diff options
Diffstat (limited to 'libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt')
-rw-r--r-- | libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt | 6170 |
1 files changed, 3085 insertions, 3085 deletions
diff --git a/libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt b/libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt index 98394c8..9702669 100644 --- a/libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt +++ b/libraries/irrlicht-1.8/source/Irrlicht/jpeglib/libjpeg.txt | |||
@@ -1,3085 +1,3085 @@ | |||
1 | USING THE IJG JPEG LIBRARY | 1 | USING THE IJG JPEG LIBRARY |
2 | 2 | ||
3 | Copyright (C) 1994-2011, Thomas G. Lane, Guido Vollbeding. | 3 | Copyright (C) 1994-2011, Thomas G. Lane, Guido Vollbeding. |
4 | This file is part of the Independent JPEG Group's software. | 4 | This file is part of the Independent JPEG Group's software. |
5 | For conditions of distribution and use, see the accompanying README file. | 5 | For conditions of distribution and use, see the accompanying README file. |
6 | 6 | ||
7 | 7 | ||
8 | This file describes how to use the IJG JPEG library within an application | 8 | This file describes how to use the IJG JPEG library within an application |
9 | program. Read it if you want to write a program that uses the library. | 9 | program. Read it if you want to write a program that uses the library. |
10 | 10 | ||
11 | The file example.c provides heavily commented skeleton code for calling the | 11 | The file example.c provides heavily commented skeleton code for calling the |
12 | JPEG library. Also see jpeglib.h (the include file to be used by application | 12 | JPEG library. Also see jpeglib.h (the include file to be used by application |
13 | programs) for full details about data structures and function parameter lists. | 13 | programs) for full details about data structures and function parameter lists. |
14 | The library source code, of course, is the ultimate reference. | 14 | The library source code, of course, is the ultimate reference. |
15 | 15 | ||
16 | Note that there have been *major* changes from the application interface | 16 | Note that there have been *major* changes from the application interface |
17 | presented by IJG version 4 and earlier versions. The old design had several | 17 | presented by IJG version 4 and earlier versions. The old design had several |
18 | inherent limitations, and it had accumulated a lot of cruft as we added | 18 | inherent limitations, and it had accumulated a lot of cruft as we added |
19 | features while trying to minimize application-interface changes. We have | 19 | features while trying to minimize application-interface changes. We have |
20 | sacrificed backward compatibility in the version 5 rewrite, but we think the | 20 | sacrificed backward compatibility in the version 5 rewrite, but we think the |
21 | improvements justify this. | 21 | improvements justify this. |
22 | 22 | ||
23 | 23 | ||
24 | TABLE OF CONTENTS | 24 | TABLE OF CONTENTS |
25 | ----------------- | 25 | ----------------- |
26 | 26 | ||
27 | Overview: | 27 | Overview: |
28 | Functions provided by the library | 28 | Functions provided by the library |
29 | Outline of typical usage | 29 | Outline of typical usage |
30 | Basic library usage: | 30 | Basic library usage: |
31 | Data formats | 31 | Data formats |
32 | Compression details | 32 | Compression details |
33 | Decompression details | 33 | Decompression details |
34 | Mechanics of usage: include files, linking, etc | 34 | Mechanics of usage: include files, linking, etc |
35 | Advanced features: | 35 | Advanced features: |
36 | Compression parameter selection | 36 | Compression parameter selection |
37 | Decompression parameter selection | 37 | Decompression parameter selection |
38 | Special color spaces | 38 | Special color spaces |
39 | Error handling | 39 | Error handling |
40 | Compressed data handling (source and destination managers) | 40 | Compressed data handling (source and destination managers) |
41 | I/O suspension | 41 | I/O suspension |
42 | Progressive JPEG support | 42 | Progressive JPEG support |
43 | Buffered-image mode | 43 | Buffered-image mode |
44 | Abbreviated datastreams and multiple images | 44 | Abbreviated datastreams and multiple images |
45 | Special markers | 45 | Special markers |
46 | Raw (downsampled) image data | 46 | Raw (downsampled) image data |
47 | Really raw data: DCT coefficients | 47 | Really raw data: DCT coefficients |
48 | Progress monitoring | 48 | Progress monitoring |
49 | Memory management | 49 | Memory management |
50 | Memory usage | 50 | Memory usage |
51 | Library compile-time options | 51 | Library compile-time options |
52 | Portability considerations | 52 | Portability considerations |
53 | Notes for MS-DOS implementors | 53 | Notes for MS-DOS implementors |
54 | 54 | ||
55 | You should read at least the overview and basic usage sections before trying | 55 | You should read at least the overview and basic usage sections before trying |
56 | to program with the library. The sections on advanced features can be read | 56 | to program with the library. The sections on advanced features can be read |
57 | if and when you need them. | 57 | if and when you need them. |
58 | 58 | ||
59 | 59 | ||
60 | OVERVIEW | 60 | OVERVIEW |
61 | ======== | 61 | ======== |
62 | 62 | ||
63 | Functions provided by the library | 63 | Functions provided by the library |
64 | --------------------------------- | 64 | --------------------------------- |
65 | 65 | ||
66 | The IJG JPEG library provides C code to read and write JPEG-compressed image | 66 | The IJG JPEG library provides C code to read and write JPEG-compressed image |
67 | files. The surrounding application program receives or supplies image data a | 67 | files. The surrounding application program receives or supplies image data a |
68 | scanline at a time, using a straightforward uncompressed image format. All | 68 | scanline at a time, using a straightforward uncompressed image format. All |
69 | details of color conversion and other preprocessing/postprocessing can be | 69 | details of color conversion and other preprocessing/postprocessing can be |
70 | handled by the library. | 70 | handled by the library. |
71 | 71 | ||
72 | The library includes a substantial amount of code that is not covered by the | 72 | The library includes a substantial amount of code that is not covered by the |
73 | JPEG standard but is necessary for typical applications of JPEG. These | 73 | JPEG standard but is necessary for typical applications of JPEG. These |
74 | functions preprocess the image before JPEG compression or postprocess it after | 74 | functions preprocess the image before JPEG compression or postprocess it after |
75 | decompression. They include colorspace conversion, downsampling/upsampling, | 75 | decompression. They include colorspace conversion, downsampling/upsampling, |
76 | and color quantization. The application indirectly selects use of this code | 76 | and color quantization. The application indirectly selects use of this code |
77 | by specifying the format in which it wishes to supply or receive image data. | 77 | by specifying the format in which it wishes to supply or receive image data. |
78 | For example, if colormapped output is requested, then the decompression | 78 | For example, if colormapped output is requested, then the decompression |
79 | library automatically invokes color quantization. | 79 | library automatically invokes color quantization. |
80 | 80 | ||
81 | A wide range of quality vs. speed tradeoffs are possible in JPEG processing, | 81 | A wide range of quality vs. speed tradeoffs are possible in JPEG processing, |
82 | and even more so in decompression postprocessing. The decompression library | 82 | and even more so in decompression postprocessing. The decompression library |
83 | provides multiple implementations that cover most of the useful tradeoffs, | 83 | provides multiple implementations that cover most of the useful tradeoffs, |
84 | ranging from very-high-quality down to fast-preview operation. On the | 84 | ranging from very-high-quality down to fast-preview operation. On the |
85 | compression side we have generally not provided low-quality choices, since | 85 | compression side we have generally not provided low-quality choices, since |
86 | compression is normally less time-critical. It should be understood that the | 86 | compression is normally less time-critical. It should be understood that the |
87 | low-quality modes may not meet the JPEG standard's accuracy requirements; | 87 | low-quality modes may not meet the JPEG standard's accuracy requirements; |
88 | nonetheless, they are useful for viewers. | 88 | nonetheless, they are useful for viewers. |
89 | 89 | ||
90 | A word about functions *not* provided by the library. We handle a subset of | 90 | A word about functions *not* provided by the library. We handle a subset of |
91 | the ISO JPEG standard; most baseline, extended-sequential, and progressive | 91 | the ISO JPEG standard; most baseline, extended-sequential, and progressive |
92 | JPEG processes are supported. (Our subset includes all features now in common | 92 | JPEG processes are supported. (Our subset includes all features now in common |
93 | use.) Unsupported ISO options include: | 93 | use.) Unsupported ISO options include: |
94 | * Hierarchical storage | 94 | * Hierarchical storage |
95 | * Lossless JPEG | 95 | * Lossless JPEG |
96 | * DNL marker | 96 | * DNL marker |
97 | * Nonintegral subsampling ratios | 97 | * Nonintegral subsampling ratios |
98 | We support both 8- and 12-bit data precision, but this is a compile-time | 98 | We support both 8- and 12-bit data precision, but this is a compile-time |
99 | choice rather than a run-time choice; hence it is difficult to use both | 99 | choice rather than a run-time choice; hence it is difficult to use both |
100 | precisions in a single application. | 100 | precisions in a single application. |
101 | 101 | ||
102 | By itself, the library handles only interchange JPEG datastreams --- in | 102 | By itself, the library handles only interchange JPEG datastreams --- in |
103 | particular the widely used JFIF file format. The library can be used by | 103 | particular the widely used JFIF file format. The library can be used by |
104 | surrounding code to process interchange or abbreviated JPEG datastreams that | 104 | surrounding code to process interchange or abbreviated JPEG datastreams that |
105 | are embedded in more complex file formats. (For example, this library is | 105 | are embedded in more complex file formats. (For example, this library is |
106 | used by the free LIBTIFF library to support JPEG compression in TIFF.) | 106 | used by the free LIBTIFF library to support JPEG compression in TIFF.) |
107 | 107 | ||
108 | 108 | ||
109 | Outline of typical usage | 109 | Outline of typical usage |
110 | ------------------------ | 110 | ------------------------ |
111 | 111 | ||
112 | The rough outline of a JPEG compression operation is: | 112 | The rough outline of a JPEG compression operation is: |
113 | 113 | ||
114 | Allocate and initialize a JPEG compression object | 114 | Allocate and initialize a JPEG compression object |
115 | Specify the destination for the compressed data (eg, a file) | 115 | Specify the destination for the compressed data (eg, a file) |
116 | Set parameters for compression, including image size & colorspace | 116 | Set parameters for compression, including image size & colorspace |
117 | jpeg_start_compress(...); | 117 | jpeg_start_compress(...); |
118 | while (scan lines remain to be written) | 118 | while (scan lines remain to be written) |
119 | jpeg_write_scanlines(...); | 119 | jpeg_write_scanlines(...); |
120 | jpeg_finish_compress(...); | 120 | jpeg_finish_compress(...); |
121 | Release the JPEG compression object | 121 | Release the JPEG compression object |
122 | 122 | ||
123 | A JPEG compression object holds parameters and working state for the JPEG | 123 | A JPEG compression object holds parameters and working state for the JPEG |
124 | library. We make creation/destruction of the object separate from starting | 124 | library. We make creation/destruction of the object separate from starting |
125 | or finishing compression of an image; the same object can be re-used for a | 125 | or finishing compression of an image; the same object can be re-used for a |
126 | series of image compression operations. This makes it easy to re-use the | 126 | series of image compression operations. This makes it easy to re-use the |
127 | same parameter settings for a sequence of images. Re-use of a JPEG object | 127 | same parameter settings for a sequence of images. Re-use of a JPEG object |
128 | also has important implications for processing abbreviated JPEG datastreams, | 128 | also has important implications for processing abbreviated JPEG datastreams, |
129 | as discussed later. | 129 | as discussed later. |
130 | 130 | ||
131 | The image data to be compressed is supplied to jpeg_write_scanlines() from | 131 | The image data to be compressed is supplied to jpeg_write_scanlines() from |
132 | in-memory buffers. If the application is doing file-to-file compression, | 132 | in-memory buffers. If the application is doing file-to-file compression, |
133 | reading image data from the source file is the application's responsibility. | 133 | reading image data from the source file is the application's responsibility. |
134 | The library emits compressed data by calling a "data destination manager", | 134 | The library emits compressed data by calling a "data destination manager", |
135 | which typically will write the data into a file; but the application can | 135 | which typically will write the data into a file; but the application can |
136 | provide its own destination manager to do something else. | 136 | provide its own destination manager to do something else. |
137 | 137 | ||
138 | Similarly, the rough outline of a JPEG decompression operation is: | 138 | Similarly, the rough outline of a JPEG decompression operation is: |
139 | 139 | ||
140 | Allocate and initialize a JPEG decompression object | 140 | Allocate and initialize a JPEG decompression object |
141 | Specify the source of the compressed data (eg, a file) | 141 | Specify the source of the compressed data (eg, a file) |
142 | Call jpeg_read_header() to obtain image info | 142 | Call jpeg_read_header() to obtain image info |
143 | Set parameters for decompression | 143 | Set parameters for decompression |
144 | jpeg_start_decompress(...); | 144 | jpeg_start_decompress(...); |
145 | while (scan lines remain to be read) | 145 | while (scan lines remain to be read) |
146 | jpeg_read_scanlines(...); | 146 | jpeg_read_scanlines(...); |
147 | jpeg_finish_decompress(...); | 147 | jpeg_finish_decompress(...); |
148 | Release the JPEG decompression object | 148 | Release the JPEG decompression object |
149 | 149 | ||
150 | This is comparable to the compression outline except that reading the | 150 | This is comparable to the compression outline except that reading the |
151 | datastream header is a separate step. This is helpful because information | 151 | datastream header is a separate step. This is helpful because information |
152 | about the image's size, colorspace, etc is available when the application | 152 | about the image's size, colorspace, etc is available when the application |
153 | selects decompression parameters. For example, the application can choose an | 153 | selects decompression parameters. For example, the application can choose an |
154 | output scaling ratio that will fit the image into the available screen size. | 154 | output scaling ratio that will fit the image into the available screen size. |
155 | 155 | ||
156 | The decompression library obtains compressed data by calling a data source | 156 | The decompression library obtains compressed data by calling a data source |
157 | manager, which typically will read the data from a file; but other behaviors | 157 | manager, which typically will read the data from a file; but other behaviors |
158 | can be obtained with a custom source manager. Decompressed data is delivered | 158 | can be obtained with a custom source manager. Decompressed data is delivered |
159 | into in-memory buffers passed to jpeg_read_scanlines(). | 159 | into in-memory buffers passed to jpeg_read_scanlines(). |
160 | 160 | ||
161 | It is possible to abort an incomplete compression or decompression operation | 161 | It is possible to abort an incomplete compression or decompression operation |
162 | by calling jpeg_abort(); or, if you do not need to retain the JPEG object, | 162 | by calling jpeg_abort(); or, if you do not need to retain the JPEG object, |
163 | simply release it by calling jpeg_destroy(). | 163 | simply release it by calling jpeg_destroy(). |
164 | 164 | ||
165 | JPEG compression and decompression objects are two separate struct types. | 165 | JPEG compression and decompression objects are two separate struct types. |
