |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 533 | { |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 534 | screen_gamma = 2.5; /* A good guess for a PC monitor in a bright office */ |
| 535 | screen_gamma = 2.2; /* A good guess for a PC monitor in a dim room */ |
| 536 | screen_gamma = 2.0; /* A good guess for a PC monitor in a dark room */ |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 537 | screen_gamma = 1.7 or 1.0; /* A good guess for Mac systems */ |
| 538 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 539 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 540 | The png_set_gamma() function handles gamma transformations of the data. |
| 541 | Pass both the file gamma and the current screen_gamma. If the file does |
| 542 | not have a gamma value, you can pass one anyway if you have an idea what |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 543 | it is (usually 0.50 is a good guess for GIF images on PCs). Note |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 544 | that file gammas are inverted from screen gammas. See the discussions |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 545 | on gamma in the PNG specification for an excellent description of what |
| 546 | gamma is, and why all applications should support it. It is strongly |
| 547 | recommended that PNG viewers support gamma correction. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 548 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 549 | if (png_get_gAMA(png_ptr, info_ptr, &gamma)) |
| 550 | png_set_gamma(png_ptr, screen_gamma, gamma); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 551 | else |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 552 | png_set_gamma(png_ptr, screen_gamma, 0.50); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 553 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 554 | If you need to reduce an RGB file to a paletted file, or if a paletted |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 555 | file has more entries then will fit on your screen, png_set_dither() |
| 556 | will do that. Note that this is a simple match dither that merely |
| 557 | finds the closest color available. This should work fairly well with |
| 558 | optimized palettes, and fairly badly with linear color cubes. If you |
| 559 | pass a palette that is larger then maximum_colors, the file will |
| 560 | reduce the number of colors in the palette so it will fit into |
| 561 | maximum_colors. If there is a histogram, it will use it to make |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 562 | more intelligent choices when reducing the palette. If there is no |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 563 | histogram, it may not do as good a job. |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 564 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 565 | if (color_type & PNG_COLOR_MASK_COLOR) |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 566 | { |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 567 | if (png_get_valid(png_ptr, info_ptr, PNG_INFO_PLTE)) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 568 | { |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 569 | png_color_16p histogram; |
| 570 | |
| 571 | png_get_hIST(png_ptr, info_ptr, &histogram); |
| 572 | png_set_dither(png_ptr, palette, num_palette, max_screen_colors, |
| 573 | histogram, 1); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 574 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 575 | else |
| 576 | { |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 577 | png_color std_color_cube[MAX_SCREEN_COLORS] = |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 578 | { ... colors ... }; |
| 579 | |
| 580 | png_set_dither(png_ptr, std_color_cube, MAX_SCREEN_COLORS, |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 581 | MAX_SCREEN_COLORS, NULL,0); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 582 | } |
| 583 | } |
| 584 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 585 | PNG files describe monochrome as black being zero and white being one. |
| 586 | The following code will reverse this (make black be one and white be |
| 587 | zero): |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 588 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 589 | if (bit_depth == 1 && color_type == PNG_COLOR_GRAY) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 590 | png_set_invert_mono(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 591 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 592 | PNG files store 16 bit pixels in network byte order (big-endian, |
| 593 | ie. most significant bits first). This code chages the storage to the |
| 594 | other way (little-endian, ie. least significant bits first, eg. the |
| 595 | way PCs store them): |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 596 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 597 | if (bit_depth == 16) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 598 | png_set_swap(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 599 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 600 | If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you |
| 601 | need to change the order the pixels are packed into bytes, you can use: |
| 602 | |
| 603 | if (bit_depth < 8) |
| 604 | png_set_packswap(png_ptr); |
| 605 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 606 | The last thing to handle is interlacing; this is covered in detail below, |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 607 | but you must call the function here if you want libpng to handle expansion |
| 608 | of the interlaced image. |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 609 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 610 | number_of_passes = png_set_interlace_handling(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 611 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 612 | After setting the transformations, libpng can update your png_info |
| 613 | structure to reflect any transformations you've requested with this |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 614 | call. This is most useful to update the info structure's rowbytes |
| 615 | field so you can use it to allocate your image memory. This function |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 616 | will also update your palette with the correct screen_gamma and |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 617 | background if these have been given with the calls above. |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 618 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 619 | png_read_update_info(png_ptr, info_ptr); |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 620 | |
| 621 | After you call png_read_update_info(), you can allocate any |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 622 | memory you need to hold the image. The row data is simply |
| 623 | raw byte data for all forms of images. As the actual allocation |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 624 | varies among applications, no example will be given. If you |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 625 | are allocating one large chunk, you will need to build an |
| 626 | array of pointers to each row, as it will be needed for some |
| 627 | of the functions below. |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 628 | |
| 629 | After you've allocated memory, you can read the image data. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 630 | The simplest way to do this is in one function call. If you are |
| 631 | allocating enough memory to hold the whole image, you can just |
| 632 | call png_read_image() and libpng will read in all the image data |
| 633 | and put it in the memory area supplied. You will need to pass in |
| 634 | an array of pointers to each row. |
| 635 | |
| 636 | This function automatically handles interlacing, so you don't need |
| 637 | to call png_set_interlace_handling() or call this function multiple |
| 638 | times, or any of that other stuff necessary with png_read_rows(). |
| 639 | |
| 640 | png_read_image(png_ptr, row_pointers); |
| 641 | |
| 642 | where row_pointers is: |
| 643 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 644 | png_bytep row_pointers[height]; |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 645 | |
| 646 | You can point to void or char or whatever you use for pixels. |
| 647 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 648 | If you don't want to read in the whole image at once, you can |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 649 | use png_read_rows() instead. If there is no interlacing (check |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 650 | interlace_type == PNG_INTERLACE_TYPE_NONE), this is simple: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 651 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 652 | png_read_rows(png_ptr, row_pointers, NULL, number_of_rows); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 653 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 654 | where row_pointers is the same as in the png_read_image() call. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 655 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 656 | If you are doing this just one row at a time, you can do this with |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 657 | row_pointers: |
| 658 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 659 | png_bytep row_pointers = row; |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 660 | png_read_row(png_ptr, &row_pointers, NULL); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 661 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 662 | If the file is interlaced (info_ptr->interlace_type != 0), things get |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 663 | somewhat harder. The only current (PNG Specification version 1.0) |
| 664 | interlacing type for PNG is (interlace_type == PNG_INTERLACE_TYPE_ADAM7) |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 665 | is a somewhat complicated 2D interlace scheme, known as Adam7, that |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 666 | breaks down an image into seven smaller images of varying size, based |