166 | However, they share some common fields, and certain routines such as | 166 | However, they share some common fields, and certain routines such as |
167 | jpeg_destroy() can work on either type of object. | 167 | jpeg_destroy() can work on either type of object. |
168 | 168 | ||
169 | The JPEG library has no static variables: all state is in the compression | 169 | The JPEG library has no static variables: all state is in the compression |
170 | or decompression object. Therefore it is possible to process multiple | 170 | or decompression object. Therefore it is possible to process multiple |
171 | compression and decompression operations concurrently, using multiple JPEG | 171 | compression and decompression operations concurrently, using multiple JPEG |
172 | objects. | 172 | objects. |
173 | 173 | ||
174 | Both compression and decompression can be done in an incremental memory-to- | 174 | Both compression and decompression can be done in an incremental memory-to- |
175 | memory fashion, if suitable source/destination managers are used. See the | 175 | memory fashion, if suitable source/destination managers are used. See the |
176 | section on "I/O suspension" for more details. | 176 | section on "I/O suspension" for more details. |
177 | 177 | ||
178 | 178 | ||
179 | BASIC LIBRARY USAGE | 179 | BASIC LIBRARY USAGE |
180 | =================== | 180 | =================== |
181 | 181 | ||
182 | Data formats | 182 | Data formats |
183 | ------------ | 183 | ------------ |
184 | 184 | ||
185 | Before diving into procedural details, it is helpful to understand the | 185 | Before diving into procedural details, it is helpful to understand the |
186 | image data format that the JPEG library expects or returns. | 186 | image data format that the JPEG library expects or returns. |
187 | 187 | ||
188 | The standard input image format is a rectangular array of pixels, with each | 188 | The standard input image format is a rectangular array of pixels, with each |
189 | pixel having the same number of "component" or "sample" values (color | 189 | pixel having the same number of "component" or "sample" values (color |
190 | channels). You must specify how many components there are and the colorspace | 190 | channels). You must specify how many components there are and the colorspace |
191 | interpretation of the components. Most applications will use RGB data | 191 | interpretation of the components. Most applications will use RGB data |
192 | (three components per pixel) or grayscale data (one component per pixel). | 192 | (three components per pixel) or grayscale data (one component per pixel). |
193 | PLEASE NOTE THAT RGB DATA IS THREE SAMPLES PER PIXEL, GRAYSCALE ONLY ONE. | 193 | PLEASE NOTE THAT RGB DATA IS THREE SAMPLES PER PIXEL, GRAYSCALE ONLY ONE. |
194 | A remarkable number of people manage to miss this, only to find that their | 194 | A remarkable number of people manage to miss this, only to find that their |
195 | programs don't work with grayscale JPEG files. | 195 | programs don't work with grayscale JPEG files. |
196 | 196 | ||
197 | There is no provision for colormapped input. JPEG files are always full-color | 197 | There is no provision for colormapped input. JPEG files are always full-color |
198 | or full grayscale (or sometimes another colorspace such as CMYK). You can | 198 | or full grayscale (or sometimes another colorspace such as CMYK). You can |
199 | feed in a colormapped image by expanding it to full-color format. However | 199 | feed in a colormapped image by expanding it to full-color format. However |
200 | JPEG often doesn't work very well with source data that has been colormapped, | 200 | JPEG often doesn't work very well with source data that has been colormapped, |
201 | because of dithering noise. This is discussed in more detail in the JPEG FAQ | 201 | because of dithering noise. This is discussed in more detail in the JPEG FAQ |
202 | and the other references mentioned in the README file. | 202 | and the other references mentioned in the README file. |
203 | 203 | ||
204 | Pixels are stored by scanlines, with each scanline running from left to | 204 | Pixels are stored by scanlines, with each scanline running from left to |
205 | right. The component values for each pixel are adjacent in the row; for | 205 | right. The component values for each pixel are adjacent in the row; for |
206 | example, R,G,B,R,G,B,R,G,B,... for 24-bit RGB color. Each scanline is an | 206 | example, R,G,B,R,G,B,R,G,B,... for 24-bit RGB color. Each scanline is an |
207 | array of data type JSAMPLE --- which is typically "unsigned char", unless | 207 | array of data type JSAMPLE --- which is typically "unsigned char", unless |
208 | you've changed jmorecfg.h. (You can also change the RGB pixel layout, say | 208 | you've changed jmorecfg.h. (You can also change the RGB pixel layout, say |
209 | to B,G,R order, by modifying jmorecfg.h. But see the restrictions listed in | 209 | to B,G,R order, by modifying jmorecfg.h. But see the restrictions listed in |
210 | that file before doing so.) | 210 | that file before doing so.) |
211 | 211 | ||
212 | A 2-D array of pixels is formed by making a list of pointers to the starts of | 212 | A 2-D array of pixels is formed by making a list of pointers to the starts of |
213 | scanlines; so the scanlines need not be physically adjacent in memory. Even | 213 | scanlines; so the scanlines need not be physically adjacent in memory. Even |
214 | if you process just one scanline at a time, you must make a one-element | 214 | if you process just one scanline at a time, you must make a one-element |
215 | pointer array to conform to this structure. Pointers to JSAMPLE rows are of | 215 | pointer array to conform to this structure. Pointers to JSAMPLE rows are of |
216 | type JSAMPROW, and the pointer to the pointer array is of type JSAMPARRAY. | 216 | type JSAMPROW, and the pointer to the pointer array is of type JSAMPARRAY. |
217 | 217 | ||
218 | The library accepts or supplies one or more complete scanlines per call. | 218 | The library accepts or supplies one or more complete scanlines per call. |
219 | It is not possible to process part of a row at a time. Scanlines are always | 219 | It is not possible to process part of a row at a time. Scanlines are always |
220 | processed top-to-bottom. You can process an entire image in one call if you | 220 | processed top-to-bottom. You can process an entire image in one call if you |
221 | have it all in memory, but usually it's simplest to process one scanline at | 221 | have it all in memory, but usually it's simplest to process one scanline at |
222 | a time. | 222 | a time. |
223 | 223 | ||
224 | For best results, source data values should have the precision specified by | 224 | For best results, source data values should have the precision specified by |
225 | BITS_IN_JSAMPLE (normally 8 bits). For instance, if you choose to compress | 225 | BITS_IN_JSAMPLE (normally 8 bits). For instance, if you choose to compress |
226 | data that's only 6 bits/channel, you should left-justify each value in a | 226 | data that's only 6 bits/channel, you should left-justify each value in a |
227 | byte before passing it to the compressor. If you need to compress data | 227 | byte before passing it to the compressor. If you need to compress data |
228 | that has more than 8 bits/channel, compile with BITS_IN_JSAMPLE = 12. | 228 | that has more than 8 bits/channel, compile with BITS_IN_JSAMPLE = 12. |
229 | (See "Library compile-time options", later.) | 229 | (See "Library compile-time options", later.) |
230 | 230 | ||
231 | 231 | ||
232 | The data format returned by the decompressor is the same in all details, | 232 | The data format returned by the decompressor is the same in all details, |
233 | except that colormapped output is supported. (Again, a JPEG file is never | 233 | except that colormapped output is supported. (Again, a JPEG file is never |
234 | colormapped. But you can ask the decompressor to perform on-the-fly color | 234 | colormapped. But you can ask the decompressor to perform on-the-fly color |
235 | quantization to deliver colormapped output.) If you request colormapped | 235 | quantization to deliver colormapped output.) If you request colormapped |
236 | output then the returned data array contains a single JSAMPLE per pixel; | 236 | output then the returned data array contains a single JSAMPLE per pixel; |
237 | its value is an index into a color map. The color map is represented as | 237 | its value is an index into a color map. The color map is represented as |
238 | a 2-D JSAMPARRAY in which each row holds the values of one color component, | 238 | a 2-D JSAMPARRAY in which each row holds the values of one color component, |
239 | that is, colormap[i][j] is the value of the i'th color component for pixel | 239 | that is, colormap[i][j] is the value of the i'th color component for pixel |
240 | value (map index) j. Note that since the colormap indexes are stored in | 240 | value (map index) j. Note that since the colormap indexes are stored in |
241 | JSAMPLEs, the maximum number of colors is limited by the size of JSAMPLE | 241 | JSAMPLEs, the maximum number of colors is limited by the size of JSAMPLE |
242 | (ie, at most 256 colors for an 8-bit JPEG library). | 242 | (ie, at most 256 colors for an 8-bit JPEG library). |
243 | 243 | ||
244 | 244 | ||
245 | Compression details | 245 | Compression details |
246 | ------------------- | 246 | ------------------- |
247 | 247 | ||
248 | Here we revisit the JPEG compression outline given in the overview. | 248 | Here we revisit the JPEG compression outline given in the overview. |
249 | 249 | ||
250 | 1. Allocate and initialize a JPEG compression object. | 250 | 1. Allocate and initialize a JPEG compression object. |
251 | 251 | ||
252 | A JPEG compression object is a "struct jpeg_compress_struct". (It also has | 252 | A JPEG compression object is a "struct jpeg_compress_struct". (It also has |
253 | a bunch of subsidiary structures which are allocated via malloc(), but the | 253 | a bunch of subsidiary structures which are allocated via malloc(), but the |
254 | application doesn't control those directly.) This struct can be just a local | 254 | application doesn't control those directly.) This struct can be just a local |
255 | variable in the calling routine, if a single routine is going to execute the | 255 | variable in the calling routine, if a single routine is going to execute the |
256 | whole JPEG compression sequence. Otherwise it can be static or allocated | 256 | whole JPEG compression sequence. Otherwise it can be static or allocated |
257 | from malloc(). | 257 | from malloc(). |
258 | 258 | ||
259 | You will also need a structure representing a JPEG error handler. The part | 259 | You will also need a structure representing a JPEG error handler. The part |
260 | of this that the library cares about is a "struct jpeg_error_mgr". If you | 260 | of this that the library cares about is a "struct jpeg_error_mgr". If you |
261 | are providing your own error handler, you'll typically want to embed the | 261 | are providing your own error handler, you'll typically want to embed the |
262 | jpeg_error_mgr struct in a larger structure; this is discussed later under | 262 | jpeg_error_mgr struct in a larger structure; this is discussed later under |
263 | "Error handling". For now we'll assume you are just using the default error | 263 | "Error handling". For now we'll assume you are just using the default error |
264 | handler. The default error handler will print JPEG error/warning messages | 264 | handler. The default error handler will print JPEG error/warning messages |
265 | on stderr, and it will call exit() if a fatal error occurs. | 265 | on stderr, and it will call exit() if a fatal error occurs. |
266 | 266 | ||
267 | You must initialize the error handler structure, store a pointer to it into | 267 | You must initialize the error handler structure, store a pointer to it into |
268 | the JPEG object's "err" field, and then call jpeg_create_compress() to | 268 | the JPEG object's "err" field, and then call jpeg_create_compress() to |
269 | initialize the rest of the JPEG object. | 269 | initialize the rest of the JPEG object. |
270 | 270 | ||
271 | Typical code for this step, if you are using the default error handler, is | 271 | Typical code for this step, if you are using the default error handler, is |
272 | 272 | ||
273 | struct jpeg_compress_struct cinfo; | 273 | struct jpeg_compress_struct cinfo; |
274 | struct jpeg_error_mgr jerr; | 274 | struct jpeg_error_mgr jerr; |
275 | ... | 275 | ... |
276 | cinfo.err = jpeg_std_error(&jerr); | 276 | cinfo.err = jpeg_std_error(&jerr); |
277 | jpeg_create_compress(&cinfo); | 277 | jpeg_create_compress(&cinfo); |
278 | 278 | ||
279 | jpeg_create_compress allocates a small amount of memory, so it could fail | 279 | jpeg_create_compress allocates a small amount of memory, so it could fail |
280 | if you are out of memory. In that case it will exit via the error handler; | 280 | if you are out of memory. In that case it will exit via the error handler; |
281 | that's why the error handler must be initialized first. | 281 | that's why the error handler must be initialized first. |
282 | 282 | ||
283 | 283 | ||
284 | 2. Specify the destination for the compressed data (eg, a file). | 284 | 2. Specify the destination for the compressed data (eg, a file). |
285 | 285 | ||
286 | As previously mentioned, the JPEG library delivers compressed data to a | 286 | As previously mentioned, the JPEG library delivers compressed data to a |
287 | "data destination" module. The library includes one data destination | 287 | "data destination" module. The library includes one data destination |
288 | module which knows how to write to a stdio stream. You can use your own | 288 | module which knows how to write to a stdio stream. You can use your own |
289 | destination module if you want to do something else, as discussed later. | 289 | destination module if you want to do something else, as discussed later. |
290 | 290 | ||
291 | If you use the standard destination module, you must open the target stdio | 291 | If you use the standard destination module, you must open the target stdio |
292 | stream beforehand. Typical code for this step looks like: | 292 | stream beforehand. Typical code for this step looks like: |
293 | 293 | ||
294 | FILE * outfile; | 294 | FILE * outfile; |
295 | ... | 295 | ... |
296 | if ((outfile = fopen(filename, "wb")) == NULL) { | 296 | if ((outfile = fopen(filename, "wb")) == NULL) { |
297 | fprintf(stderr, "can't open %s\n", filename); | 297 | fprintf(stderr, "can't open %s\n", filename); |
298 | exit(1); | 298 | exit(1); |
299 | } | 299 | } |
300 | jpeg_stdio_dest(&cinfo, outfile); | 300 | jpeg_stdio_dest(&cinfo, outfile); |
301 | 301 | ||
302 | where the last line invokes the standard destination module. | 302 | where the last line invokes the standard destination module. |
303 | 303 | ||
304 | WARNING: it is critical that the binary compressed data be delivered to the | 304 | WARNING: it is critical that the binary compressed data be delivered to the |
305 | output file unchanged. On non-Unix systems the stdio library may perform | 305 | output file unchanged. On non-Unix systems the stdio library may perform |
306 | newline translation or otherwise corrupt binary data. To suppress this | 306 | newline translation or otherwise corrupt binary data. To suppress this |
307 | behavior, you may need to use a "b" option to fopen (as shown above), or use | 307 | behavior, you may need to use a "b" option to fopen (as shown above), or use |
308 | setmode() or another routine to put the stdio stream in binary mode. See | 308 | setmode() or another routine to put the stdio stream in binary mode. See |
309 | cjpeg.c and djpeg.c for code that has been found to work on many systems. | 309 | cjpeg.c and djpeg.c for code that has been found to work on many systems. |
310 | 310 | ||
311 | You can select the data destination after setting other parameters (step 3), | 311 | You can select the data destination after setting other parameters (step 3), |
312 | if that's more convenient. You may not change the destination between | 312 | if that's more convenient. You may not change the destination between |
313 | calling jpeg_start_compress() and jpeg_finish_compress(). | 313 | calling jpeg_start_compress() and jpeg_finish_compress(). |
314 | 314 | ||
315 | 315 | ||
316 | 3. Set parameters for compression, including image size & colorspace. | 316 | 3. Set parameters for compression, including image size & colorspace. |
317 | 317 | ||
318 | You must supply information about the source image by setting the following | 318 | You must supply information about the source image by setting the following |
319 | fields in the JPEG object (cinfo structure): | 319 | fields in the JPEG object (cinfo structure): |
320 | 320 | ||
321 | image_width Width of image, in pixels | 321 | image_width Width of image, in pixels |
322 | image_height Height of image, in pixels | 322 | image_height Height of image, in pixels |
323 | input_components Number of color channels (samples per pixel) | 323 | input_components Number of color channels (samples per pixel) |
324 | in_color_space Color space of source image | 324 | in_color_space Color space of source image |
325 | 325 | ||
326 | The image dimensions are, hopefully, obvious. JPEG supports image dimensions | 326 | The image dimensions are, hopefully, obvious. JPEG supports image dimensions |
327 | of 1 to 64K pixels in either direction. The input color space is typically | 327 | of 1 to 64K pixels in either direction. The input color space is typically |
328 | RGB or grayscale, and input_components is 3 or 1 accordingly. (See "Special | 328 | RGB or grayscale, and input_components is 3 or 1 accordingly. (See "Special |
329 | color spaces", later, for more info.) The in_color_space field must be | 329 | color spaces", later, for more info.) The in_color_space field must be |
330 | assigned one of the J_COLOR_SPACE enum constants, typically JCS_RGB or | 330 | assigned one of the J_COLOR_SPACE enum constants, typically JCS_RGB or |
331 | JCS_GRAYSCALE. | 331 | JCS_GRAYSCALE. |
332 | 332 | ||
333 | JPEG has a large number of compression parameters that determine how the | 333 | JPEG has a large number of compression parameters that determine how the |
334 | image is encoded. Most applications don't need or want to know about all | 334 | image is encoded. Most applications don't need or want to know about all |
335 | these parameters. You can set all the parameters to reasonable defaults by | 335 | these parameters. You can set all the parameters to reasonable defaults by |
336 | calling jpeg_set_defaults(); then, if there are particular values you want | 336 | calling jpeg_set_defaults(); then, if there are particular values you want |
337 | to change, you can do so after that. The "Compression parameter selection" | 337 | to change, you can do so after that. The "Compression parameter selection" |
338 | section tells about all the parameters. | 338 | section tells about all the parameters. |
339 | 339 | ||
340 | You must set in_color_space correctly before calling jpeg_set_defaults(), | 340 | You must set in_color_space correctly before calling jpeg_set_defaults(), |
341 | because the defaults depend on the source image colorspace. However the | 341 | because the defaults depend on the source image colorspace. However the |
342 | other three source image parameters need not be valid until you call | 342 | other three source image parameters need not be valid until you call |
343 | jpeg_start_compress(). There's no harm in calling jpeg_set_defaults() more | 343 | jpeg_start_compress(). There's no harm in calling jpeg_set_defaults() more |
344 | than once, if that happens to be convenient. | 344 | than once, if that happens to be convenient. |
345 | 345 | ||
346 | Typical code for a 24-bit RGB source image is | 346 | Typical code for a 24-bit RGB source image is |
347 | 347 | ||
348 | cinfo.image_width = Width; /* image width and height, in pixels */ | 348 | cinfo.image_width = Width; /* image width and height, in pixels */ |
349 | cinfo.image_height = Height; | 349 | cinfo.image_height = Height; |
350 | cinfo.input_components = 3; /* # of color components per pixel */ | 350 | cinfo.input_components = 3; /* # of color components per pixel */ |
351 | cinfo.in_color_space = JCS_RGB; /* colorspace of input image */ | 351 | cinfo.in_color_space = JCS_RGB; /* colorspace of input image */ |
352 | 352 | ||
353 | jpeg_set_defaults(&cinfo); | 353 | jpeg_set_defaults(&cinfo); |
354 | /* Make optional parameter settings here */ | 354 | /* Make optional parameter settings here */ |
355 | 355 | ||
356 | 356 | ||
357 | 4. jpeg_start_compress(...); | 357 | 4. jpeg_start_compress(...); |
358 | 358 | ||
359 | After you have established the data destination and set all the necessary | 359 | After you have established the data destination and set all the necessary |
360 | source image info and other parameters, call jpeg_start_compress() to begin | 360 | source image info and other parameters, call jpeg_start_compress() to begin |
361 | a compression cycle. This will initialize internal state, allocate working | 361 | a compression cycle. This will initialize internal state, allocate working |
362 | storage, and emit the first few bytes of the JPEG datastream header. | 362 | storage, and emit the first few bytes of the JPEG datastream header. |
363 | 363 | ||
364 | Typical code: | 364 | Typical code: |
365 | 365 | ||
366 | jpeg_start_compress(&cinfo, TRUE); | 366 | jpeg_start_compress(&cinfo, TRUE); |
367 | 367 | ||
368 | The "TRUE" parameter ensures that a complete JPEG interchange datastream | 368 | The "TRUE" parameter ensures that a complete JPEG interchange datastream |
369 | will be written. This is appropriate in most cases. If you think you might | 369 | will be written. This is appropriate in most cases. If you think you might |
370 | want to use an abbreviated datastream, read the section on abbreviated | 370 | want to use an abbreviated datastream, read the section on abbreviated |
371 | datastreams, below. | 371 | datastreams, below. |
372 | 372 | ||
373 | Once you have called jpeg_start_compress(), you may not alter any JPEG | 373 | Once you have called jpeg_start_compress(), you may not alter any JPEG |
374 | parameters or other fields of the JPEG object until you have completed | 374 | parameters or other fields of the JPEG object until you have completed |
375 | the compression cycle. | 375 | the compression cycle. |
376 | 376 | ||
377 | 377 | ||
378 | 5. while (scan lines remain to be written) | 378 | 5. while (scan lines remain to be written) |
379 | jpeg_write_scanlines(...); | 379 | jpeg_write_scanlines(...); |
380 | 380 | ||
381 | Now write all the required image data by calling jpeg_write_scanlines() | 381 | Now write all the required image data by calling jpeg_write_scanlines() |
382 | one or more times. You can pass one or more scanlines in each call, up | 382 | one or more times. You can pass one or more scanlines in each call, up |
383 | to the total image height. In most applications it is convenient to pass | 383 | to the total image height. In most applications it is convenient to pass |
384 | just one or a few scanlines at a time. The expected format for the passed | 384 | just one or a few scanlines at a time. The expected format for the passed |
385 | data is discussed under "Data formats", above. | 385 | data is discussed under "Data formats", above. |
386 | 386 | ||
387 | Image data should be written in top-to-bottom scanline order. The JPEG spec | 387 | Image data should be written in top-to-bottom scanline order. The JPEG spec |
388 | contains some weasel wording about how top and bottom are application-defined | 388 | contains some weasel wording about how top and bottom are application-defined |
389 | terms (a curious interpretation of the English language...) but if you want | 389 | terms (a curious interpretation of the English language...) but if you want |
390 | your files to be compatible with everyone else's, you WILL use top-to-bottom | 390 | your files to be compatible with everyone else's, you WILL use top-to-bottom |
391 | order. If the source data must be read in bottom-to-top order, you can use | 391 | order. If the source data must be read in bottom-to-top order, you can use |
392 | the JPEG library's virtual array mechanism to invert the data efficiently. | 392 | the JPEG library's virtual array mechanism to invert the data efficiently. |
393 | Examples of this can be found in the sample application cjpeg. | 393 | Examples of this can be found in the sample application cjpeg. |
394 | 394 | ||
395 | The library maintains a count of the number of scanlines written so far | 395 | The library maintains a count of the number of scanlines written so far |
396 | in the next_scanline field of the JPEG object. Usually you can just use | 396 | in the next_scanline field of the JPEG object. Usually you can just use |
397 | this variable as the loop counter, so that the loop test looks like | 397 | this variable as the loop counter, so that the loop test looks like |
398 | "while (cinfo.next_scanline < cinfo.image_height)". | 398 | "while (cinfo.next_scanline < cinfo.image_height)". |
399 | 399 | ||
400 | Code for this step depends heavily on the way that you store the source data. | 400 | Code for this step depends heavily on the way that you store the source data. |
401 | example.c shows the following code for the case of a full-size 2-D source | 401 | example.c shows the following code for the case of a full-size 2-D source |
402 | array containing 3-byte RGB pixels: | 402 | array containing 3-byte RGB pixels: |
403 | 403 | ||
404 | JSAMPROW row_pointer[1]; /* pointer to a single row */ | 404 | JSAMPROW row_pointer[1]; /* pointer to a single row */ |
405 | int row_stride; /* physical row width in buffer */ | 405 | int row_stride; /* physical row width in buffer */ |
406 | 406 | ||
407 | row_stride = image_width * 3; /* JSAMPLEs per row in image_buffer */ | 407 | row_stride = image_width * 3; /* JSAMPLEs per row in image_buffer */ |
408 | 408 | ||
409 | while (cinfo.next_scanline < cinfo.image_height) { | 409 | while (cinfo.next_scanline < cinfo.image_height) { |
410 | row_pointer[0] = & image_buffer[cinfo.next_scanline * row_stride]; | 410 | row_pointer[0] = & image_buffer[cinfo.next_scanline * row_stride]; |
411 | jpeg_write_scanlines(&cinfo, row_pointer, 1); | 411 | jpeg_write_scanlines(&cinfo, row_pointer, 1); |
412 | } | 412 | } |
413 | 413 | ||
414 | jpeg_write_scanlines() returns the number of scanlines actually written. | 414 | jpeg_write_scanlines() returns the number of scanlines actually written. |
415 | This will normally be equal to the number passed in, so you can usually | 415 | This will normally be equal to the number passed in, so you can usually |
416 | ignore the return value. It is different in just two cases: | 416 | ignore the return value. It is different in just two cases: |
417 | * If you try to write more scanlines than the declared image height, | 417 | * If you try to write more scanlines than the declared image height, |
418 | the additional scanlines are ignored. | 418 | the additional scanlines are ignored. |
419 | * If you use a suspending data destination manager, output buffer overrun | 419 | * If you use a suspending data destination manager, output buffer overrun |
420 | will cause the compressor to return before accepting all the passed lines. | 420 | will cause the compressor to return before accepting all the passed lines. |
421 | This feature is discussed under "I/O suspension", below. The normal | 421 | This feature is discussed under "I/O suspension", below. The normal |
422 | stdio destination manager will NOT cause this to happen. | 422 | stdio destination manager will NOT cause this to happen. |
423 | In any case, the return value is the same as the change in the value of | 423 | In any case, the return value is the same as the change in the value of |
424 | next_scanline. | 424 | next_scanline. |
425 | 425 | ||
426 | 426 | ||
427 | 6. jpeg_finish_compress(...); | 427 | 6. jpeg_finish_compress(...); |
428 | 428 | ||
429 | After all the image data has been written, call jpeg_finish_compress() to | 429 | After all the image data has been written, call jpeg_finish_compress() to |
430 | complete the compression cycle. This step is ESSENTIAL to ensure that the | 430 | complete the compression cycle. This step is ESSENTIAL to ensure that the |
431 | last bufferload of data is written to the data destination. | 431 | last bufferload of data is written to the data destination. |
432 | jpeg_finish_compress() also releases working memory associated with the JPEG | 432 | jpeg_finish_compress() also releases working memory associated with the JPEG |
433 | object. | 433 | object. |
434 | 434 | ||
435 | Typical code: | 435 | Typical code: |
436 | 436 | ||
437 | jpeg_finish_compress(&cinfo); | 437 | jpeg_finish_compress(&cinfo); |
438 | 438 | ||
439 | If using the stdio destination manager, don't forget to close the output | 439 | If using the stdio destination manager, don't forget to close the output |
440 | stdio stream (if necessary) afterwards. | 440 | stdio stream (if necessary) afterwards. |
441 | 441 | ||
442 | If you have requested a multi-pass operating mode, such as Huffman code | 442 | If you have requested a multi-pass operating mode, such as Huffman code |
443 | optimization, jpeg_finish_compress() will perform the additional passes using | 443 | optimization, jpeg_finish_compress() will perform the additional passes using |
444 | data buffered by the first pass. In this case jpeg_finish_compress() may take | 444 | data buffered by the first pass. In this case jpeg_finish_compress() may take |
445 | quite a while to complete. With the default compression parameters, this will | 445 | quite a while to complete. With the default compression parameters, this will |
446 | not happen. | 446 | not happen. |
447 | 447 | ||
448 | It is an error to call jpeg_finish_compress() before writing the necessary | 448 | It is an error to call jpeg_finish_compress() before writing the necessary |
449 | total number of scanlines. If you wish to abort compression, call | 449 | total number of scanlines. If you wish to abort compression, call |
450 | jpeg_abort() as discussed below. | 450 | jpeg_abort() as discussed below. |
451 | 451 | ||
452 | After completing a compression cycle, you may dispose of the JPEG object | 452 | After completing a compression cycle, you may dispose of the JPEG object |
453 | as discussed next, or you may use it to compress another image. In that case | 453 | as discussed next, or you may use it to compress another image. In that case |
454 | return to step 2, 3, or 4 as appropriate. If you do not change the | 454 | return to step 2, 3, or 4 as appropriate. If you do not change the |
455 | destination manager, the new datastream will be written to the same target. | 455 | destination manager, the new datastream will be written to the same target. |
456 | If you do not change any JPEG parameters, the new datastream will be written | 456 | If you do not change any JPEG parameters, the new datastream will be written |
457 | with the same parameters as before. Note that you can change the input image | 457 | with the same parameters as before. Note that you can change the input image |
458 | dimensions freely between cycles, but if you change the input colorspace, you | 458 | dimensions freely between cycles, but if you change the input colorspace, you |
459 | should call jpeg_set_defaults() to adjust for the new colorspace; and then | 459 | should call jpeg_set_defaults() to adjust for the new colorspace; and then |
460 | you'll need to repeat all of step 3. | 460 | you'll need to repeat all of step 3. |
461 | 461 | ||
462 | 462 | ||
463 | 7. Release the JPEG compression object. | 463 | 7. Release the JPEG compression object. |
464 | 464 | ||
465 | When you are done with a JPEG compression object, destroy it by calling | 465 | When you are done with a JPEG compression object, destroy it by calling |
466 | jpeg_destroy_compress(). This will free all subsidiary memory (regardless of | 466 | jpeg_destroy_compress(). This will free all subsidiary memory (regardless of |
467 | the previous state of the object). Or you can call jpeg_destroy(), which | 467 | the previous state of the object). Or you can call jpeg_destroy(), which |
468 | works for either compression or decompression objects --- this may be more | 468 | works for either compression or decompression objects --- this may be more |
469 | convenient if you are sharing code between compression and decompression | 469 | convenient if you are sharing code between compression and decompression |
470 | cases. (Actually, these routines are equivalent except for the declared type | 470 | cases. (Actually, these routines are equivalent except for the declared type |
471 | of the passed pointer. To avoid gripes from ANSI C compilers, jpeg_destroy() | 471 | of the passed pointer. To avoid gripes from ANSI C compilers, jpeg_destroy() |
472 | should be passed a j_common_ptr.) | 472 | should be passed a j_common_ptr.) |
473 | 473 | ||
474 | If you allocated the jpeg_compress_struct structure from malloc(), freeing | 474 | If you allocated the jpeg_compress_struct structure from malloc(), freeing |
475 | it is your responsibility --- jpeg_destroy() won't. Ditto for the error | 475 | it is your responsibility --- jpeg_destroy() won't. Ditto for the error |
476 | handler structure. | 476 | handler structure. |
477 | 477 | ||
478 | Typical code: | 478 | Typical code: |
479 | 479 | ||
480 | jpeg_destroy_compress(&cinfo); | 480 | jpeg_destroy_compress(&cinfo); |
481 | 481 | ||
482 | 482 | ||
483 | 8. Aborting. | 483 | 8. Aborting. |
484 | 484 | ||