| 667 | on an 8x8 grid. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 668 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 669 | libpng can fill out those images or it can give them to you "as is". |
| 670 | If you want them filled out, there are two ways to do that. The one |
| 671 | mentioned in the PNG specification is to expand each pixel to cover |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 672 | those pixels that have not been read yet (the "rectangle" method). |
| 673 | This results in a blocky image for the first pass, which gradually |
| 674 | smooths out as more pixels are read. The other method is the "sparkle" |
| 675 | method, where pixels are drawn only in their final locations, with the |
| 676 | rest of the image remaining whatever colors they were initialized to |
| 677 | before the start of the read. The first method usually looks better, |
| 678 | but tends to be slower, as there are more pixels to put in the rows. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 679 | |
| 680 | If you don't want libpng to handle the interlacing details, just call |
| 681 | png_read_rows() seven times to read in all seven images. Each of the |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 682 | images is a valid image by itself, or they can all be combined on an |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 683 | 8x8 grid to form a single image (although if you intend to combine them |
| 684 | you would be far better off using the libpng interlace handling). |
| 685 | |
| 686 | The first pass will return an image 1/8 as wide as the entire image |
| 687 | (every 8th column starting in column 0) and 1/8 as high as the original |
| 688 | (every 8th row starting in row 0), the second will be 1/8 as wide |
| 689 | (starting in column 4) and 1/8 as high (also starting in row 0). The |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 690 | third pass will be 1/4 as wide (every 4th pixel starting in column 0) and |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 691 | 1/8 as high (every 8th row starting in row 4), and the fourth pass will |
| 692 | be 1/4 as wide and 1/4 as high (every 4th column starting in column 2, |
| 693 | and every 4th row starting in row 0). The fifth pass will return an |
| 694 | image 1/2 as wide, and 1/4 as high (starting at column 0 and row 2), |
| 695 | while the sixth pass will be 1/2 as wide and 1/2 as high as the original |
| 696 | (starting in column 1 and row 0). The seventh and final pass will be as |
| 697 | wide as the original, and 1/2 as high, containing all of the odd |
| 698 | numbered scanlines. Phew! |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 699 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 700 | If you want libpng to expand the images, call this before calling |
| 701 | png_start_read_image() or png_read_update_info(): |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 702 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 703 | if (interlace_type == PNG_INTERLACE_TYPE_ADAM7) |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 704 | number_of_passes = png_set_interlace_handling(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 705 | |
| 706 | This will return the number of passes needed. Currently, this |
| 707 | is seven, but may change if another interlace type is added. |
| 708 | This function can be called even if the file is not interlaced, |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 709 | where it will return one pass. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 710 | |
| 711 | If you are not going to display the image after each pass, but are |
| 712 | going to wait until the entire image is read in, use the sparkle |
| 713 | effect. This effect is faster and the end result of either method |
| 714 | is exactly the same. If you are planning on displaying the image |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 715 | after each pass, the "rectangle" effect is generally considered the |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 716 | better looking one. |
| 717 | |
| 718 | If you only want the "sparkle" effect, just call png_read_rows() as |
| 719 | normal, with the third parameter NULL. Make sure you make pass over |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 720 | the image number_of_passes times, and you don't change the data in the |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 721 | rows between calls. You can change the locations of the data, just |
| 722 | not the data. Each pass only writes the pixels appropriate for that |
| 723 | pass, and assumes the data from previous passes is still valid. |
| 724 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 725 | png_read_rows(png_ptr, row_pointers, NULL, number_of_rows); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 726 | |
| 727 | If you only want the first effect (the rectangles), do the same as |
| 728 | before except pass the row buffer in the third parameter, and leave |
| 729 | the second parameter NULL. |
| 730 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 731 | png_read_rows(png_ptr, NULL, row_pointers, number_of_rows); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 732 | |
| 733 | After you are finished reading the image, you can finish reading |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 734 | the file. If you are interested in comments or time, which may be |
| 735 | stored either before or after the image data, you should pass the |
| 736 | info_ptr pointer from the png_read_info() call, or you can pass a |
| 737 | separate png_info struct if you want to keep the comments from |
| 738 | before and after the image separate. If you are not interested, you |
| 739 | can pass NULL. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 740 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 741 | png_read_end(png_ptr, end_info); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 742 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 743 | When you are done, you can free all memory allocated by libpng like this: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 744 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 745 | png_destroy_read_struct(&png_ptr, &info_ptr, &end_info); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 746 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 747 | For a more compact example of reading a PNG image, see the file example.c. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 748 | |
| 749 | |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 750 | Reading PNG files progressively: |
| 751 | |
| 752 | The progressive reader is slightly different then the non-progressive |
| 753 | reader. Instead of calling png_read_info(), png_read_rows(), and |
| 754 | png_read_end(), you make one call to png_process_data(), which calls |
| 755 | callbacks when it has the info, a row, or the end of the image. You |
| 756 | set up these callbacks with png_set_progressive_read_fn(). You don't |
| 757 | have to worry about the input/output functions of libpng, as you are |
| 758 | giving the library the data directly in png_process_data(). I will |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 759 | assume that you have read the section on reading PNG files above, |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 760 | so I will only highlight the differences (although I will show |
| 761 | all of the code). |
| 762 | |
| 763 | png_structp png_ptr; |
| 764 | png_infop info_ptr; |
| 765 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 766 | /* An example code fragment of how you would initialize the progressive |
| 767 | reader in your application. */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 768 | int |
| 769 | initialize_png_reader() |
| 770 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 771 | png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING, |
| 772 | (void *)user_error_ptr, user_error_fn, user_warning_fn); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 773 | if (!png_ptr) |
| 774 | return -1; |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 775 | info_ptr = png_create_info_struct(png_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 776 | if (!info_ptr) |
| 777 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 778 | png_destroy_read_struct(&png_ptr, (png_infopp)NULL, (png_infopp)NULL); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 779 | return -1; |
| 780 | } |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 781 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 782 | if (setjmp(png_ptr->jmpbuf)) |
| 783 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 784 | png_destroy_read_struct(&png_ptr, &info_ptr, (png_infopp)NULL); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 785 | return -1; |
| 786 | } |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 787 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 788 | /* This one's new. You can provide functions to be called |
| 789 | when the header info is valid, when each row is completed, |
| 790 | and when the image is finished. If you aren't using all |
| 791 | functions, you can specify a NULL parameter. You can use |
| 792 | any struct as the user_ptr (cast to a void pointer for the |
| 793 | function call), and retrieve the pointer from inside the |
| 794 | callbacks using the function png_get_progressive_ptr(png_ptr); |
| 795 | which will return a void pointer, which you have to cast |
| 796 | appropriately. |
| 797 | */ |
| 798 | png_set_progressive_read_fn(png_ptr, (void *)user_ptr, |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 799 | info_callback, row_callback, end_callback); |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 800 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 801 | return 0; |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 802 | } |
| 803 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 804 | /* A code fragment that you call as you recieve blocks of data */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 805 | int |
| 806 | process_data(png_bytep buffer, png_uint_32 length) |
| 807 | { |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 808 | if (setjmp(png_ptr->jmpbuf)) |
| 809 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 810 | png_destroy_read_struct(&png_ptr, &info_ptr, (png_infopp)NULL); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 811 | return -1; |