485 | If you decide to abort a compression cycle before finishing, you can clean up | 485 | If you decide to abort a compression cycle before finishing, you can clean up |
486 | in either of two ways: | 486 | in either of two ways: |
487 | 487 | ||
488 | * If you don't need the JPEG object any more, just call | 488 | * If you don't need the JPEG object any more, just call |
489 | jpeg_destroy_compress() or jpeg_destroy() to release memory. This is | 489 | jpeg_destroy_compress() or jpeg_destroy() to release memory. This is |
490 | legitimate at any point after calling jpeg_create_compress() --- in fact, | 490 | legitimate at any point after calling jpeg_create_compress() --- in fact, |
491 | it's safe even if jpeg_create_compress() fails. | 491 | it's safe even if jpeg_create_compress() fails. |
492 | 492 | ||
493 | * If you want to re-use the JPEG object, call jpeg_abort_compress(), or call | 493 | * If you want to re-use the JPEG object, call jpeg_abort_compress(), or call |
494 | jpeg_abort() which works on both compression and decompression objects. | 494 | jpeg_abort() which works on both compression and decompression objects. |
495 | This will return the object to an idle state, releasing any working memory. | 495 | This will return the object to an idle state, releasing any working memory. |
496 | jpeg_abort() is allowed at any time after successful object creation. | 496 | jpeg_abort() is allowed at any time after successful object creation. |
497 | 497 | ||
498 | Note that cleaning up the data destination, if required, is your | 498 | Note that cleaning up the data destination, if required, is your |
499 | responsibility; neither of these routines will call term_destination(). | 499 | responsibility; neither of these routines will call term_destination(). |
500 | (See "Compressed data handling", below, for more about that.) | 500 | (See "Compressed data handling", below, for more about that.) |
501 | 501 | ||
502 | jpeg_destroy() and jpeg_abort() are the only safe calls to make on a JPEG | 502 | jpeg_destroy() and jpeg_abort() are the only safe calls to make on a JPEG |
503 | object that has reported an error by calling error_exit (see "Error handling" | 503 | object that has reported an error by calling error_exit (see "Error handling" |
504 | for more info). The internal state of such an object is likely to be out of | 504 | for more info). The internal state of such an object is likely to be out of |
505 | whack. Either of these two routines will return the object to a known state. | 505 | whack. Either of these two routines will return the object to a known state. |
506 | 506 | ||
507 | 507 | ||
508 | Decompression details | 508 | Decompression details |
509 | --------------------- | 509 | --------------------- |
510 | 510 | ||
511 | Here we revisit the JPEG decompression outline given in the overview. | 511 | Here we revisit the JPEG decompression outline given in the overview. |
512 | 512 | ||
513 | 1. Allocate and initialize a JPEG decompression object. | 513 | 1. Allocate and initialize a JPEG decompression object. |
514 | 514 | ||
515 | This is just like initialization for compression, as discussed above, | 515 | This is just like initialization for compression, as discussed above, |
516 | except that the object is a "struct jpeg_decompress_struct" and you | 516 | except that the object is a "struct jpeg_decompress_struct" and you |
517 | call jpeg_create_decompress(). Error handling is exactly the same. | 517 | call jpeg_create_decompress(). Error handling is exactly the same. |
518 | 518 | ||
519 | Typical code: | 519 | Typical code: |
520 | 520 | ||
521 | struct jpeg_decompress_struct cinfo; | 521 | struct jpeg_decompress_struct cinfo; |
522 | struct jpeg_error_mgr jerr; | 522 | struct jpeg_error_mgr jerr; |
523 | ... | 523 | ... |
524 | cinfo.err = jpeg_std_error(&jerr); | 524 | cinfo.err = jpeg_std_error(&jerr); |
525 | jpeg_create_decompress(&cinfo); | 525 | jpeg_create_decompress(&cinfo); |
526 | 526 | ||
527 | (Both here and in the IJG code, we usually use variable name "cinfo" for | 527 | (Both here and in the IJG code, we usually use variable name "cinfo" for |
528 | both compression and decompression objects.) | 528 | both compression and decompression objects.) |
529 | 529 | ||
530 | 530 | ||
531 | 2. Specify the source of the compressed data (eg, a file). | 531 | 2. Specify the source of the compressed data (eg, a file). |
532 | 532 | ||
533 | As previously mentioned, the JPEG library reads compressed data from a "data | 533 | As previously mentioned, the JPEG library reads compressed data from a "data |
534 | source" module. The library includes one data source module which knows how | 534 | source" module. The library includes one data source module which knows how |
535 | to read from a stdio stream. You can use your own source module if you want | 535 | to read from a stdio stream. You can use your own source module if you want |
536 | to do something else, as discussed later. | 536 | to do something else, as discussed later. |
537 | 537 | ||
538 | If you use the standard source module, you must open the source stdio stream | 538 | If you use the standard source module, you must open the source stdio stream |
539 | beforehand. Typical code for this step looks like: | 539 | beforehand. Typical code for this step looks like: |
540 | 540 | ||
541 | FILE * infile; | 541 | FILE * infile; |
542 | ... | 542 | ... |
543 | if ((infile = fopen(filename, "rb")) == NULL) { | 543 | if ((infile = fopen(filename, "rb")) == NULL) { |
544 | fprintf(stderr, "can't open %s\n", filename); | 544 | fprintf(stderr, "can't open %s\n", filename); |
545 | exit(1); | 545 | exit(1); |
546 | } | 546 | } |
547 | jpeg_stdio_src(&cinfo, infile); | 547 | jpeg_stdio_src(&cinfo, infile); |
548 | 548 | ||
549 | where the last line invokes the standard source module. | 549 | where the last line invokes the standard source module. |
550 | 550 | ||
551 | WARNING: it is critical that the binary compressed data be read unchanged. | 551 | WARNING: it is critical that the binary compressed data be read unchanged. |
552 | On non-Unix systems the stdio library may perform newline translation or | 552 | On non-Unix systems the stdio library may perform newline translation or |
553 | otherwise corrupt binary data. To suppress this behavior, you may need to use | 553 | otherwise corrupt binary data. To suppress this behavior, you may need to use |
554 | a "b" option to fopen (as shown above), or use setmode() or another routine to | 554 | a "b" option to fopen (as shown above), or use setmode() or another routine to |
555 | put the stdio stream in binary mode. See cjpeg.c and djpeg.c for code that | 555 | put the stdio stream in binary mode. See cjpeg.c and djpeg.c for code that |
556 | has been found to work on many systems. | 556 | has been found to work on many systems. |
557 | 557 | ||
558 | You may not change the data source between calling jpeg_read_header() and | 558 | You may not change the data source between calling jpeg_read_header() and |
559 | jpeg_finish_decompress(). If you wish to read a series of JPEG images from | 559 | jpeg_finish_decompress(). If you wish to read a series of JPEG images from |
560 | a single source file, you should repeat the jpeg_read_header() to | 560 | a single source file, you should repeat the jpeg_read_header() to |
561 | jpeg_finish_decompress() sequence without reinitializing either the JPEG | 561 | jpeg_finish_decompress() sequence without reinitializing either the JPEG |
562 | object or the data source module; this prevents buffered input data from | 562 | object or the data source module; this prevents buffered input data from |
563 | being discarded. | 563 | being discarded. |
564 | 564 | ||
565 | 565 | ||
566 | 3. Call jpeg_read_header() to obtain image info. | 566 | 3. Call jpeg_read_header() to obtain image info. |
567 | 567 | ||
568 | Typical code for this step is just | 568 | Typical code for this step is just |
569 | 569 | ||
570 | jpeg_read_header(&cinfo, TRUE); | 570 | jpeg_read_header(&cinfo, TRUE); |
571 | 571 | ||
572 | This will read the source datastream header markers, up to the beginning | 572 | This will read the source datastream header markers, up to the beginning |
573 | of the compressed data proper. On return, the image dimensions and other | 573 | of the compressed data proper. On return, the image dimensions and other |
574 | info have been stored in the JPEG object. The application may wish to | 574 | info have been stored in the JPEG object. The application may wish to |
575 | consult this information before selecting decompression parameters. | 575 | consult this information before selecting decompression parameters. |
576 | 576 | ||
577 | More complex code is necessary if | 577 | More complex code is necessary if |
578 | * A suspending data source is used --- in that case jpeg_read_header() | 578 | * A suspending data source is used --- in that case jpeg_read_header() |
579 | may return before it has read all the header data. See "I/O suspension", | 579 | may return before it has read all the header data. See "I/O suspension", |
580 | below. The normal stdio source manager will NOT cause this to happen. | 580 | below. The normal stdio source manager will NOT cause this to happen. |
581 | * Abbreviated JPEG files are to be processed --- see the section on | 581 | * Abbreviated JPEG files are to be processed --- see the section on |
582 | abbreviated datastreams. Standard applications that deal only in | 582 | abbreviated datastreams. Standard applications that deal only in |
583 | interchange JPEG files need not be concerned with this case either. | 583 | interchange JPEG files need not be concerned with this case either. |
584 | 584 | ||
585 | It is permissible to stop at this point if you just wanted to find out the | 585 | It is permissible to stop at this point if you just wanted to find out the |
586 | image dimensions and other header info for a JPEG file. In that case, | 586 | image dimensions and other header info for a JPEG file. In that case, |
587 | call jpeg_destroy() when you are done with the JPEG object, or call | 587 | call jpeg_destroy() when you are done with the JPEG object, or call |
588 | jpeg_abort() to return it to an idle state before selecting a new data | 588 | jpeg_abort() to return it to an idle state before selecting a new data |
589 | source and reading another header. | 589 | source and reading another header. |
590 | 590 | ||
591 | 591 | ||
592 | 4. Set parameters for decompression. | 592 | 4. Set parameters for decompression. |
593 | 593 | ||
594 | jpeg_read_header() sets appropriate default decompression parameters based on | 594 | jpeg_read_header() sets appropriate default decompression parameters based on |
595 | the properties of the image (in particular, its colorspace). However, you | 595 | the properties of the image (in particular, its colorspace). However, you |
596 | may well want to alter these defaults before beginning the decompression. | 596 | may well want to alter these defaults before beginning the decompression. |
597 | For example, the default is to produce full color output from a color file. | 597 | For example, the default is to produce full color output from a color file. |
598 | If you want colormapped output you must ask for it. Other options allow the | 598 | If you want colormapped output you must ask for it. Other options allow the |
599 | returned image to be scaled and allow various speed/quality tradeoffs to be | 599 | returned image to be scaled and allow various speed/quality tradeoffs to be |
600 | selected. "Decompression parameter selection", below, gives details. | 600 | selected. "Decompression parameter selection", below, gives details. |
601 | 601 | ||
602 | If the defaults are appropriate, nothing need be done at this step. | 602 | If the defaults are appropriate, nothing need be done at this step. |
603 | 603 | ||
604 | Note that all default values are set by each call to jpeg_read_header(). | 604 | Note that all default values are set by each call to jpeg_read_header(). |
605 | If you reuse a decompression object, you cannot expect your parameter | 605 | If you reuse a decompression object, you cannot expect your parameter |
606 | settings to be preserved across cycles, as you can for compression. | 606 | settings to be preserved across cycles, as you can for compression. |
607 | You must set desired parameter values each time. | 607 | You must set desired parameter values each time. |
608 | 608 | ||
609 | 609 | ||
610 | 5. jpeg_start_decompress(...); | 610 | 5. jpeg_start_decompress(...); |
611 | 611 | ||
612 | Once the parameter values are satisfactory, call jpeg_start_decompress() to | 612 | Once the parameter values are satisfactory, call jpeg_start_decompress() to |
613 | begin decompression. This will initialize internal state, allocate working | 613 | begin decompression. This will initialize internal state, allocate working |
614 | memory, and prepare for returning data. | 614 | memory, and prepare for returning data. |
615 | 615 | ||
616 | Typical code is just | 616 | Typical code is just |
617 | 617 | ||
618 | jpeg_start_decompress(&cinfo); | 618 | jpeg_start_decompress(&cinfo); |
619 | 619 | ||
620 | If you have requested a multi-pass operating mode, such as 2-pass color | 620 | If you have requested a multi-pass operating mode, such as 2-pass color |
621 | quantization, jpeg_start_decompress() will do everything needed before data | 621 | quantization, jpeg_start_decompress() will do everything needed before data |
622 | output can begin. In this case jpeg_start_decompress() may take quite a while | 622 | output can begin. In this case jpeg_start_decompress() may take quite a while |
623 | to complete. With a single-scan (non progressive) JPEG file and default | 623 | to complete. With a single-scan (non progressive) JPEG file and default |
624 | decompression parameters, this will not happen; jpeg_start_decompress() will | 624 | decompression parameters, this will not happen; jpeg_start_decompress() will |
625 | return quickly. | 625 | return quickly. |
626 | 626 | ||
627 | After this call, the final output image dimensions, including any requested | 627 | After this call, the final output image dimensions, including any requested |
628 | scaling, are available in the JPEG object; so is the selected colormap, if | 628 | scaling, are available in the JPEG object; so is the selected colormap, if |
629 | colormapped output has been requested. Useful fields include | 629 | colormapped output has been requested. Useful fields include |
630 | 630 | ||
631 | output_width image width and height, as scaled | 631 | output_width image width and height, as scaled |
632 | output_height | 632 | output_height |
633 | out_color_components # of color components in out_color_space | 633 | out_color_components # of color components in out_color_space |
634 | output_components # of color components returned per pixel | 634 | output_components # of color components returned per pixel |
635 | colormap the selected colormap, if any | 635 | colormap the selected colormap, if any |
636 | actual_number_of_colors number of entries in colormap | 636 | actual_number_of_colors number of entries in colormap |
637 | 637 | ||
638 | output_components is 1 (a colormap index) when quantizing colors; otherwise it | 638 | output_components is 1 (a colormap index) when quantizing colors; otherwise it |
639 | equals out_color_components. It is the number of JSAMPLE values that will be | 639 | equals out_color_components. It is the number of JSAMPLE values that will be |
640 | emitted per pixel in the output arrays. | 640 | emitted per pixel in the output arrays. |
641 | 641 | ||
642 | Typically you will need to allocate data buffers to hold the incoming image. | 642 | Typically you will need to allocate data buffers to hold the incoming image. |
643 | You will need output_width * output_components JSAMPLEs per scanline in your | 643 | You will need output_width * output_components JSAMPLEs per scanline in your |
644 | output buffer, and a total of output_height scanlines will be returned. | 644 | output buffer, and a total of output_height scanlines will be returned. |
645 | 645 | ||
646 | Note: if you are using the JPEG library's internal memory manager to allocate | 646 | Note: if you are using the JPEG library's internal memory manager to allocate |
647 | data buffers (as djpeg does), then the manager's protocol requires that you | 647 | data buffers (as djpeg does), then the manager's protocol requires that you |
648 | request large buffers *before* calling jpeg_start_decompress(). This is a | 648 | request large buffers *before* calling jpeg_start_decompress(). This is a |
649 | little tricky since the output_XXX fields are not normally valid then. You | 649 | little tricky since the output_XXX fields are not normally valid then. You |
650 | can make them valid by calling jpeg_calc_output_dimensions() after setting the | 650 | can make them valid by calling jpeg_calc_output_dimensions() after setting the |
651 | relevant parameters (scaling, output color space, and quantization flag). | 651 | relevant parameters (scaling, output color space, and quantization flag). |
652 | 652 | ||
653 | 653 | ||
654 | 6. while (scan lines remain to be read) | 654 | 6. while (scan lines remain to be read) |
655 | jpeg_read_scanlines(...); | 655 | jpeg_read_scanlines(...); |
656 | 656 | ||
657 | Now you can read the decompressed image data by calling jpeg_read_scanlines() | 657 | Now you can read the decompressed image data by calling jpeg_read_scanlines() |
658 | one or more times. At each call, you pass in the maximum number of scanlines | 658 | one or more times. At each call, you pass in the maximum number of scanlines |
659 | to be read (ie, the height of your working buffer); jpeg_read_scanlines() | 659 | to be read (ie, the height of your working buffer); jpeg_read_scanlines() |
660 | will return up to that many lines. The return value is the number of lines | 660 | will return up to that many lines. The return value is the number of lines |
661 | actually read. The format of the returned data is discussed under "Data | 661 | actually read. The format of the returned data is discussed under "Data |
662 | formats", above. Don't forget that grayscale and color JPEGs will return | 662 | formats", above. Don't forget that grayscale and color JPEGs will return |
663 | different data formats! | 663 | different data formats! |
664 | 664 | ||
665 | Image data is returned in top-to-bottom scanline order. If you must write | 665 | Image data is returned in top-to-bottom scanline order. If you must write |
666 | out the image in bottom-to-top order, you can use the JPEG library's virtual | 666 | out the image in bottom-to-top order, you can use the JPEG library's virtual |
667 | array mechanism to invert the data efficiently. Examples of this can be | 667 | array mechanism to invert the data efficiently. Examples of this can be |
668 | found in the sample application djpeg. | 668 | found in the sample application djpeg. |
669 | 669 | ||
670 | The library maintains a count of the number of scanlines returned so far | 670 | The library maintains a count of the number of scanlines returned so far |
671 | in the output_scanline field of the JPEG object. Usually you can just use | 671 | in the output_scanline field of the JPEG object. Usually you can just use |
672 | this variable as the loop counter, so that the loop test looks like | 672 | this variable as the loop counter, so that the loop test looks like |
673 | "while (cinfo.output_scanline < cinfo.output_height)". (Note that the test | 673 | "while (cinfo.output_scanline < cinfo.output_height)". (Note that the test |
674 | should NOT be against image_height, unless you never use scaling. The | 674 | should NOT be against image_height, unless you never use scaling. The |
675 | image_height field is the height of the original unscaled image.) | 675 | image_height field is the height of the original unscaled image.) |
676 | The return value always equals the change in the value of output_scanline. | 676 | The return value always equals the change in the value of output_scanline. |
677 | 677 | ||
678 | If you don't use a suspending data source, it is safe to assume that | 678 | If you don't use a suspending data source, it is safe to assume that |
679 | jpeg_read_scanlines() reads at least one scanline per call, until the | 679 | jpeg_read_scanlines() reads at least one scanline per call, until the |
680 | bottom of the image has been reached. | 680 | bottom of the image has been reached. |
681 | 681 | ||
682 | If you use a buffer larger than one scanline, it is NOT safe to assume that | 682 | If you use a buffer larger than one scanline, it is NOT safe to assume that |
683 | jpeg_read_scanlines() fills it. (The current implementation returns only a | 683 | jpeg_read_scanlines() fills it. (The current implementation returns only a |
684 | few scanlines per call, no matter how large a buffer you pass.) So you must | 684 | few scanlines per call, no matter how large a buffer you pass.) So you must |
685 | always provide a loop that calls jpeg_read_scanlines() repeatedly until the | 685 | always provide a loop that calls jpeg_read_scanlines() repeatedly until the |
686 | whole image has been read. | 686 | whole image has been read. |
687 | 687 | ||
688 | 688 | ||
689 | 7. jpeg_finish_decompress(...); | 689 | 7. jpeg_finish_decompress(...); |
690 | 690 | ||
691 | After all the image data has been read, call jpeg_finish_decompress() to | 691 | After all the image data has been read, call jpeg_finish_decompress() to |
692 | complete the decompression cycle. This causes working memory associated | 692 | complete the decompression cycle. This causes working memory associated |
693 | with the JPEG object to be released. | 693 | with the JPEG object to be released. |
694 | 694 | ||
695 | Typical code: | 695 | Typical code: |
696 | 696 | ||
697 | jpeg_finish_decompress(&cinfo); | 697 | jpeg_finish_decompress(&cinfo); |
698 | 698 | ||
699 | If using the stdio source manager, don't forget to close the source stdio | 699 | If using the stdio source manager, don't forget to close the source stdio |
700 | stream if necessary. | 700 | stream if necessary. |
701 | 701 | ||
702 | It is an error to call jpeg_finish_decompress() before reading the correct | 702 | It is an error to call jpeg_finish_decompress() before reading the correct |
703 | total number of scanlines. If you wish to abort decompression, call | 703 | total number of scanlines. If you wish to abort decompression, call |
704 | jpeg_abort() as discussed below. | 704 | jpeg_abort() as discussed below. |
705 | 705 | ||
706 | After completing a decompression cycle, you may dispose of the JPEG object as | 706 | After completing a decompression cycle, you may dispose of the JPEG object as |
707 | discussed next, or you may use it to decompress another image. In that case | 707 | discussed next, or you may use it to decompress another image. In that case |
708 | return to step 2 or 3 as appropriate. If you do not change the source | 708 | return to step 2 or 3 as appropriate. If you do not change the source |
709 | manager, the next image will be read from the same source. | 709 | manager, the next image will be read from the same source. |
710 | 710 | ||
711 | 711 | ||
712 | 8. Release the JPEG decompression object. | 712 | 8. Release the JPEG decompression object. |
713 | 713 | ||
714 | When you are done with a JPEG decompression object, destroy it by calling | 714 | When you are done with a JPEG decompression object, destroy it by calling |
715 | jpeg_destroy_decompress() or jpeg_destroy(). The previous discussion of | 715 | jpeg_destroy_decompress() or jpeg_destroy(). The previous discussion of |
716 | destroying compression objects applies here too. | 716 | destroying compression objects applies here too. |
717 | 717 | ||
718 | Typical code: | 718 | Typical code: |
719 | 719 | ||
720 | jpeg_destroy_decompress(&cinfo); | 720 | jpeg_destroy_decompress(&cinfo); |
721 | 721 | ||
722 | 722 | ||
723 | 9. Aborting. | 723 | 9. Aborting. |
724 | 724 | ||
725 | You can abort a decompression cycle by calling jpeg_destroy_decompress() or | 725 | You can abort a decompression cycle by calling jpeg_destroy_decompress() or |
726 | jpeg_destroy() if you don't need the JPEG object any more, or | 726 | jpeg_destroy() if you don't need the JPEG object any more, or |
727 | jpeg_abort_decompress() or jpeg_abort() if you want to reuse the object. | 727 | jpeg_abort_decompress() or jpeg_abort() if you want to reuse the object. |
728 | The previous discussion of aborting compression cycles applies here too. | 728 | The previous discussion of aborting compression cycles applies here too. |
729 | 729 | ||
730 | 730 | ||
731 | Mechanics of usage: include files, linking, etc | 731 | Mechanics of usage: include files, linking, etc |
732 | ----------------------------------------------- | 732 | ----------------------------------------------- |
733 | 733 | ||
734 | Applications using the JPEG library should include the header file jpeglib.h | 734 | Applications using the JPEG library should include the header file jpeglib.h |
735 | to obtain declarations of data types and routines. Before including | 735 | to obtain declarations of data types and routines. Before including |
736 | jpeglib.h, include system headers that define at least the typedefs FILE and | 736 | jpeglib.h, include system headers that define at least the typedefs FILE and |
737 | size_t. On ANSI-conforming systems, including <stdio.h> is sufficient; on | 737 | size_t. On ANSI-conforming systems, including <stdio.h> is sufficient; on |
738 | older Unix systems, you may need <sys/types.h> to define size_t. | 738 | older Unix systems, you may need <sys/types.h> to define size_t. |
739 | 739 | ||
740 | If the application needs to refer to individual JPEG library error codes, also | 740 | If the application needs to refer to individual JPEG library error codes, also |
741 | include jerror.h to define those symbols. | 741 | include jerror.h to define those symbols. |
742 | 742 | ||
743 | jpeglib.h indirectly includes the files jconfig.h and jmorecfg.h. If you are | 743 | jpeglib.h indirectly includes the files jconfig.h and jmorecfg.h. If you are |
744 | installing the JPEG header files in a system directory, you will want to | 744 | installing the JPEG header files in a system directory, you will want to |
745 | install all four files: jpeglib.h, jerror.h, jconfig.h, jmorecfg.h. | 745 | install all four files: jpeglib.h, jerror.h, jconfig.h, jmorecfg.h. |
746 | 746 | ||
747 | The most convenient way to include the JPEG code into your executable program | 747 | The most convenient way to include the JPEG code into your executable program |
748 | is to prepare a library file ("libjpeg.a", or a corresponding name on non-Unix | 748 | is to prepare a library file ("libjpeg.a", or a corresponding name on non-Unix |
749 | machines) and reference it at your link step. If you use only half of the | 749 | machines) and reference it at your link step. If you use only half of the |
750 | library (only compression or only decompression), only that much code will be | 750 | library (only compression or only decompression), only that much code will be |
751 | included from the library, unless your linker is hopelessly brain-damaged. | 751 | included from the library, unless your linker is hopelessly brain-damaged. |
752 | The supplied makefiles build libjpeg.a automatically (see install.txt). | 752 | The supplied makefiles build libjpeg.a automatically (see install.txt). |
753 | 753 | ||
754 | While you can build the JPEG library as a shared library if the whim strikes | 754 | While you can build the JPEG library as a shared library if the whim strikes |
755 | you, we don't really recommend it. The trouble with shared libraries is that | 755 | you, we don't really recommend it. The trouble with shared libraries is that |
756 | at some point you'll probably try to substitute a new version of the library | 756 | at some point you'll probably try to substitute a new version of the library |
757 | without recompiling the calling applications. That generally doesn't work | 757 | without recompiling the calling applications. That generally doesn't work |
758 | because the parameter struct declarations usually change with each new | 758 | because the parameter struct declarations usually change with each new |
759 | version. In other words, the library's API is *not* guaranteed binary | 759 | version. In other words, the library's API is *not* guaranteed binary |
760 | compatible across versions; we only try to ensure source-code compatibility. | 760 | compatible across versions; we only try to ensure source-code compatibility. |
761 | (In hindsight, it might have been smarter to hide the parameter structs from | 761 | (In hindsight, it might have been smarter to hide the parameter structs from |
762 | applications and introduce a ton of access functions instead. Too late now, | 762 | applications and introduce a ton of access functions instead. Too late now, |
763 | however.) | 763 | however.) |
764 | 764 | ||
765 | On some systems your application may need to set up a signal handler to ensure | 765 | On some systems your application may need to set up a signal handler to ensure |
766 | that temporary files are deleted if the program is interrupted. This is most | 766 | that temporary files are deleted if the program is interrupted. This is most |
767 | critical if you are on MS-DOS and use the jmemdos.c memory manager back end; | 767 | critical if you are on MS-DOS and use the jmemdos.c memory manager back end; |
768 | it will try to grab extended memory for temp files, and that space will NOT be | 768 | it will try to grab extended memory for temp files, and that space will NOT be |
769 | freed automatically. See cjpeg.c or djpeg.c for an example signal handler. | 769 | freed automatically. See cjpeg.c or djpeg.c for an example signal handler. |
770 | 770 | ||
771 | It may be worth pointing out that the core JPEG library does not actually | 771 | It may be worth pointing out that the core JPEG library does not actually |
772 | require the stdio library: only the default source/destination managers and | 772 | require the stdio library: only the default source/destination managers and |
773 | error handler need it. You can use the library in a stdio-less environment | 773 | error handler need it. You can use the library in a stdio-less environment |
774 | if you replace those modules and use jmemnobs.c (or another memory manager of | 774 | if you replace those modules and use jmemnobs.c (or another memory manager of |
775 | your own devising). More info about the minimum system library requirements | 775 | your own devising). More info about the minimum system library requirements |
776 | may be found in jinclude.h. | 776 | may be found in jinclude.h. |
777 | 777 | ||
778 | 778 | ||
779 | ADVANCED FEATURES | 779 | ADVANCED FEATURES |
780 | ================= | 780 | ================= |
781 | 781 | ||
782 | Compression parameter selection | 782 | Compression parameter selection |
783 | ------------------------------- | 783 | ------------------------------- |
784 | 784 | ||
785 | This section describes all the optional parameters you can set for JPEG | 785 | This section describes all the optional parameters you can set for JPEG |
786 | compression, as well as the "helper" routines provided to assist in this | 786 | compression, as well as the "helper" routines provided to assist in this |
787 | task. Proper setting of some parameters requires detailed understanding | 787 | task. Proper setting of some parameters requires detailed understanding |
788 | of the JPEG standard; if you don't know what a parameter is for, it's best | 788 | of the JPEG standard; if you don't know what a parameter is for, it's best |
789 | not to mess with it! See REFERENCES in the README file for pointers to | 789 | not to mess with it! See REFERENCES in the README file for pointers to |
790 | more info about JPEG. | 790 | more info about JPEG. |
791 | 791 | ||
792 | It's a good idea to call jpeg_set_defaults() first, even if you plan to set | 792 | It's a good idea to call jpeg_set_defaults() first, even if you plan to set |
793 | all the parameters; that way your code is more likely to work with future JPEG | 793 | all the parameters; that way your code is more likely to work with future JPEG |
794 | libraries that have additional parameters. For the same reason, we recommend | 794 | libraries that have additional parameters. For the same reason, we recommend |
795 | you use a helper routine where one is provided, in preference to twiddling | 795 | you use a helper routine where one is provided, in preference to twiddling |
796 | cinfo fields directly. | 796 | cinfo fields directly. |
797 | 797 | ||
798 | The helper routines are: | 798 | The helper routines are: |
799 | 799 | ||
800 | jpeg_set_defaults (j_compress_ptr cinfo) | 800 | jpeg_set_defaults (j_compress_ptr cinfo) |
801 | This routine sets all JPEG parameters to reasonable defaults, using | 801 | This routine sets all JPEG parameters to reasonable defaults, using |
802 | only the input image's color space (field in_color_space, which must | 802 | only the input image's color space (field in_color_space, which must |
803 | already be set in cinfo). Many applications will only need to use | 803 | already be set in cinfo). Many applications will only need to use |
804 | this routine and perhaps jpeg_set_quality(). | 804 | this routine and perhaps jpeg_set_quality(). |
805 | 805 | ||
806 | jpeg_set_colorspace (j_compress_ptr cinfo, J_COLOR_SPACE colorspace) | 806 | jpeg_set_colorspace (j_compress_ptr cinfo, J_COLOR_SPACE colorspace) |
807 | Sets the JPEG file's colorspace (field jpeg_color_space) as specified, | 807 | Sets the JPEG file's colorspace (field jpeg_color_space) as specified, |
808 | and sets other color-space-dependent parameters appropriately. See | 808 | and sets other color-space-dependent parameters appropriately. See |