| 812 | } |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 813 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 814 | /* This one's new also. Simply give it a chunk of data |
| 815 | from the file stream (in order, of course). On machines |
| 816 | with segmented memory models machines, don't give it any |
| 817 | more than 64K. The library seems to run fine with sizes |
| 818 | of 4K. Although you can give it much less if necessary |
| 819 | (I assume you can give it chunks of 1 byte, I haven't |
| 820 | tried less then 256 bytes yet). When this function returns, |
| 821 | you may want to display any rows that were generated in the |
| 822 | row callback if you don't already do so there. |
| 823 | */ |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 824 | png_process_data(png_ptr, info_ptr, buffer, length); |
| 825 | return 0; |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 826 | } |
| 827 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 828 | /* This function is called (as set by png_set_progressive_fn() above) |
| 829 | when enough data has been supplied so all of the header has been read. |
| 830 | */ |
| 831 | void |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 832 | info_callback(png_structp png_ptr, png_infop info) |
| 833 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 834 | /* Do any setup here, including setting any of the transformations |
| 835 | mentioned in the Reading PNG files section. For now, you _must_ |
| 836 | call either png_start_read_image() or png_read_update_info() |
| 837 | after all the transformations are set (even if you don't set |
| 838 | any). You may start getting rows before png_process_data() |
| 839 | returns, so this is your last chance to prepare for that. |
| 840 | */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 841 | } |
| 842 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 843 | /* This function is called when each row of image data is complete */ |
| 844 | void |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 845 | row_callback(png_structp png_ptr, png_bytep new_row, |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 846 | png_uint_32 row_num, int pass) |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 847 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 848 | /* If the image is interlaced, and you turned on the interlace |
| 849 | handler, this function will be called for every row in every pass. |
| 850 | Some of these rows will not be changed from the previous pass. |
| 851 | When the row is not changed, the new_row variable will be NULL. |
| 852 | The rows and passes are called in order, so you don't really |
| 853 | need the row_num and pass, but I'm supplying them because it |
| 854 | may make your life easier. |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 855 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 856 | For the non-NULL rows of interlaced images, you must call |
| 857 | png_progressive_combine_row() passing in the row and the |
| 858 | old row. You can call this function for NULL rows (it will |
| 859 | just return) and for non-interlaced images (it just does the |
| 860 | memcpy for you) if it will make the code easier. Thus, you |
| 861 | can just do this for all cases: |
| 862 | */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 863 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 864 | png_progressive_combine_row(png_ptr, old_row, new_row); |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 865 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 866 | /* where old_row is what was displayed for previous rows. Note |
| 867 | that the first pass (pass == 0, really) will completely cover |
| 868 | the old row, so the rows do not have to be initialized. After |
| 869 | the first pass (and only for interlaced images), you will have |
| 870 | to pass the current row, and the function will combine the |
| 871 | old row and the new row. |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 872 | */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 873 | } |
| 874 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 875 | void |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 876 | end_callback(png_structp png_ptr, png_infop info) |
| 877 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 878 | /* This function is called after the whole image has been read, |
| 879 | including any chunks after the image (up to and including |
| 880 | the IEND). You will usually have the same info chunk as you |
| 881 | had in the header, although some data may have been added |
| 882 | to the comments and time fields. |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 883 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 884 | Most people won't do much here, perhaps setting a flag that |
| 885 | marks the image as finished. |
| 886 | */ |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 887 | } |
| 888 | |
| 889 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 890 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 891 | IV. Writing |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 892 | |
| 893 | Much of this is very similar to reading. However, everything of |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 894 | importance is repeated here, so you won't have to constantly look |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 895 | back up in the reading section to understand writing. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 896 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 897 | You will want to do the I/O initialization before you get into libpng, |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 898 | so if it doesn't work, you don't have anything to undo. If you are not |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 899 | using the standard I/O functions, you will need to replace them with |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 900 | custom writing functions. See the discussion under Customizing libpng. |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 901 | |
| 902 | FILE *fp = fopen(file_name, "wb"); |
| 903 | if (!fp) |
| 904 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 905 | return; |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 906 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 907 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 908 | Next, png_struct and png_info need to be allocated and initialized. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 909 | As these can be both relatively large, you may not want to store these |
| 910 | on the stack, unless you have stack space to spare. Of course, you |
| 911 | will want to check if they return NULL. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 912 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 913 | png_structp png_ptr = png_create_write_struct(PNG_LIBPNG_VER_STRING, |
| 914 | (void *)user_error_ptr, user_error_fn, user_warning_fn); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 915 | if (!png_ptr) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 916 | return; |
| 917 | |
| 918 | png_infop info_ptr = png_create_info_struct(png_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 919 | if (!info_ptr) |
| 920 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 921 | png_destroy_write_struct(&png_ptr, (png_infopp)NULL); |
| 922 | return; |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 923 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 924 | |
| 925 | After you have these structures, you will need to set up the |
| 926 | error handling. When libpng encounters an error, it expects to |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 927 | longjmp() back to your routine. Therefore, you will need to call |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 928 | setjmp and pass the jmpbuf field of your png_struct. If you |
| 929 | write the file from different routines, you will need to update |
| 930 | the jmpbuf field every time you enter a new routine that will |
| 931 | call a png_ function. See your documentation of setjmp/longjmp |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 932 | for your compiler for more information on setjmp/longjmp. See |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 933 | the discussion on libpng error handling in the Customizing Libpng |
| 934 | section below for more information on the libpng error handling. |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 935 | |
| 936 | if (setjmp(png_ptr->jmpbuf)) |
| 937 | { |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 938 | png_destroy_write_struct(&png_ptr, &info_ptr); |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 939 | fclose(fp); |
| 940 | return; |
| 941 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 942 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 943 | Now you need to set up the output code. The default for libpng is to |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 944 | use the C function fwrite(). If you use this, you will need to pass a |
| 945 | valid FILE * in the function png_init_io(). Be sure that the file is |
| 946 | opened in binary mode. Again, if you wish to handle writing data in |
| 947 | another way, see the discussion on libpng I/O handling in the Customizing |
| 948 | Libpng section below. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 949 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 950 | png_init_io(png_ptr, fp); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 951 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 952 | You now have the option of modifying how the compression library will |
| 953 | run. The following functions are mainly for testing, but may be useful |
Andreas Dilger | 02ad0ef | 1997-01-17 01:34:35 -0600 | [diff] [blame] | 954 | in some cases, like if you need to write PNG files extremely fast and |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 955 | are willing to give up some compression, or if you want to get the |
| 956 | maximum possible compression at the expense of slower writing. If you |
| 957 | have no special needs in this area, let the library do what it wants by |
| 958 | not calling this function at all, as it has been tuned to deliver a good |
| 959 | speed/compression ratio. The second parameter to png_set_filter() is |