809 | "Special color spaces", below, before using this. A large number of | 809 | "Special color spaces", below, before using this. A large number of |
810 | parameters, including all per-component parameters, are set by this | 810 | parameters, including all per-component parameters, are set by this |
811 | routine; if you want to twiddle individual parameters you should call | 811 | routine; if you want to twiddle individual parameters you should call |
812 | jpeg_set_colorspace() before rather than after. | 812 | jpeg_set_colorspace() before rather than after. |
813 | 813 | ||
814 | jpeg_default_colorspace (j_compress_ptr cinfo) | 814 | jpeg_default_colorspace (j_compress_ptr cinfo) |
815 | Selects an appropriate JPEG colorspace based on cinfo->in_color_space, | 815 | Selects an appropriate JPEG colorspace based on cinfo->in_color_space, |
816 | and calls jpeg_set_colorspace(). This is actually a subroutine of | 816 | and calls jpeg_set_colorspace(). This is actually a subroutine of |
817 | jpeg_set_defaults(). It's broken out in case you want to change | 817 | jpeg_set_defaults(). It's broken out in case you want to change |
818 | just the colorspace-dependent JPEG parameters. | 818 | just the colorspace-dependent JPEG parameters. |
819 | 819 | ||
820 | jpeg_set_quality (j_compress_ptr cinfo, int quality, boolean force_baseline) | 820 | jpeg_set_quality (j_compress_ptr cinfo, int quality, boolean force_baseline) |
821 | Constructs JPEG quantization tables appropriate for the indicated | 821 | Constructs JPEG quantization tables appropriate for the indicated |
822 | quality setting. The quality value is expressed on the 0..100 scale | 822 | quality setting. The quality value is expressed on the 0..100 scale |
823 | recommended by IJG (cjpeg's "-quality" switch uses this routine). | 823 | recommended by IJG (cjpeg's "-quality" switch uses this routine). |
824 | Note that the exact mapping from quality values to tables may change | 824 | Note that the exact mapping from quality values to tables may change |
825 | in future IJG releases as more is learned about DCT quantization. | 825 | in future IJG releases as more is learned about DCT quantization. |
826 | If the force_baseline parameter is TRUE, then the quantization table | 826 | If the force_baseline parameter is TRUE, then the quantization table |
827 | entries are constrained to the range 1..255 for full JPEG baseline | 827 | entries are constrained to the range 1..255 for full JPEG baseline |
828 | compatibility. In the current implementation, this only makes a | 828 | compatibility. In the current implementation, this only makes a |
829 | difference for quality settings below 25, and it effectively prevents | 829 | difference for quality settings below 25, and it effectively prevents |
830 | very small/low quality files from being generated. The IJG decoder | 830 | very small/low quality files from being generated. The IJG decoder |
831 | is capable of reading the non-baseline files generated at low quality | 831 | is capable of reading the non-baseline files generated at low quality |
832 | settings when force_baseline is FALSE, but other decoders may not be. | 832 | settings when force_baseline is FALSE, but other decoders may not be. |
833 | 833 | ||
834 | jpeg_set_linear_quality (j_compress_ptr cinfo, int scale_factor, | 834 | jpeg_set_linear_quality (j_compress_ptr cinfo, int scale_factor, |
835 | boolean force_baseline) | 835 | boolean force_baseline) |
836 | Same as jpeg_set_quality() except that the generated tables are the | 836 | Same as jpeg_set_quality() except that the generated tables are the |
837 | sample tables given in the JPEC spec section K.1, multiplied by the | 837 | sample tables given in the JPEC spec section K.1, multiplied by the |
838 | specified scale factor (which is expressed as a percentage; thus | 838 | specified scale factor (which is expressed as a percentage; thus |
839 | scale_factor = 100 reproduces the spec's tables). Note that larger | 839 | scale_factor = 100 reproduces the spec's tables). Note that larger |
840 | scale factors give lower quality. This entry point is useful for | 840 | scale factors give lower quality. This entry point is useful for |
841 | conforming to the Adobe PostScript DCT conventions, but we do not | 841 | conforming to the Adobe PostScript DCT conventions, but we do not |
842 | recommend linear scaling as a user-visible quality scale otherwise. | 842 | recommend linear scaling as a user-visible quality scale otherwise. |
843 | force_baseline again constrains the computed table entries to 1..255. | 843 | force_baseline again constrains the computed table entries to 1..255. |
844 | 844 | ||
845 | int jpeg_quality_scaling (int quality) | 845 | int jpeg_quality_scaling (int quality) |
846 | Converts a value on the IJG-recommended quality scale to a linear | 846 | Converts a value on the IJG-recommended quality scale to a linear |
847 | scaling percentage. Note that this routine may change or go away | 847 | scaling percentage. Note that this routine may change or go away |
848 | in future releases --- IJG may choose to adopt a scaling method that | 848 | in future releases --- IJG may choose to adopt a scaling method that |
849 | can't be expressed as a simple scalar multiplier, in which case the | 849 | can't be expressed as a simple scalar multiplier, in which case the |
850 | premise of this routine collapses. Caveat user. | 850 | premise of this routine collapses. Caveat user. |
851 | 851 | ||
852 | jpeg_default_qtables (j_compress_ptr cinfo, boolean force_baseline) | 852 | jpeg_default_qtables (j_compress_ptr cinfo, boolean force_baseline) |
853 | Set default quantization tables with linear q_scale_factor[] values | 853 | Set default quantization tables with linear q_scale_factor[] values |
854 | (see below). | 854 | (see below). |
855 | 855 | ||
856 | jpeg_add_quant_table (j_compress_ptr cinfo, int which_tbl, | 856 | jpeg_add_quant_table (j_compress_ptr cinfo, int which_tbl, |
857 | const unsigned int *basic_table, | 857 | const unsigned int *basic_table, |
858 | int scale_factor, boolean force_baseline) | 858 | int scale_factor, boolean force_baseline) |
859 | Allows an arbitrary quantization table to be created. which_tbl | 859 | Allows an arbitrary quantization table to be created. which_tbl |
860 | indicates which table slot to fill. basic_table points to an array | 860 | indicates which table slot to fill. basic_table points to an array |
861 | of 64 unsigned ints given in normal array order. These values are | 861 | of 64 unsigned ints given in normal array order. These values are |
862 | multiplied by scale_factor/100 and then clamped to the range 1..65535 | 862 | multiplied by scale_factor/100 and then clamped to the range 1..65535 |
863 | (or to 1..255 if force_baseline is TRUE). | 863 | (or to 1..255 if force_baseline is TRUE). |
864 | CAUTION: prior to library version 6a, jpeg_add_quant_table expected | 864 | CAUTION: prior to library version 6a, jpeg_add_quant_table expected |
865 | the basic table to be given in JPEG zigzag order. If you need to | 865 | the basic table to be given in JPEG zigzag order. If you need to |
866 | write code that works with either older or newer versions of this | 866 | write code that works with either older or newer versions of this |
867 | routine, you must check the library version number. Something like | 867 | routine, you must check the library version number. Something like |
868 | "#if JPEG_LIB_VERSION >= 61" is the right test. | 868 | "#if JPEG_LIB_VERSION >= 61" is the right test. |
869 | 869 | ||
870 | jpeg_simple_progression (j_compress_ptr cinfo) | 870 | jpeg_simple_progression (j_compress_ptr cinfo) |
871 | Generates a default scan script for writing a progressive-JPEG file. | 871 | Generates a default scan script for writing a progressive-JPEG file. |
872 | This is the recommended method of creating a progressive file, | 872 | This is the recommended method of creating a progressive file, |
873 | unless you want to make a custom scan sequence. You must ensure that | 873 | unless you want to make a custom scan sequence. You must ensure that |
874 | the JPEG color space is set correctly before calling this routine. | 874 | the JPEG color space is set correctly before calling this routine. |
875 | 875 | ||
876 | 876 | ||
877 | Compression parameters (cinfo fields) include: | 877 | Compression parameters (cinfo fields) include: |
878 | 878 | ||
879 | int block_size | 879 | int block_size |
880 | Set DCT block size. All N from 1 to 16 are possible. | 880 | Set DCT block size. All N from 1 to 16 are possible. |
881 | Default is 8 (baseline format). | 881 | Default is 8 (baseline format). |
882 | Larger values produce higher compression, | 882 | Larger values produce higher compression, |
883 | smaller values produce higher quality. | 883 | smaller values produce higher quality. |
884 | An exact DCT stage is possible with 1 or 2. | 884 | An exact DCT stage is possible with 1 or 2. |
885 | With the default quality of 75 and default Luminance qtable | 885 | With the default quality of 75 and default Luminance qtable |
886 | the DCT+Quantization stage is lossless for value 1. | 886 | the DCT+Quantization stage is lossless for value 1. |
887 | Note that values other than 8 require a SmartScale capable decoder, | 887 | Note that values other than 8 require a SmartScale capable decoder, |
888 | introduced with IJG JPEG 8. Setting the block_size parameter for | 888 | introduced with IJG JPEG 8. Setting the block_size parameter for |
889 | compression works with version 8c and later. | 889 | compression works with version 8c and later. |
890 | 890 | ||
891 | J_DCT_METHOD dct_method | 891 | J_DCT_METHOD dct_method |
892 | Selects the algorithm used for the DCT step. Choices are: | 892 | Selects the algorithm used for the DCT step. Choices are: |
893 | JDCT_ISLOW: slow but accurate integer algorithm | 893 | JDCT_ISLOW: slow but accurate integer algorithm |
894 | JDCT_IFAST: faster, less accurate integer method | 894 | JDCT_IFAST: faster, less accurate integer method |
895 | JDCT_FLOAT: floating-point method | 895 | JDCT_FLOAT: floating-point method |
896 | JDCT_DEFAULT: default method (normally JDCT_ISLOW) | 896 | JDCT_DEFAULT: default method (normally JDCT_ISLOW) |
897 | JDCT_FASTEST: fastest method (normally JDCT_IFAST) | 897 | JDCT_FASTEST: fastest method (normally JDCT_IFAST) |
898 | The FLOAT method is very slightly more accurate than the ISLOW method, | 898 | The FLOAT method is very slightly more accurate than the ISLOW method, |
899 | but may give different results on different machines due to varying | 899 | but may give different results on different machines due to varying |
900 | roundoff behavior. The integer methods should give the same results | 900 | roundoff behavior. The integer methods should give the same results |
901 | on all machines. On machines with sufficiently fast FP hardware, the | 901 | on all machines. On machines with sufficiently fast FP hardware, the |
902 | floating-point method may also be the fastest. The IFAST method is | 902 | floating-point method may also be the fastest. The IFAST method is |
903 | considerably less accurate than the other two; its use is not | 903 | considerably less accurate than the other two; its use is not |
904 | recommended if high quality is a concern. JDCT_DEFAULT and | 904 | recommended if high quality is a concern. JDCT_DEFAULT and |
905 | JDCT_FASTEST are macros configurable by each installation. | 905 | JDCT_FASTEST are macros configurable by each installation. |
906 | 906 | ||
907 | unsigned int scale_num, scale_denom | 907 | unsigned int scale_num, scale_denom |
908 | Scale the image by the fraction scale_num/scale_denom. Default is | 908 | Scale the image by the fraction scale_num/scale_denom. Default is |
909 | 1/1, or no scaling. Currently, the supported scaling ratios are | 909 | 1/1, or no scaling. Currently, the supported scaling ratios are |
910 | M/N with all N from 1 to 16, where M is the destination DCT size, | 910 | M/N with all N from 1 to 16, where M is the destination DCT size, |
911 | which is 8 by default (see block_size parameter above). | 911 | which is 8 by default (see block_size parameter above). |
912 | (The library design allows for arbitrary scaling ratios but this | 912 | (The library design allows for arbitrary scaling ratios but this |
913 | is not likely to be implemented any time soon.) | 913 | is not likely to be implemented any time soon.) |
914 | 914 | ||
915 | J_COLOR_SPACE jpeg_color_space | 915 | J_COLOR_SPACE jpeg_color_space |
916 | int num_components | 916 | int num_components |
917 | The JPEG color space and corresponding number of components; see | 917 | The JPEG color space and corresponding number of components; see |
918 | "Special color spaces", below, for more info. We recommend using | 918 | "Special color spaces", below, for more info. We recommend using |
919 | jpeg_set_color_space() if you want to change these. | 919 | jpeg_set_color_space() if you want to change these. |
920 | 920 | ||
921 | boolean optimize_coding | 921 | boolean optimize_coding |
922 | TRUE causes the compressor to compute optimal Huffman coding tables | 922 | TRUE causes the compressor to compute optimal Huffman coding tables |
923 | for the image. This requires an extra pass over the data and | 923 | for the image. This requires an extra pass over the data and |
924 | therefore costs a good deal of space and time. The default is | 924 | therefore costs a good deal of space and time. The default is |
925 | FALSE, which tells the compressor to use the supplied or default | 925 | FALSE, which tells the compressor to use the supplied or default |
926 | Huffman tables. In most cases optimal tables save only a few percent | 926 | Huffman tables. In most cases optimal tables save only a few percent |
927 | of file size compared to the default tables. Note that when this is | 927 | of file size compared to the default tables. Note that when this is |
928 | TRUE, you need not supply Huffman tables at all, and any you do | 928 | TRUE, you need not supply Huffman tables at all, and any you do |
929 | supply will be overwritten. | 929 | supply will be overwritten. |
930 | 930 | ||
931 | unsigned int restart_interval | 931 | unsigned int restart_interval |
932 | int restart_in_rows | 932 | int restart_in_rows |
933 | To emit restart markers in the JPEG file, set one of these nonzero. | 933 | To emit restart markers in the JPEG file, set one of these nonzero. |
934 | Set restart_interval to specify the exact interval in MCU blocks. | 934 | Set restart_interval to specify the exact interval in MCU blocks. |
935 | Set restart_in_rows to specify the interval in MCU rows. (If | 935 | Set restart_in_rows to specify the interval in MCU rows. (If |
936 | restart_in_rows is not 0, then restart_interval is set after the | 936 | restart_in_rows is not 0, then restart_interval is set after the |
937 | image width in MCUs is computed.) Defaults are zero (no restarts). | 937 | image width in MCUs is computed.) Defaults are zero (no restarts). |
938 | One restart marker per MCU row is often a good choice. | 938 | One restart marker per MCU row is often a good choice. |
939 | NOTE: the overhead of restart markers is higher in grayscale JPEG | 939 | NOTE: the overhead of restart markers is higher in grayscale JPEG |
940 | files than in color files, and MUCH higher in progressive JPEGs. | 940 | files than in color files, and MUCH higher in progressive JPEGs. |
941 | If you use restarts, you may want to use larger intervals in those | 941 | If you use restarts, you may want to use larger intervals in those |
942 | cases. | 942 | cases. |
943 | 943 | ||
944 | const jpeg_scan_info * scan_info | 944 | const jpeg_scan_info * scan_info |
945 | int num_scans | 945 | int num_scans |
946 | By default, scan_info is NULL; this causes the compressor to write a | 946 | By default, scan_info is NULL; this causes the compressor to write a |
947 | single-scan sequential JPEG file. If not NULL, scan_info points to | 947 | single-scan sequential JPEG file. If not NULL, scan_info points to |
948 | an array of scan definition records of length num_scans. The | 948 | an array of scan definition records of length num_scans. The |
949 | compressor will then write a JPEG file having one scan for each scan | 949 | compressor will then write a JPEG file having one scan for each scan |
950 | definition record. This is used to generate noninterleaved or | 950 | definition record. This is used to generate noninterleaved or |
951 | progressive JPEG files. The library checks that the scan array | 951 | progressive JPEG files. The library checks that the scan array |
952 | defines a valid JPEG scan sequence. (jpeg_simple_progression creates | 952 | defines a valid JPEG scan sequence. (jpeg_simple_progression creates |
953 | a suitable scan definition array for progressive JPEG.) This is | 953 | a suitable scan definition array for progressive JPEG.) This is |
954 | discussed further under "Progressive JPEG support". | 954 | discussed further under "Progressive JPEG support". |
955 | 955 | ||
956 | boolean do_fancy_downsampling | 956 | boolean do_fancy_downsampling |
957 | If TRUE, use direct DCT scaling with DCT size > 8 for downsampling | 957 | If TRUE, use direct DCT scaling with DCT size > 8 for downsampling |
958 | of chroma components. | 958 | of chroma components. |
959 | If FALSE, use only DCT size <= 8 and simple separate downsampling. | 959 | If FALSE, use only DCT size <= 8 and simple separate downsampling. |
960 | Default is TRUE. | 960 | Default is TRUE. |
961 | For better image stability in multiple generation compression cycles | 961 | For better image stability in multiple generation compression cycles |
962 | it is preferable that this value matches the corresponding | 962 | it is preferable that this value matches the corresponding |
963 | do_fancy_upsampling value in decompression. | 963 | do_fancy_upsampling value in decompression. |
964 | 964 | ||
965 | int smoothing_factor | 965 | int smoothing_factor |
966 | If non-zero, the input image is smoothed; the value should be 1 for | 966 | If non-zero, the input image is smoothed; the value should be 1 for |
967 | minimal smoothing to 100 for maximum smoothing. Consult jcsample.c | 967 | minimal smoothing to 100 for maximum smoothing. Consult jcsample.c |
968 | for details of the smoothing algorithm. The default is zero. | 968 | for details of the smoothing algorithm. The default is zero. |
969 | 969 | ||
970 | boolean write_JFIF_header | 970 | boolean write_JFIF_header |
971 | If TRUE, a JFIF APP0 marker is emitted. jpeg_set_defaults() and | 971 | If TRUE, a JFIF APP0 marker is emitted. jpeg_set_defaults() and |
972 | jpeg_set_colorspace() set this TRUE if a JFIF-legal JPEG color space | 972 | jpeg_set_colorspace() set this TRUE if a JFIF-legal JPEG color space |
973 | (ie, YCbCr or grayscale) is selected, otherwise FALSE. | 973 | (ie, YCbCr or grayscale) is selected, otherwise FALSE. |
974 | 974 | ||
975 | UINT8 JFIF_major_version | 975 | UINT8 JFIF_major_version |
976 | UINT8 JFIF_minor_version | 976 | UINT8 JFIF_minor_version |
977 | The version number to be written into the JFIF marker. | 977 | The version number to be written into the JFIF marker. |
978 | jpeg_set_defaults() initializes the version to 1.01 (major=minor=1). | 978 | jpeg_set_defaults() initializes the version to 1.01 (major=minor=1). |
979 | You should set it to 1.02 (major=1, minor=2) if you plan to write | 979 | You should set it to 1.02 (major=1, minor=2) if you plan to write |
980 | any JFIF 1.02 extension markers. | 980 | any JFIF 1.02 extension markers. |
981 | 981 | ||
982 | UINT8 density_unit | 982 | UINT8 density_unit |
983 | UINT16 X_density | 983 | UINT16 X_density |
984 | UINT16 Y_density | 984 | UINT16 Y_density |
985 | The resolution information to be written into the JFIF marker; | 985 | The resolution information to be written into the JFIF marker; |
986 | not used otherwise. density_unit may be 0 for unknown, | 986 | not used otherwise. density_unit may be 0 for unknown, |
987 | 1 for dots/inch, or 2 for dots/cm. The default values are 0,1,1 | 987 | 1 for dots/inch, or 2 for dots/cm. The default values are 0,1,1 |
988 | indicating square pixels of unknown size. | 988 | indicating square pixels of unknown size. |
989 | 989 | ||
990 | boolean write_Adobe_marker | 990 | boolean write_Adobe_marker |
991 | If TRUE, an Adobe APP14 marker is emitted. jpeg_set_defaults() and | 991 | If TRUE, an Adobe APP14 marker is emitted. jpeg_set_defaults() and |
992 | jpeg_set_colorspace() set this TRUE if JPEG color space RGB, CMYK, | 992 | jpeg_set_colorspace() set this TRUE if JPEG color space RGB, CMYK, |
993 | or YCCK is selected, otherwise FALSE. It is generally a bad idea | 993 | or YCCK is selected, otherwise FALSE. It is generally a bad idea |
994 | to set both write_JFIF_header and write_Adobe_marker. In fact, | 994 | to set both write_JFIF_header and write_Adobe_marker. In fact, |
995 | you probably shouldn't change the default settings at all --- the | 995 | you probably shouldn't change the default settings at all --- the |
996 | default behavior ensures that the JPEG file's color space can be | 996 | default behavior ensures that the JPEG file's color space can be |
997 | recognized by the decoder. | 997 | recognized by the decoder. |
998 | 998 | ||
999 | JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS] | 999 | JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS] |
1000 | Pointers to coefficient quantization tables, one per table slot, | 1000 | Pointers to coefficient quantization tables, one per table slot, |
1001 | or NULL if no table is defined for a slot. Usually these should | 1001 | or NULL if no table is defined for a slot. Usually these should |
1002 | be set via one of the above helper routines; jpeg_add_quant_table() | 1002 | be set via one of the above helper routines; jpeg_add_quant_table() |
1003 | is general enough to define any quantization table. The other | 1003 | is general enough to define any quantization table. The other |
1004 | routines will set up table slot 0 for luminance quality and table | 1004 | routines will set up table slot 0 for luminance quality and table |
1005 | slot 1 for chrominance. | 1005 | slot 1 for chrominance. |
1006 | 1006 | ||
1007 | int q_scale_factor[NUM_QUANT_TBLS] | 1007 | int q_scale_factor[NUM_QUANT_TBLS] |
1008 | Linear quantization scaling factors (percentage, initialized 100) | 1008 | Linear quantization scaling factors (percentage, initialized 100) |
1009 | for use with jpeg_default_qtables(). | 1009 | for use with jpeg_default_qtables(). |
1010 | See rdswitch.c and cjpeg.c for an example of usage. | 1010 | See rdswitch.c and cjpeg.c for an example of usage. |
1011 | Note that the q_scale_factor[] fields are the "linear" scales, so you | 1011 | Note that the q_scale_factor[] fields are the "linear" scales, so you |
1012 | have to convert from user-defined ratings via jpeg_quality_scaling(). | 1012 | have to convert from user-defined ratings via jpeg_quality_scaling(). |
1013 | Here is an example code which corresponds to cjpeg -quality 90,70: | 1013 | Here is an example code which corresponds to cjpeg -quality 90,70: |
1014 | 1014 | ||
1015 | jpeg_set_defaults(cinfo); | 1015 | jpeg_set_defaults(cinfo); |
1016 | 1016 | ||
1017 | /* Set luminance quality 90. */ | 1017 | /* Set luminance quality 90. */ |
1018 | cinfo->q_scale_factor[0] = jpeg_quality_scaling(90); | 1018 | cinfo->q_scale_factor[0] = jpeg_quality_scaling(90); |
1019 | /* Set chrominance quality 70. */ | 1019 | /* Set chrominance quality 70. */ |
1020 | cinfo->q_scale_factor[1] = jpeg_quality_scaling(70); | 1020 | cinfo->q_scale_factor[1] = jpeg_quality_scaling(70); |
1021 | 1021 | ||
1022 | jpeg_default_qtables(cinfo, force_baseline); | 1022 | jpeg_default_qtables(cinfo, force_baseline); |
1023 | 1023 | ||
1024 | CAUTION: You must also set 1x1 subsampling for efficient separate | 1024 | CAUTION: You must also set 1x1 subsampling for efficient separate |
1025 | color quality selection, since the default value used by library | 1025 | color quality selection, since the default value used by library |
1026 | is 2x2: | 1026 | is 2x2: |
1027 | 1027 | ||
1028 | cinfo->comp_info[0].v_samp_factor = 1; | 1028 | cinfo->comp_info[0].v_samp_factor = 1; |
1029 | cinfo->comp_info[0].h_samp_factor = 1; | 1029 | cinfo->comp_info[0].h_samp_factor = 1; |
1030 | 1030 | ||
1031 | JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS] | 1031 | JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS] |
1032 | JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS] | 1032 | JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS] |
1033 | Pointers to Huffman coding tables, one per table slot, or NULL if | 1033 | Pointers to Huffman coding tables, one per table slot, or NULL if |
1034 | no table is defined for a slot. Slots 0 and 1 are filled with the | 1034 | no table is defined for a slot. Slots 0 and 1 are filled with the |
1035 | JPEG sample tables by jpeg_set_defaults(). If you need to allocate | 1035 | JPEG sample tables by jpeg_set_defaults(). If you need to allocate |
1036 | more table structures, jpeg_alloc_huff_table() may be used. | 1036 | more table structures, jpeg_alloc_huff_table() may be used. |
1037 | Note that optimal Huffman tables can be computed for an image | 1037 | Note that optimal Huffman tables can be computed for an image |
1038 | by setting optimize_coding, as discussed above; there's seldom | 1038 | by setting optimize_coding, as discussed above; there's seldom |
1039 | any need to mess with providing your own Huffman tables. | 1039 | any need to mess with providing your own Huffman tables. |
1040 | 1040 | ||
1041 | 1041 | ||
1042 | The actual dimensions of the JPEG image that will be written to the file are | 1042 | The actual dimensions of the JPEG image that will be written to the file are |
1043 | given by the following fields. These are computed from the input image | 1043 | given by the following fields. These are computed from the input image |
1044 | dimensions and the compression parameters by jpeg_start_compress(). You can | 1044 | dimensions and the compression parameters by jpeg_start_compress(). You can |
1045 | also call jpeg_calc_jpeg_dimensions() to obtain the values that will result | 1045 | also call jpeg_calc_jpeg_dimensions() to obtain the values that will result |
1046 | from the current parameter settings. This can be useful if you are trying | 1046 | from the current parameter settings. This can be useful if you are trying |
1047 | to pick a scaling ratio that will get close to a desired target size. | 1047 | to pick a scaling ratio that will get close to a desired target size. |
1048 | 1048 | ||
1049 | JDIMENSION jpeg_width Actual dimensions of output image. | 1049 | JDIMENSION jpeg_width Actual dimensions of output image. |
1050 | JDIMENSION jpeg_height | 1050 | JDIMENSION jpeg_height |
1051 | 1051 | ||
1052 | 1052 | ||
1053 | Per-component parameters are stored in the struct cinfo.comp_info[i] for | 1053 | Per-component parameters are stored in the struct cinfo.comp_info[i] for |
1054 | component number i. Note that components here refer to components of the | 1054 | component number i. Note that components here refer to components of the |
1055 | JPEG color space, *not* the source image color space. A suitably large | 1055 | JPEG color space, *not* the source image color space. A suitably large |
1056 | comp_info[] array is allocated by jpeg_set_defaults(); if you choose not | 1056 | comp_info[] array is allocated by jpeg_set_defaults(); if you choose not |
1057 | to use that routine, it's up to you to allocate the array. | 1057 | to use that routine, it's up to you to allocate the array. |
1058 | 1058 | ||
1059 | int component_id | 1059 | int component_id |
1060 | The one-byte identifier code to be recorded in the JPEG file for | 1060 | The one-byte identifier code to be recorded in the JPEG file for |
1061 | this component. For the standard color spaces, we recommend you | 1061 | this component. For the standard color spaces, we recommend you |
1062 | leave the default values alone. | 1062 | leave the default values alone. |
1063 | 1063 | ||
1064 | int h_samp_factor | 1064 | int h_samp_factor |
1065 | int v_samp_factor | 1065 | int v_samp_factor |
1066 | Horizontal and vertical sampling factors for the component; must | 1066 | Horizontal and vertical sampling factors for the component; must |
1067 | be 1..4 according to the JPEG standard. Note that larger sampling | 1067 | be 1..4 according to the JPEG standard. Note that larger sampling |
1068 | factors indicate a higher-resolution component; many people find | 1068 | factors indicate a higher-resolution component; many people find |
1069 | this behavior quite unintuitive. The default values are 2,2 for | 1069 | this behavior quite unintuitive. The default values are 2,2 for |
1070 | luminance components and 1,1 for chrominance components, except | 1070 | luminance components and 1,1 for chrominance components, except |
1071 | for grayscale where 1,1 is used. | 1071 | for grayscale where 1,1 is used. |
1072 | 1072 | ||
1073 | int quant_tbl_no | 1073 | int quant_tbl_no |
1074 | Quantization table number for component. The default value is | 1074 | Quantization table number for component. The default value is |
1075 | 0 for luminance components and 1 for chrominance components. | 1075 | 0 for luminance components and 1 for chrominance components. |
1076 | 1076 | ||
1077 | int dc_tbl_no | 1077 | int dc_tbl_no |
1078 | int ac_tbl_no | 1078 | int ac_tbl_no |
1079 | DC and AC entropy coding table numbers. The default values are | 1079 | DC and AC entropy coding table numbers. The default values are |
1080 | 0 for luminance components and 1 for chrominance components. | 1080 | 0 for luminance components and 1 for chrominance components. |
1081 | 1081 | ||
1082 | int component_index | 1082 | int component_index |
1083 | Must equal the component's index in comp_info[]. (Beginning in | 1083 | Must equal the component's index in comp_info[]. (Beginning in |
1084 | release v6, the compressor library will fill this in automatically; | 1084 | release v6, the compressor library will fill this in automatically; |
1085 | you don't have to.) | 1085 | you don't have to.) |
1086 | 1086 | ||
1087 | 1087 | ||
1088 | Decompression parameter selection | 1088 | Decompression parameter selection |
1089 | --------------------------------- | 1089 | --------------------------------- |
1090 | 1090 | ||
1091 | Decompression parameter selection is somewhat simpler than compression | 1091 | Decompression parameter selection is somewhat simpler than compression |
1092 | parameter selection, since all of the JPEG internal parameters are | 1092 | parameter selection, since all of the JPEG internal parameters are |
1093 | recorded in the source file and need not be supplied by the application. | 1093 | recorded in the source file and need not be supplied by the application. |
1094 | (Unless you are working with abbreviated files, in which case see | 1094 | (Unless you are working with abbreviated files, in which case see |
1095 | "Abbreviated datastreams", below.) Decompression parameters control | 1095 | "Abbreviated datastreams", below.) Decompression parameters control |
1096 | the postprocessing done on the image to deliver it in a format suitable | 1096 | the postprocessing done on the image to deliver it in a format suitable |
1097 | for the application's use. Many of the parameters control speed/quality | 1097 | for the application's use. Many of the parameters control speed/quality |
1098 | tradeoffs, in which faster decompression may be obtained at the price of | 1098 | tradeoffs, in which faster decompression may be obtained at the price of |
1099 | a poorer-quality image. The defaults select the highest quality (slowest) | 1099 | a poorer-quality image. The defaults select the highest quality (slowest) |
1100 | processing. | 1100 | processing. |
1101 | 1101 | ||
1102 | The following fields in the JPEG object are set by jpeg_read_header() and | 1102 | The following fields in the JPEG object are set by jpeg_read_header() and |
1103 | may be useful to the application in choosing decompression parameters: | 1103 | may be useful to the application in choosing decompression parameters: |
1104 | 1104 | ||
1105 | JDIMENSION image_width Width and height of image | 1105 | JDIMENSION image_width Width and height of image |
1106 | JDIMENSION image_height | 1106 | JDIMENSION image_height |
1107 | int num_components Number of color components | 1107 | int num_components Number of color components |
1108 | J_COLOR_SPACE jpeg_color_space Colorspace of image | 1108 | J_COLOR_SPACE jpeg_color_space Colorspace of image |
1109 | boolean saw_JFIF_marker TRUE if a JFIF APP0 marker was seen | 1109 | boolean saw_JFIF_marker TRUE if a JFIF APP0 marker was seen |
1110 | UINT8 JFIF_major_version Version information from JFIF marker | 1110 | UINT8 JFIF_major_version Version information from JFIF marker |
1111 | UINT8 JFIF_minor_version | 1111 | UINT8 JFIF_minor_version |
1112 | UINT8 density_unit Resolution data from JFIF marker | 1112 | UINT8 density_unit Resolution data from JFIF marker |
1113 | UINT16 X_density | 1113 | UINT16 X_density |
1114 | UINT16 Y_density | 1114 | UINT16 Y_density |
1115 | boolean saw_Adobe_marker TRUE if an Adobe APP14 marker was seen | 1115 | boolean saw_Adobe_marker TRUE if an Adobe APP14 marker was seen |
1116 | UINT8 Adobe_transform Color transform code from Adobe marker | 1116 | UINT8 Adobe_transform Color transform code from Adobe marker |
1117 | 1117 | ||
1118 | The JPEG color space, unfortunately, is something of a guess since the JPEG | 1118 | The JPEG color space, unfortunately, is something of a guess since the JPEG |
1119 | standard proper does not provide a way to record it. In practice most files | 1119 | standard proper does not provide a way to record it. In practice most files |