| 960 | the filter method, for which the only valid value is '0' (as of the |
| 961 | 06/96 PNG specification. The third parameter is a flag that indicates |
| 962 | which filter type(s) are to be tested for each scanline. See the |
| 963 | Compression Library for details on the specific filter types. |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 964 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 965 | |
| 966 | /* turn on or off filtering, and/or choose specific filters */ |
| 967 | png_set_filter(png_ptr, 0, |
| 968 | PNG_FILTER_NONE | PNG_FILTER_SUB | PNG_FILTER_PAETH); |
| 969 | |
| 970 | The png_set_compression_???() functions interface to the zlib compression |
| 971 | library, and should mostly be ignored unless you really know what you are |
| 972 | doing. The only generally useful call is png_set_compression_level() |
| 973 | which changes how much time zlib spends on trying to compress the image |
| 974 | data. See the Compression Library for details on the compression levels. |
| 975 | |
| 976 | /* set the zlib compression level */ |
| 977 | png_set_compression_level(png_ptr, Z_BEST_COMPRESSION); |
| 978 | |
| 979 | /* set other zlib parameters */ |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 980 | png_set_compression_mem_level(png_ptr, 8); |
| 981 | png_set_compression_strategy(png_ptr, Z_DEFAULT_STRATEGY); |
| 982 | png_set_compression_window_bits(png_ptr, 15); |
| 983 | png_set_compression_method(png_ptr, 8); |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 984 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 985 | You now need to fill in the png_info structure with all the data you |
| 986 | wish to write before the actual image. Note that the only thing you |
| 987 | are allowed to write after the image is the text chunks and the time |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 988 | chunk (as of PNG Specification 1.0, anyway). See png_write_end() and |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 989 | the latest PNG specification for more information on that. If you |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 990 | wish to write them before the image, fill them in now, and flag that |
| 991 | data as being valid. If you want to wait until after the data, don't |
| 992 | fill them until png_write_end(). For all the fields in png_info and |
| 993 | their data types, see png.h. For explanations of what the fields |
| 994 | contain, see the PNG specification. |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 995 | |
| 996 | Some of the more important parts of the png_info are: |
| 997 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 998 | png_set_IHDR(png_ptr, info_ptr, width, height, bit_depth, color_type, |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 999 | interlace_type, compression_type, filter_type) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1000 | width - holds the width of the image in pixels (up to 2^31). |
| 1001 | height - holds the height of the image in pixels (up to 2^31). |
| 1002 | bit_depth - holds the bit depth of one of the image channels. |
| 1003 | (valid values are 1, 2, 4, 8, 16 and depend also on the |
| 1004 | color_type. See also significant bits (sBIT) below). |
| 1005 | color_type - describes which color/alpha channels are present. |
| 1006 | PNG_COLOR_TYPE_GRAY (bit depths 1, 2, 4, 8, 16) |
| 1007 | PNG_COLOR_TYPE_GRAY_ALPHA (bit depths 8, 16) |
| 1008 | PNG_COLOR_TYPE_PALETTE (bit depths 1, 2, 4, 8) |
| 1009 | PNG_COLOR_TYPE_RGB (bit_depths 8, 16) |
| 1010 | PNG_COLOR_TYPE_RGB_ALPHA (bit_depths 8, 16) |
| 1011 | |
| 1012 | PNG_COLOR_MASK_PALETTE |
| 1013 | PNG_COLOR_MASK_COLOR |
| 1014 | PNG_COLOR_MASK_ALPHA |
| 1015 | |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1016 | interlace_type - PNG_INTERLACE_TYPE_NONE or PNG_INTERLACE_TYPE_ADAM7 |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1017 | compression_type - (must be PNG_COMPRESSION_TYPE_DEFAULT for PNG 1.0) |
| 1018 | filter_type - (must be PNG_FILTER_TYPE_DEFAULT for PNG 1.0) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1019 | |
| 1020 | png_set_PLTE(png_ptr, info_ptr, palette, num_palette); |
| 1021 | palette - the palette for the file (array of png_color) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1022 | num_palette - number of entries in the palette |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1023 | |
| 1024 | png_set_gAMA(png_ptr, info_ptr, gamma); |
| 1025 | gamma - the gamma the image was created at (PNG_INFO_gAMA) |
| 1026 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1027 | png_set_sRGB(png_ptr, info_ptr, srgb_intent); |
| 1028 | srgb_intent - the rendering intent (PNG_INFO_sRGB) |
| 1029 | The presence of the sRGB chunk means that the pixel |
| 1030 | data is in the sRGB color space. This chunk also |
| 1031 | implies specific values of gAMA and cHRM. |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1032 | Rendering intent is the CSS-1 property that has been |
| 1033 | defined by the International Color Consortium |
| 1034 | (http://www.color.org). It can be one of |
| 1035 | PNG_SRGB_INTENT_SATURATION, PNG_SRGB_INTENT_PERCEPTUAL, |
| 1036 | PNG_SRGB_INTENT_ABSOLUTE, or PNG_SRGB_INTENT_RELATIVE. |
| 1037 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1038 | |
| 1039 | png_set_sRGB_gAMA_and_cHRM(png_ptr, info_ptr, srgb_intent); |
| 1040 | srgb_intent - the rendering intent (PNG_INFO_sRGB) |
| 1041 | The presence of the sRGB chunk means that the pixel |
| 1042 | data is in the sRGB color space. This chunk also |
| 1043 | causes gAMA and cHRM chunks with the specific values |
| 1044 | that are consistent with sRGB to be written. |
| 1045 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1046 | png_set_sBIT(png_ptr, info_ptr, sig_bit); |
| 1047 | sig_bit - the number of significant bits for (PNG_INFO_sBIT) |
| 1048 | the gray, red, green, and blue channels, whichever |
| 1049 | are appropriate for the given color type (png_color_16) |
| 1050 | |
| 1051 | png_set_tRNS(png_ptr, info_ptr, trans, num_trans, trans_values); |
| 1052 | trans - array of transparent entries for palette (PNG_INFO_tRNS) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1053 | trans_values - transparent pixel for non-paletted images (PNG_INFO_tRNS) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1054 | num_trans - number of transparent entries (PNG_INFO_tRNS) |
| 1055 | |
| 1056 | png_set_hIST(png_ptr, info_ptr, hist); (PNG_INFO_hIST) |
| 1057 | hist - histogram of palette (array of png_color_16) |
| 1058 | |
| 1059 | png_set_tIME(png_ptr, info_ptr, mod_time); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1060 | mod_time - time image was last modified (PNG_VALID_tIME) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1061 | |
| 1062 | png_set_bKGD(png_ptr, info_ptr, background); |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1063 | background - background color (PNG_VALID_bKGD) |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1064 | |
| 1065 | png_set_text(png_ptr, info_ptr, text_ptr, num_text); |
| 1066 | text_ptr - array of png_text holding image comments |
| 1067 | text_ptr[i]->key - keyword for comment. |
| 1068 | text_ptr[i]->text - text comments for current keyword. |
| 1069 | text_ptr[i]->compression - type of compression used on "text" |
| 1070 | PNG_TEXT_COMPRESSION_NONE or |
| 1071 | PNG_TEXT_COMPRESSION_zTXt |
| 1072 | num_text - number of comments in text_ptr |
| 1073 | |
| 1074 | png_set_oFFs(png_ptr, info_ptr, offset_x, offset_y, unit_type); |
| 1075 | offset_x - positive offset from the left edge of the screen |
| 1076 | offset_y - positive offset from the top edge of the screen |
| 1077 | unit_type - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER |
| 1078 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1079 | png_set_pHYs(png_ptr, info_ptr, res_x, res_y, unit_type); |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1080 | res_x - pixels/unit physical resolution in x direction |
| 1081 | res_y - pixels/unit physical resolution in x direction |
| 1082 | unit_type - PNG_RESOLUTION_UNKOWN, PNG_RESOLUTION_METER |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1083 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1084 | In PNG files, the alpha channel in an image is the level of opacity. |
| 1085 | If your data is supplied as a level of transparency, you can invert the |
| 1086 | alpha channel before you write it, so that 0 is fully transparent and 255 |
| 1087 | (in 8-bit or paletted images) or 65535 (in 16-bit images) is fully opaque, |
| 1088 | with |
| 1089 | |
| 1090 | png_set_invert_alpha(png_ptr); |
| 1091 | |
| 1092 | This must appear here instead of later with the other transformations |
| 1093 | because in the case of paletted images the tRNS chunk data has to |
| 1094 | be inverted before the tRNS chunk is written. If your image is not a |
| 1095 | paletted image, the tRNS data (which in such cases represents a single |
| 1096 | color to be rendered as transparent) won't be changed. |
| 1097 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1098 | A quick word about text and num_text. text is an array of png_text |
| 1099 | structures. num_text is the number of valid structures in the array. |
| 1100 | If you want, you can use max_text to hold the size of the array, but |
| 1101 | libpng ignores it for writing (it does use it for reading). Each |
| 1102 | png_text structure holds a keyword-text value, and a compression type. |
| 1103 | The compression types have the same valid numbers as the compression |
| 1104 | types of the image data. Currently, the only valid number is zero. |
| 1105 | However, you can store text either compressed or uncompressed, unlike |
| 1106 | images which always have to be compressed. So if you don't want the |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1107 | text compressed, set the compression type to PNG_TEXT_COMPRESSION_NONE. |
| 1108 | Until text gets around 1000 bytes, it is not worth compressing it. |
| 1109 | After the text has been written out to the file, the compression type |
| 1110 | is set to PNG_TEXT_COMPRESSION_NONE_WR or PNG_TEXT_COMPRESSION_zTXt_WR, |
| 1111 | so that it isn't written out again at the end (in case you are calling |
| 1112 | png_write_end() with the same struct. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1113 | |
| 1114 | The keywords that are given in the PNG Specification are: |
| 1115 | |
| 1116 | Title Short (one line) title or caption for image |
| 1117 | Author Name of image's creator |
| 1118 | Description Description of image (possibly long) |
| 1119 | Copyright Copyright notice |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 1120 | Creation Time Time of original image creation (usually |
| 1121 | RFC 1123 format, see below) |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1122 | Software Software used to create the image |
| 1123 | Disclaimer Legal disclaimer |
| 1124 | Warning Warning of nature of content |
| 1125 | Source Device used to create the image |
| 1126 | Comment Miscellaneous comment; conversion from other |
| 1127 | image format |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1128 | |
| 1129 | The keyword-text pairs work like this. Keywords should be short |
| 1130 | simple descriptions of what the comment is about. Some typical |
| 1131 | keywords are found in the PNG specification, as is some recomendations |
| 1132 | on keywords. You can repeat keywords in a file. You can even write |
| 1133 | some text before the image and some after. For example, you may want |
| 1134 | to put a description of the image before the image, but leave the |
| 1135 | disclaimer until after, so viewers working over modem connections |
| 1136 | don't have to wait for the disclaimer to go over the modem before |
| 1137 | they start seeing the image. Finally, keywords should be full |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1138 | words, not abbreviations. Keywords and text are in the ISO 8859-1 |
| 1139 | (Latin-1) character set (a superset of regular ASCII) and can not |
| 1140 | contain NUL characters, and should not contain control or other |
| 1141 | unprintable characters. To make the comments widely readable, stick |
| 1142 | with basic ASCII, and avoid machine specific character set extensions |
| 1143 | like the IBM-PC character set. The keyword must be present, but |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1144 | you can leave off the text string on non-compressed pairs. |
| 1145 | Compressed pairs must have a text string, as only the text string |
| 1146 | is compressed anyway, so the compression would be meaningless. |
| 1147 | |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1148 | PNG supports modification time via the png_time structure. Two |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1149 | conversion routines are proved, png_convert_from_time_t() for |
| 1150 | time_t and png_convert_from_struct_tm() for struct tm. The |
| 1151 | time_t routine uses gmtime(). You don't have to use either of |
| 1152 | these, but if you wish to fill in the png_time structure directly, |
| 1153 | you should provide the time in universal time (GMT) if possible |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1154 | instead of your local time. Note that the year number is the full |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1155 | year (ie 1996, rather than 96 - PNG is year 2000 compliant!), and |
| 1156 | that months start with 1. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1157 | |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 1158 | If you want to store the time of the original image creation, you should |
| 1159 | use a plain tEXt chunk with the "Creation Time" keyword. This is |
| 1160 | necessary because the "creation time" of a PNG image is somewhat vague, |
| 1161 | depending on whether you mean the PNG file, the time the image was |
| 1162 | created in a non-PNG format, a still photo from which the image was |
| 1163 | scanned, or possibly the subject matter itself. In order to facilitate |
| 1164 | machine-readable dates, it is recommended that the "Creation Time" |
| 1165 | tEXt chunk use RFC 1123 format dates (eg 22 May 1997 18:07:10 GMT"), |
| 1166 | although this isn't a requirement. Unlike the tIME chunk, the |
| 1167 | "Creation Time" tEXt chunk is not expected to be automatically changed |
| 1168 | by the software. To facilitate the use of RFC 1123 dates, a function |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1169 | png_convert_to_rfc1123(png_timep) is provided to convert from PNG |
| 1170 | time to an RFC 1123 format string. |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 1171 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1172 | You are now ready to write all the file information up to the actual |
| 1173 | image data. You do this with a call to png_write_info(). |
| 1174 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1175 | png_write_info(png_ptr, info_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1176 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1177 | After you've written the file information, you can set up the library |
| 1178 | to handle any special transformations of the image data. The various |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1179 | ways to transform the data will be described in the order that they |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1180 | should occur. This is important, as some of these change the color |
| 1181 | type and/or bit depth of the data, and some others only work on |
| 1182 | certain color types and bit depths. Even though each transformation |
| 1183 | checks to see if it has data that it can do somthing with, you should |
| 1184 | make sure to only enable a transformation if it will be valid for the |
| 1185 | data. For example, don't swap red and blue on grayscale data. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1186 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1187 | PNG files store RGB pixels packed into 3 bytes. This code tells |
| 1188 | the library to expect input data with 4 bytes per pixel |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1189 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1190 | png_set_filler(png_ptr, 0, PNG_FILLER_BEFORE); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1191 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1192 | where the 0 is the value that will be put in the 4th byte, and the |
| 1193 | location is either PNG_FILLER_BEFORE or PNG_FILLER_AFTER, depending |
| 1194 | upon whether the filler byte is stored XRGB or RGBX. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1195 | |
| 1196 | PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as |
| 1197 | they can, resulting in, for example, 8 pixels per byte for 1 bit files. |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1198 | If the data is supplied at 1 pixel per byte, use this code, which will |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1199 | correctly pack the pixels into a single byte: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1200 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1201 | png_set_packing(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1202 | |
| 1203 | PNG files reduce possible bit depths to 1, 2, 4, 8, and 16. If your |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1204 | data is of another bit depth, you can write an sBIT chunk into the |
| 1205 | file so that decoders can get the original data if desired. |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1206 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1207 | /* Set the true bit depth of the image data */ |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1208 | if (color_type & PNG_COLOR_MASK_COLOR) |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1209 | { |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1210 | sig_bit.red = true_bit_depth; |
| 1211 | sig_bit.green = true_bit_depth; |
| 1212 | sig_bit.blue = true_bit_depth; |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1213 | } |
| 1214 | else |
| 1215 | { |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1216 | sig_bit.gray = true_bit_depth; |
| 1217 | } |
| 1218 | if (color_type & PNG_COLOR_MASK_ALPHA) |
| 1219 | { |
| 1220 | sig_bit.alpha = true_bit_depth; |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1221 | } |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1222 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1223 | png_set_sBIT(png_ptr, info_ptr, &sig_bit); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1224 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1225 | If the data is stored in the row buffer in a bit depth other than |
| 1226 | one supported by PNG (ie 3 bit data in the range 0-7 for a 4-bit PNG), |
| 1227 | this will scale the values to appear to be the correct bit depth as |
| 1228 | is required by PNG. |
| 1229 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1230 | png_set_shift(png_ptr, &sig_bit); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1231 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1232 | PNG files store 16 bit pixels in network byte order (big-endian, |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1233 | ie. most significant bits first). This code would be used if they are |
| 1234 | supplied the other way (little-endian, ie. least significant bits |
| 1235 | first, eg. the way PCs store them): |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1236 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1237 | if (bit_depth > 8) |
| 1238 | png_set_swap(png_ptr); |
| 1239 | |
| 1240 | If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you |
| 1241 | need to change the order the pixels are packed into bytes, you can use: |
| 1242 | |
| 1243 | if (bit_depth < 8) |
| 1244 | png_set_packswap(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1245 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1246 | PNG files store 3 color pixels in red, green, blue order. This code |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1247 | would be used if they are supplied as blue, green, red: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1248 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1249 | png_set_bgr(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1250 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1251 | PNG files describe monochrome as black being zero and white being |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1252 | one. This code would be used if the pixels are supplied with this reversed |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1253 | (black being one and white being zero): |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1254 | |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1255 | png_set_invert_mono(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1256 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1257 | It is possible to have libpng flush any pending output, either manually, |
| 1258 | or automatically after a certain number of lines have been written. To |
| 1259 | flush the output stream a single time call: |
| 1260 | |
| 1261 | png_write_flush(png_ptr); |
| 1262 | |
| 1263 | and to have libpng flush the output stream periodically after a certain |
| 1264 | number of scanlines have been written, call: |
| 1265 | |
| 1266 | png_set_flush(png_ptr, nrows); |
| 1267 | |
| 1268 | Note that the distance between rows is from the last time png_write_flush() |
| 1269 | was called, or the first row of the image if it has never been called. |
| 1270 | So if you write 50 lines, and then png_set_flush 25, it will flush the |
| 1271 | output on the next scanline, and every 25 lines thereafter, unless |
| 1272 | png_write_flush()ls is called before 25 more lines have been written. |
| 1273 | If nrows is too small (less than about 10 lines for a 640 pixel wide |
| 1274 | RGB image) the image compression may decrease noticably (although this |
| 1275 | may be acceptable for real-time applications). Infrequent flushing will |
| 1276 | only degrade the compression performance by a few percent over images |
| 1277 | that do not use flushing. |
| 1278 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1279 | That's it for the transformations. Now you can write the image data. |
| 1280 | The simplest way to do this is in one function call. If have the |
| 1281 | whole image in memory, you can just call png_write_image() and libpng |
| 1282 | will write the image. You will need to pass in an array of pointers to |
| 1283 | each row. This function automatically handles interlacing, so you don't |
| 1284 | need to call png_set_interlace_handling() or call this function multiple |
| 1285 | times, or any of that other stuff necessary with png_write_rows(). |
| 1286 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1287 | png_write_image(png_ptr, row_pointers); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1288 | |
| 1289 | where row_pointers is: |
| 1290 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1291 | png_bytef *row_pointers[height]; |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1292 | |
| 1293 | You can point to void or char or whatever you use for pixels. |
| 1294 | |
| 1295 | If you can't want to write the whole image at once, you can |
| 1296 | use png_write_rows() instead. If the file is not interlaced, |
| 1297 | this is simple: |
| 1298 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1299 | png_write_rows(png_ptr, row_pointers, number_of_rows); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1300 | |
| 1301 | row_pointers is the same as in the png_write_image() call. |
| 1302 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1303 | If you are just writing one row at a time, you can do this with |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1304 | row_pointers: |
| 1305 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1306 | png_bytep row_pointer = row; |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1307 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1308 | png_write_row(png_ptr, &row_pointer); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1309 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1310 | When the file is interlaced, things can get a good deal more |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1311 | complicated. The only currently (as of 6/96 -- PNG Specification |
| 1312 | version 1.0) defined interlacing scheme for PNG files is a |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1313 | compilcated interlace scheme, known as Adam7, that breaks down an |
| 1314 | image into seven smaller images of varying size. libpng will build |
| 1315 | these images for you, or you can do them yourself. If you want to |
| 1316 | build them yourself, see the PNG specification for details of which |
| 1317 | pixels to write when. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1318 | |
| 1319 | If you don't want libpng to handle the interlacing details, just |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1320 | use png_set_interlace_handling() and call png_write_rows() the |
| 1321 | correct number of times to write all seven sub-images. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1322 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1323 | If you want libpng to build the sub-images, call this before you start |
| 1324 | writing any rows: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1325 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1326 | number_of_passes = png_set_interlace_handling(png_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1327 | |
| 1328 | This will return the number of passes needed. Currently, this |
| 1329 | is seven, but may change if another interlace type is added. |
| 1330 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1331 | Then write the complete image number_of_passes times. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1332 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1333 | png_write_rows(png_ptr, row_pointers, number_of_rows); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1334 | |
| 1335 | As some of these rows are not used, and thus return immediately, |
| 1336 | you may want to read about interlacing in the PNG specification, |
| 1337 | and only update the rows that are actually used. |
| 1338 | |
| 1339 | After you are finished writing the image, you should finish writing |
| 1340 | the file. If you are interested in writing comments or time, you should |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1341 | pass an appropriately filled png_info pointer. If you are not interested, |
| 1342 | you can pass NULL. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1343 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1344 | png_write_end(png_ptr, info_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1345 | |
| 1346 | When you are done, you can free all memory used by libpng like this: |
| 1347 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1348 | png_destroy_write_struct(&png_ptr, &info_ptr); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1349 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1350 | You must free any data you allocated for info_ptr, such as comments, |
| 1351 | palette, or histogram, before the call to png_destroy_write_struct(); |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1352 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1353 | For a more compact example of writing a PNG image, see the file example.c. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1354 | |
| 1355 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1356 | V. Modifying/Customizing libpng: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1357 | |
| 1358 | There are two issues here. The first is changing how libpng does |
| 1359 | standard things like memory allocation, input/output, and error handling. |
| 1360 | The second deals with more complicated things like adding new chunks, |
| 1361 | adding new transformations, and generally changing how libpng works. |
| 1362 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1363 | All of the memory allocation, input/output, and error handling in libpng |
| 1364 | goes through callbacks which are user setable. The default routines are |
| 1365 | in pngmem.c, pngrio.c, pngwio.c, and pngerror.c respectively. To change |
| 1366 | these functions, call the approprate png_set_???_fn() function. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1367 | |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1368 | Memory allocation is done through the functions png_large_malloc(), |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1369 | png_malloc(), png_realloc(), png_large_free(), and png_free(). These |
| 1370 | currently just call the standard C functions. The large functions must |
| 1371 | handle exactly 64K, but they don't have to handle more than that. If |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1372 | your pointers can't access more then 64K at a time, you will want to set |
| 1373 | MAXSEG_64K in zlib.h. Since it is unlikely that the method of handling |
| 1374 | memory allocation on a platform will change between applications, these |
| 1375 | functions must be modified in the library at compile time. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1376 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1377 | Input/Output in libpng is done through png_read() and png_write(), |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1378 | which currently just call fread() and fwrite(). The FILE * is stored in |
| 1379 | png_struct and is initialized via png_init_io(). If you wish to change |
| 1380 | the method of I/O, the library supplies callbacks that you can set |
| 1381 | through the function png_set_read_fn() and png_set_write_fn() at run |
| 1382 | time, instead of calling the png_init_io() function. These functions |
| 1383 | also provide a void pointer that can be retrieved via the function |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1384 | png_get_io_ptr(). For example: |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1385 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1386 | png_set_read_fn(png_structp png_ptr, voidp io_ptr, |
| 1387 | png_rw_ptr read_data_fn) |
Guy Schalnat | 0f71645 | 1995-11-28 11:22:13 -0600 | [diff] [blame] | 1388 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1389 | png_set_write_fn(png_structp png_ptr, voidp io_ptr, |
| 1390 | png_rw_ptr write_data_fn, png_flush_ptr output_flush_fn); |
Guy Schalnat | 0f71645 | 1995-11-28 11:22:13 -0600 | [diff] [blame] | 1391 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1392 | voidp io_ptr = png_get_io_ptr(png_ptr); |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1393 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1394 | The replacement I/O functions should have prototypes as follows: |
| 1395 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1396 | void user_read_data(png_structp png_ptr, png_bytep data, |
| 1397 | png_uint_32 length); |
| 1398 | void user_write_data(png_structp png_ptr, png_bytep data, |
| 1399 | png_uint_32 length); |
| 1400 | void user_flush_data(png_structp png_ptr); |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1401 | |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1402 | Supplying NULL for the read, write, or flush functions sets them back |
| 1403 | to using the default C stream functions. It is an error to read from |
| 1404 | a write stream, and vice versa. |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1405 | |
| 1406 | Error handling in libpng is done through png_error() and png_warning(). |
| 1407 | Errors handled through png_error() are fatal, meaning that png_error() |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1408 | should never return to its caller. Currently, this is handled via |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1409 | setjmp() and longjmp(), but you could change this to do things like |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1410 | exit() if you should wish. On non-fatal errors, png_warning() is called |
| 1411 | to print a warning message, and then control returns to the calling code. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1412 | By default png_error() and png_warning() print a message on stderr via |
| 1413 | fprintf() unless the library is compiled with PNG_NO_STDIO defined. If |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1414 | you wish to change the behavior of the error functions, you will need to |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1415 | set up your own message callbacks. These functions are normally supplied |
| 1416 | at the time that the png_struct is created. It is also possible to change |
| 1417 | these functions after png_create_???_struct() has been called by calling: |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1418 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1419 | png_set_error_fn(png_structp png_ptr, png_voidp error_ptr, |
| 1420 | png_error_ptr error_fn, png_error_ptr warning_fn); |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1421 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1422 | png_voidp error_ptr = png_get_error_ptr(png_ptr); |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1423 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1424 | If NULL is supplied for either error_fn or warning_fn, then the libpng |
| 1425 | default function will be used, calling fprintf() and/or longjmp() if a |
| 1426 | problem is encountered. The replacement error functions should have |
| 1427 | parameters as follows: |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1428 | |
Glenn Randers-Pehrson | b6ce43d | 1998-01-01 07:13:13 -0600 | [diff] [blame] | 1429 | void user_error_fn(png_structp png_ptr, png_const_charp error_msg); |
| 1430 | void user_warning_fn(png_structp png_ptr, png_const_charp warning_msg); |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1431 | |
| 1432 | The motivation behind using setjmp() and longjmp() is the C++ throw and |
| 1433 | catch exception handling methods. This makes the code much easier to write, |
| 1434 | as there is no need to check every return code of every function call. |
| 1435 | However, there are some uncertainties about the status of local variables |
| 1436 | after a longjmp, so the user may want to be careful about doing anything after |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1437 | setjmp returns non-zero besides returning itself. Consult your compiler |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1438 | documentation for more details. |
Guy Schalnat | 0f71645 | 1995-11-28 11:22:13 -0600 | [diff] [blame] | 1439 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1440 | If you need to read or write custom chunks, you will need to get deeper |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1441 | into the libpng code, as a mechanism has not yet been supplied for user |
| 1442 | callbacks with custom chunks. First, read the PNG specification, and have |
| 1443 | a first level of understanding of how it works. Pay particular attention |
| 1444 | to the sections that describe chunk names, and look at how other chunks |
| 1445 | were designed, so you can do things similarly. Second, check out the |
| 1446 | sections of libpng that read and write chunks. Try to find a chunk that |
| 1447 | is similar to yours and copy off of it. More details can be found in the |
| 1448 | comments inside the code. A way of handling unknown chunks in a generic |
| 1449 | method, potentially via callback functions, would be best. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1450 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1451 | If you wish to write your own transformation for the data, look through |
| 1452 | the part of the code that does the transformations, and check out some of |
| 1453 | the simpler ones to get an idea of how they work. Try to find a similar |
| 1454 | transformation to the one you want to add and copy off of it. More details |
| 1455 | can be found in the comments inside the code itself. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1456 | |
| 1457 | Configuring for 16 bit platforms: |
| 1458 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1459 | You may need to change the png_large_malloc() and png_large_free() |
| 1460 | routines in pngmem.c, as these are requred to allocate 64K, although |
| 1461 | there is already support for many of the common DOS compilers. Also, |
| 1462 | you will want to look into zconf.h to tell zlib (and thus libpng) that |
| 1463 | it cannot allocate more then 64K at a time. Even if you can, the memory |
| 1464 | won't be accessable. So limit zlib and libpng to 64K by defining MAXSEG_64K. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1465 | |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1466 | Configuring for DOS: |
| 1467 | |
| 1468 | For DOS users which only have access to the lower 640K, you will |
| 1469 | have to limit zlib's memory usage via a png_set_compression_mem_level() |
| 1470 | call. See zlib.h or zconf.h in the zlib library for more information. |
| 1471 | |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1472 | Configuring for Medium Model: |
| 1473 | |
| 1474 | Libpng's support for medium model has been tested on most of the popular |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1475 | compilers. Make sure MAXSEG_64K gets defined, USE_FAR_KEYWORD gets |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1476 | defined, and FAR gets defined to far in pngconf.h, and you should be |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1477 | all set. Everything in the library (except for zlib's structure) is |
| 1478 | expecting far data. You must use the typedefs with the p or pp on |
| 1479 | the end for pointers (or at least look at them and be careful). Make |
| 1480 | note that the row's of data are defined as png_bytepp which is a |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1481 | unsigned char far * far *. |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1482 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1483 | Configuring for gui/windowing platforms: |
| 1484 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1485 | You will need to write new error and warning functions that use the GUI |
| 1486 | interface, as described previously, and set them to be the error and |
| 1487 | warning functions at the time that png_create_???_struct() is called, |
| 1488 | in order to have them available during the structure initialization. |
| 1489 | They can be changed later via png_set_error_fn(). On some compliers, |
| 1490 | you may also have to change the memory allocators (png_malloc, etc.). |
Guy Schalnat | 6d76471 | 1995-12-19 03:22:19 -0600 | [diff] [blame] | 1491 | |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1492 | Configuring for compiler xxx: |
| 1493 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1494 | All includes for libpng are in pngconf.h. If you need to add/change/delete |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1495 | an include, this is the place to do it. The includes that are not |
| 1496 | needed outside libpng are protected by the PNG_INTERNAL definition, |
| 1497 | which is only defined for those routines inside libpng itself. The |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1498 | files in libpng proper only include png.h, which includes pngconf.h. |
| 1499 | |
| 1500 | Configuring zlib: |
| 1501 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1502 | There are special functions to configure the compression. Perhaps the |
| 1503 | most useful one changes the compression level, which currently uses |
| 1504 | input compression values in the range 0 - 9. The library normally |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1505 | uses the default compression level (Z_DEFAULT_COMPRESSION = 6). Tests |
| 1506 | have shown that for a large majority of images, compression values in |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1507 | the range 3-6 compress nearly as well as higher levels, and do so much |
| 1508 | faster. For online applications it may be desirable to have maximum speed |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1509 | (Z_BEST_SPEED = 1). With versions of zlib after v0.99, you can also |
| 1510 | specify no compression (Z_NO_COMPRESSION = 0), but this would create |
| 1511 | files larger than just storing the raw bitmap. You can specify the |
| 1512 | compression level by calling: |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1513 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1514 | png_set_compression_level(png_ptr, level); |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1515 | |
| 1516 | Another useful one is to reduce the memory level used by the library. |
| 1517 | The memory level defaults to 8, but it can be lowered if you are |
| 1518 | short on memory (running DOS, for example, where you only have 640K). |
| 1519 | |
Guy Schalnat | b2e01bd | 1996-01-26 01:38:47 -0600 | [diff] [blame] | 1520 | png_set_compression_mem_level(png_ptr, level); |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1521 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1522 | The other functions are for configuring zlib. They are not recommended |
| 1523 | for normal use and may result in writing an invalid PNG file. See |
| 1524 | zlib.h for more information on what these mean. |
| 1525 | |
| 1526 | png_set_compression_strategy(png_ptr, strategy); |
| 1527 | png_set_compression_window_bits(png_ptr, window_bits); |
| 1528 | png_set_compression_method(png_ptr, method); |
| 1529 | |
| 1530 | Controlling row filtering: |
| 1531 | |
| 1532 | If you want to control whether libpng uses filtering or not, which |
| 1533 | filters are used, and how it goes about picking row filters, you |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1534 | can call one of these functions. The selection and configuration |
| 1535 | of row filters can have a significant impact on the size and |
| 1536 | encoding speed and a somewhat lesser impact on the decoding speed |
| 1537 | of an image. Filtering is enabled by default for RGB and grayscale |
| 1538 | images (with and without alpha), and for 8-bit paletted images, but |
| 1539 | not for paletted images with bit depths less than 8 bits/pixel. |
| 1540 | |
| 1541 | The 'method' parameter sets the main filtering method, which is |
| 1542 | currently only '0' in the PNG 1.0 specification. The 'filters' |
| 1543 | parameter sets which filter(s), if any, should be used for each |
| 1544 | scanline. Possible values are PNG_ALL_FILTERS and PNG_NO_FILTERS |
| 1545 | to turn filtering on and off, respectively. |
| 1546 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1547 | Individual filter types are PNG_FILTER_NONE, PNG_FILTER_SUB, |
| 1548 | PNG_FILTER_UP, PNG_FILTER_AVG, PNG_FILTER_PAETH, which can be bitwise |
| 1549 | ORed together '|' to specify one or more filters to use. These |
| 1550 | filters are described in more detail in the PNG specification. If |
| 1551 | you intend to change the filter type during the course of writing |
| 1552 | the image, you should start with flags set for all of the filters |
| 1553 | you intend to use so that libpng can initialize its internal |
| 1554 | structures appropriately for all of the filter types. |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1555 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1556 | filters = PNG_FILTER_NONE | PNG_FILTER_SUB | PNG_FILTER_UP; |
| 1557 | png_set_filter(png_ptr, PNG_FILTER_TYPE_BASE, filters); |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1558 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1559 | It is also possible to influence how libpng chooses from among the |
| 1560 | available filters. This is done in two ways - by telling it how |
| 1561 | important it is to keep the same filter for successive rows, and |
| 1562 | by telling it the relative computational costs of the filters. |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1563 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1564 | double weights[3] = {1.5, 1.3, 1.1}, |
| 1565 | costs[PNG_FILTER_VALUE_LAST] = {1.0, 1.3, 1.3, 1.5, 1.7}; |
Guy Schalnat | 4ee97b0 | 1996-01-16 01:51:56 -0600 | [diff] [blame] | 1566 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1567 | png_set_filter_selection(png_ptr, PNG_FILTER_SELECTION_WEIGHTED, |
| 1568 | 3, weights, costs); |
| 1569 | |
| 1570 | The weights are multiplying factors which indicate to libpng that row |
| 1571 | should be the same for successive rows unless another row filter is that |
| 1572 | many times better than the previous filter. In the above example, if |
| 1573 | the previous 3 filters were SUB, SUB, NONE, the SUB filter could have a |
| 1574 | "sum of absolute differences" 1.5 x 1.3 times higher than other filters |
| 1575 | and still be chosen, while the NONE filter could have a sum 1.1 times |
| 1576 | higher than other filters and still be chosen. Unspecified weights are |
| 1577 | taken to be 1.0, and the specified weights should probably be declining |
| 1578 | like those above in order to emphasize recent filters over older filters. |
| 1579 | |
| 1580 | The filter costs specify for each filter type a relative decoding cost |
| 1581 | to be considered when selecting row filters. This means that filters |
| 1582 | with higher costs are less likely to be chosen over filters with lower |
| 1583 | costs, unless their "sum of absolute differences" is that much smaller. |
| 1584 | The costs do not necessarily reflect the exact computational speeds of |
| 1585 | the various filters, since this would unduely influence the final image |
| 1586 | size. |
| 1587 | |
| 1588 | Note that the numbers above were invented purely for this example and |
| 1589 | are given only to help explain the function usage. Little testing has |
| 1590 | been done to find optimum values for either the costs or the weights. |
Guy Schalnat | 0d58058 | 1995-07-20 02:43:20 -0500 | [diff] [blame] | 1591 | |
Guy Schalnat | 51f0eb4 | 1995-09-26 05:22:39 -0500 | [diff] [blame] | 1592 | Removing unwanted object code: |
| 1593 | |
Guy Schalnat | 69b1448 | 1996-01-10 02:56:49 -0600 | [diff] [blame] | 1594 | There are a bunch of #define's in pngconf.h that control what parts of |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1595 | libpng are compiled. All the defines end in _SUPPORTED. If you are |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1596 | never going to use an ability, you can change the #define to #undef |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1597 | before recompiling libpng and save yourself code and data space. |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1598 | You can also turn them off en masse with a compiler directive that |
| 1599 | defines PNG_READ[or WRITE]_NOT_FULLY_SUPPORTED, or |
| 1600 | PNG_READ[or WRITE]_ANCILLARY_CHUNKS_NOT_SUPPORTED, or all four, |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1601 | along with directives to turn on any of the capabilities that you do |
Glenn Randers-Pehrson | 46f61e2 | 1998-01-30 21:45:12 -0600 | [diff] [blame^] | 1602 | want. The NOT_FULLY_SUPPORTED directives disable the extra |
| 1603 | transformations but still leave the library fully capable of reading |
| 1604 | and writing PNG files with all known public chunks [except for sPLT]. |
| 1605 | Use of the PNG_READ[or WRITE]_ANCILLARY_CHUNKS_NOT_SUPPORTED directive |
| 1606 | produces a library that is incapable of reading or writing ancillary chunks. |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1607 | |
| 1608 | All the reading and writing specific code are in separate files, so the |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1609 | linker should only grab the files it needs. However, if you want to |
| 1610 | make sure, or if you are building a stand alone library, all the |
| 1611 | reading files start with pngr and all the writing files start with |
| 1612 | pngw. The files that don't match either (like png.c, pngtrans.c, etc.) |
| 1613 | are used for both reading and writing, and always need to be included. |
| 1614 | The progressive reader is in pngpread.c |
| 1615 | |
| 1616 | If you are creating or distributing a dynamically linked library (a .so |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1617 | or DLL file), you should not remove or disable any parts of the library, |
| 1618 | as this will cause applications linked with different versions of the |
| 1619 | library to fail if they call functions not available in your library. |
| 1620 | The size of the library itself should not be an issue, because only |
| 1621 | those sections which are actually used will be loaded into memory. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1622 | |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1623 | |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1624 | Changes to Libpng from version 0.88 |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1625 | |
Glenn Randers-Pehrson | c4a2ae6 | 1998-01-16 22:06:18 -0600 | [diff] [blame] | 1626 | It should be noted that versions of libpng later than 0.96 are not |
| 1627 | distributed by the original libpng author, Guy Schalnat, nor by |
| 1628 | Andreas Dilger, who had taken over from Guy during 1996 and 1997, and |
| 1629 | distributed versions 0.89 through 0.96, but rather by another member |
| 1630 | of the original PNG Group, Glenn Randers-Pehrson. Guy and Andreas are |
| 1631 | still alive and well, but they have moved on to other things. |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1632 | |
| 1633 | The old libpng functions png_read_init(), png_write_init(), |
| 1634 | png_info_init(), png_read_destroy(), and png_write_destory() have been |
| 1635 | moved to PNG_INTERNAL in version 0.95 to discourage their use. The |
| 1636 | preferred method of creating and initializing the libpng structures is |
| 1637 | via the png_create_read_struct(), png_create_write_struct(), and |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1638 | png_create_info_struct() because they isolate the size of the structures |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1639 | from the application, allow version error checking, and also allow the |
| 1640 | use of custom error handling routines during the initialization, which |
| 1641 | the old functions do not. The functions png_read_destroy() and |
| 1642 | png_write_destroy() do not actually free the memory that libpng |
| 1643 | allocated for these structs, but just reset the data structures, so they |
| 1644 | can be used instead of png_destroy_read_struct() and |
| 1645 | png_destroy_write_struct() if you feel there is too much system overhead |
| 1646 | allocating and freeing the png_struct for each image read. |
Guy Schalnat | e5a3779 | 1996-06-05 15:50:50 -0500 | [diff] [blame] | 1647 | |
| 1648 | Setting the error callbacks via png_set_message_fn() before |
| 1649 | png_read_init() as was suggested in libpng-0.88 is no longer supported |
| 1650 | because this caused applications which do not use custom error functions |
| 1651 | to fail if the png_ptr was not initialized to zero. It is still possible |
| 1652 | to set the error callbacks AFTER png_read_init(), or to change them with |
| 1653 | png_set_error_fn(), which is essentially the same function, but with a |
Andreas Dilger | 47a0c42 | 1997-05-16 02:46:07 -0500 | [diff] [blame] | 1654 | new name to force compilation errors with applications that try to use |
| 1655 | the old method. |