1120 | adhere to the JFIF or Adobe conventions, and the decoder will recognize these | 1120 | adhere to the JFIF or Adobe conventions, and the decoder will recognize these |
1121 | correctly. See "Special color spaces", below, for more info. | 1121 | correctly. See "Special color spaces", below, for more info. |
1122 | 1122 | ||
1123 | 1123 | ||
1124 | The decompression parameters that determine the basic properties of the | 1124 | The decompression parameters that determine the basic properties of the |
1125 | returned image are: | 1125 | returned image are: |
1126 | 1126 | ||
1127 | J_COLOR_SPACE out_color_space | 1127 | J_COLOR_SPACE out_color_space |
1128 | Output color space. jpeg_read_header() sets an appropriate default | 1128 | Output color space. jpeg_read_header() sets an appropriate default |
1129 | based on jpeg_color_space; typically it will be RGB or grayscale. | 1129 | based on jpeg_color_space; typically it will be RGB or grayscale. |
1130 | The application can change this field to request output in a different | 1130 | The application can change this field to request output in a different |
1131 | colorspace. For example, set it to JCS_GRAYSCALE to get grayscale | 1131 | colorspace. For example, set it to JCS_GRAYSCALE to get grayscale |
1132 | output from a color file. (This is useful for previewing: grayscale | 1132 | output from a color file. (This is useful for previewing: grayscale |
1133 | output is faster than full color since the color components need not | 1133 | output is faster than full color since the color components need not |
1134 | be processed.) Note that not all possible color space transforms are | 1134 | be processed.) Note that not all possible color space transforms are |
1135 | currently implemented; you may need to extend jdcolor.c if you want an | 1135 | currently implemented; you may need to extend jdcolor.c if you want an |
1136 | unusual conversion. | 1136 | unusual conversion. |
1137 | 1137 | ||
1138 | unsigned int scale_num, scale_denom | 1138 | unsigned int scale_num, scale_denom |
1139 | Scale the image by the fraction scale_num/scale_denom. Currently, | 1139 | Scale the image by the fraction scale_num/scale_denom. Currently, |
1140 | the supported scaling ratios are M/N with all M from 1 to 16, where | 1140 | the supported scaling ratios are M/N with all M from 1 to 16, where |
1141 | N is the source DCT size, which is 8 for baseline JPEG. (The library | 1141 | N is the source DCT size, which is 8 for baseline JPEG. (The library |
1142 | design allows for arbitrary scaling ratios but this is not likely | 1142 | design allows for arbitrary scaling ratios but this is not likely |
1143 | to be implemented any time soon.) The values are initialized by | 1143 | to be implemented any time soon.) The values are initialized by |
1144 | jpeg_read_header() with the source DCT size. For baseline JPEG | 1144 | jpeg_read_header() with the source DCT size. For baseline JPEG |
1145 | this is 8/8. If you change only the scale_num value while leaving | 1145 | this is 8/8. If you change only the scale_num value while leaving |
1146 | the other unchanged, then this specifies the DCT scaled size to be | 1146 | the other unchanged, then this specifies the DCT scaled size to be |
1147 | applied on the given input. For baseline JPEG this is equivalent | 1147 | applied on the given input. For baseline JPEG this is equivalent |
1148 | to M/8 scaling, since the source DCT size for baseline JPEG is 8. | 1148 | to M/8 scaling, since the source DCT size for baseline JPEG is 8. |
1149 | Smaller scaling ratios permit significantly faster decoding since | 1149 | Smaller scaling ratios permit significantly faster decoding since |
1150 | fewer pixels need be processed and a simpler IDCT method can be used. | 1150 | fewer pixels need be processed and a simpler IDCT method can be used. |
1151 | 1151 | ||
1152 | boolean quantize_colors | 1152 | boolean quantize_colors |
1153 | If set TRUE, colormapped output will be delivered. Default is FALSE, | 1153 | If set TRUE, colormapped output will be delivered. Default is FALSE, |
1154 | meaning that full-color output will be delivered. | 1154 | meaning that full-color output will be delivered. |
1155 | 1155 | ||
1156 | The next three parameters are relevant only if quantize_colors is TRUE. | 1156 | The next three parameters are relevant only if quantize_colors is TRUE. |
1157 | 1157 | ||
1158 | int desired_number_of_colors | 1158 | int desired_number_of_colors |
1159 | Maximum number of colors to use in generating a library-supplied color | 1159 | Maximum number of colors to use in generating a library-supplied color |
1160 | map (the actual number of colors is returned in a different field). | 1160 | map (the actual number of colors is returned in a different field). |
1161 | Default 256. Ignored when the application supplies its own color map. | 1161 | Default 256. Ignored when the application supplies its own color map. |
1162 | 1162 | ||
1163 | boolean two_pass_quantize | 1163 | boolean two_pass_quantize |
1164 | If TRUE, an extra pass over the image is made to select a custom color | 1164 | If TRUE, an extra pass over the image is made to select a custom color |
1165 | map for the image. This usually looks a lot better than the one-size- | 1165 | map for the image. This usually looks a lot better than the one-size- |
1166 | fits-all colormap that is used otherwise. Default is TRUE. Ignored | 1166 | fits-all colormap that is used otherwise. Default is TRUE. Ignored |
1167 | when the application supplies its own color map. | 1167 | when the application supplies its own color map. |
1168 | 1168 | ||
1169 | J_DITHER_MODE dither_mode | 1169 | J_DITHER_MODE dither_mode |
1170 | Selects color dithering method. Supported values are: | 1170 | Selects color dithering method. Supported values are: |
1171 | JDITHER_NONE no dithering: fast, very low quality | 1171 | JDITHER_NONE no dithering: fast, very low quality |
1172 | JDITHER_ORDERED ordered dither: moderate speed and quality | 1172 | JDITHER_ORDERED ordered dither: moderate speed and quality |
1173 | JDITHER_FS Floyd-Steinberg dither: slow, high quality | 1173 | JDITHER_FS Floyd-Steinberg dither: slow, high quality |
1174 | Default is JDITHER_FS. (At present, ordered dither is implemented | 1174 | Default is JDITHER_FS. (At present, ordered dither is implemented |
1175 | only in the single-pass, standard-colormap case. If you ask for | 1175 | only in the single-pass, standard-colormap case. If you ask for |
1176 | ordered dither when two_pass_quantize is TRUE or when you supply | 1176 | ordered dither when two_pass_quantize is TRUE or when you supply |
1177 | an external color map, you'll get F-S dithering.) | 1177 | an external color map, you'll get F-S dithering.) |
1178 | 1178 | ||
1179 | When quantize_colors is TRUE, the target color map is described by the next | 1179 | When quantize_colors is TRUE, the target color map is described by the next |
1180 | two fields. colormap is set to NULL by jpeg_read_header(). The application | 1180 | two fields. colormap is set to NULL by jpeg_read_header(). The application |
1181 | can supply a color map by setting colormap non-NULL and setting | 1181 | can supply a color map by setting colormap non-NULL and setting |
1182 | actual_number_of_colors to the map size. Otherwise, jpeg_start_decompress() | 1182 | actual_number_of_colors to the map size. Otherwise, jpeg_start_decompress() |
1183 | selects a suitable color map and sets these two fields itself. | 1183 | selects a suitable color map and sets these two fields itself. |
1184 | [Implementation restriction: at present, an externally supplied colormap is | 1184 | [Implementation restriction: at present, an externally supplied colormap is |
1185 | only accepted for 3-component output color spaces.] | 1185 | only accepted for 3-component output color spaces.] |
1186 | 1186 | ||
1187 | JSAMPARRAY colormap | 1187 | JSAMPARRAY colormap |
1188 | The color map, represented as a 2-D pixel array of out_color_components | 1188 | The color map, represented as a 2-D pixel array of out_color_components |
1189 | rows and actual_number_of_colors columns. Ignored if not quantizing. | 1189 | rows and actual_number_of_colors columns. Ignored if not quantizing. |
1190 | CAUTION: if the JPEG library creates its own colormap, the storage | 1190 | CAUTION: if the JPEG library creates its own colormap, the storage |
1191 | pointed to by this field is released by jpeg_finish_decompress(). | 1191 | pointed to by this field is released by jpeg_finish_decompress(). |
1192 | Copy the colormap somewhere else first, if you want to save it. | 1192 | Copy the colormap somewhere else first, if you want to save it. |
1193 | 1193 | ||
1194 | int actual_number_of_colors | 1194 | int actual_number_of_colors |
1195 | The number of colors in the color map. | 1195 | The number of colors in the color map. |
1196 | 1196 | ||
1197 | Additional decompression parameters that the application may set include: | 1197 | Additional decompression parameters that the application may set include: |
1198 | 1198 | ||
1199 | J_DCT_METHOD dct_method | 1199 | J_DCT_METHOD dct_method |
1200 | Selects the algorithm used for the DCT step. Choices are the same | 1200 | Selects the algorithm used for the DCT step. Choices are the same |
1201 | as described above for compression. | 1201 | as described above for compression. |
1202 | 1202 | ||
1203 | boolean do_fancy_upsampling | 1203 | boolean do_fancy_upsampling |
1204 | If TRUE, use direct DCT scaling with DCT size > 8 for upsampling | 1204 | If TRUE, use direct DCT scaling with DCT size > 8 for upsampling |
1205 | of chroma components. | 1205 | of chroma components. |
1206 | If FALSE, use only DCT size <= 8 and simple separate upsampling. | 1206 | If FALSE, use only DCT size <= 8 and simple separate upsampling. |
1207 | Default is TRUE. | 1207 | Default is TRUE. |
1208 | For better image stability in multiple generation compression cycles | 1208 | For better image stability in multiple generation compression cycles |
1209 | it is preferable that this value matches the corresponding | 1209 | it is preferable that this value matches the corresponding |
1210 | do_fancy_downsampling value in compression. | 1210 | do_fancy_downsampling value in compression. |
1211 | 1211 | ||
1212 | boolean do_block_smoothing | 1212 | boolean do_block_smoothing |
1213 | If TRUE, interblock smoothing is applied in early stages of decoding | 1213 | If TRUE, interblock smoothing is applied in early stages of decoding |
1214 | progressive JPEG files; if FALSE, not. Default is TRUE. Early | 1214 | progressive JPEG files; if FALSE, not. Default is TRUE. Early |
1215 | progression stages look "fuzzy" with smoothing, "blocky" without. | 1215 | progression stages look "fuzzy" with smoothing, "blocky" without. |
1216 | In any case, block smoothing ceases to be applied after the first few | 1216 | In any case, block smoothing ceases to be applied after the first few |
1217 | AC coefficients are known to full accuracy, so it is relevant only | 1217 | AC coefficients are known to full accuracy, so it is relevant only |
1218 | when using buffered-image mode for progressive images. | 1218 | when using buffered-image mode for progressive images. |
1219 | 1219 | ||
1220 | boolean enable_1pass_quant | 1220 | boolean enable_1pass_quant |
1221 | boolean enable_external_quant | 1221 | boolean enable_external_quant |
1222 | boolean enable_2pass_quant | 1222 | boolean enable_2pass_quant |
1223 | These are significant only in buffered-image mode, which is | 1223 | These are significant only in buffered-image mode, which is |
1224 | described in its own section below. | 1224 | described in its own section below. |
1225 | 1225 | ||
1226 | 1226 | ||
1227 | The output image dimensions are given by the following fields. These are | 1227 | The output image dimensions are given by the following fields. These are |
1228 | computed from the source image dimensions and the decompression parameters | 1228 | computed from the source image dimensions and the decompression parameters |
1229 | by jpeg_start_decompress(). You can also call jpeg_calc_output_dimensions() | 1229 | by jpeg_start_decompress(). You can also call jpeg_calc_output_dimensions() |
1230 | to obtain the values that will result from the current parameter settings. | 1230 | to obtain the values that will result from the current parameter settings. |
1231 | This can be useful if you are trying to pick a scaling ratio that will get | 1231 | This can be useful if you are trying to pick a scaling ratio that will get |
1232 | close to a desired target size. It's also important if you are using the | 1232 | close to a desired target size. It's also important if you are using the |
1233 | JPEG library's memory manager to allocate output buffer space, because you | 1233 | JPEG library's memory manager to allocate output buffer space, because you |
1234 | are supposed to request such buffers *before* jpeg_start_decompress(). | 1234 | are supposed to request such buffers *before* jpeg_start_decompress(). |
1235 | 1235 | ||
1236 | JDIMENSION output_width Actual dimensions of output image. | 1236 | JDIMENSION output_width Actual dimensions of output image. |
1237 | JDIMENSION output_height | 1237 | JDIMENSION output_height |
1238 | int out_color_components Number of color components in out_color_space. | 1238 | int out_color_components Number of color components in out_color_space. |
1239 | int output_components Number of color components returned. | 1239 | int output_components Number of color components returned. |
1240 | int rec_outbuf_height Recommended height of scanline buffer. | 1240 | int rec_outbuf_height Recommended height of scanline buffer. |
1241 | 1241 | ||
1242 | When quantizing colors, output_components is 1, indicating a single color map | 1242 | When quantizing colors, output_components is 1, indicating a single color map |
1243 | index per pixel. Otherwise it equals out_color_components. The output arrays | 1243 | index per pixel. Otherwise it equals out_color_components. The output arrays |
1244 | are required to be output_width * output_components JSAMPLEs wide. | 1244 | are required to be output_width * output_components JSAMPLEs wide. |
1245 | 1245 | ||
1246 | rec_outbuf_height is the recommended minimum height (in scanlines) of the | 1246 | rec_outbuf_height is the recommended minimum height (in scanlines) of the |
1247 | buffer passed to jpeg_read_scanlines(). If the buffer is smaller, the | 1247 | buffer passed to jpeg_read_scanlines(). If the buffer is smaller, the |
1248 | library will still work, but time will be wasted due to unnecessary data | 1248 | library will still work, but time will be wasted due to unnecessary data |
1249 | copying. In high-quality modes, rec_outbuf_height is always 1, but some | 1249 | copying. In high-quality modes, rec_outbuf_height is always 1, but some |
1250 | faster, lower-quality modes set it to larger values (typically 2 to 4). | 1250 | faster, lower-quality modes set it to larger values (typically 2 to 4). |
1251 | If you are going to ask for a high-speed processing mode, you may as well | 1251 | If you are going to ask for a high-speed processing mode, you may as well |
1252 | go to the trouble of honoring rec_outbuf_height so as to avoid data copying. | 1252 | go to the trouble of honoring rec_outbuf_height so as to avoid data copying. |
1253 | (An output buffer larger than rec_outbuf_height lines is OK, but won't | 1253 | (An output buffer larger than rec_outbuf_height lines is OK, but won't |
1254 | provide any material speed improvement over that height.) | 1254 | provide any material speed improvement over that height.) |
1255 | 1255 | ||
1256 | 1256 | ||
1257 | Special color spaces | 1257 | Special color spaces |
1258 | -------------------- | 1258 | -------------------- |
1259 | 1259 | ||
1260 | The JPEG standard itself is "color blind" and doesn't specify any particular | 1260 | The JPEG standard itself is "color blind" and doesn't specify any particular |
1261 | color space. It is customary to convert color data to a luminance/chrominance | 1261 | color space. It is customary to convert color data to a luminance/chrominance |
1262 | color space before compressing, since this permits greater compression. The | 1262 | color space before compressing, since this permits greater compression. The |
1263 | existing de-facto JPEG file format standards specify YCbCr or grayscale data | 1263 | existing de-facto JPEG file format standards specify YCbCr or grayscale data |
1264 | (JFIF), or grayscale, RGB, YCbCr, CMYK, or YCCK (Adobe). For special | 1264 | (JFIF), or grayscale, RGB, YCbCr, CMYK, or YCCK (Adobe). For special |
1265 | applications such as multispectral images, other color spaces can be used, | 1265 | applications such as multispectral images, other color spaces can be used, |
1266 | but it must be understood that such files will be unportable. | 1266 | but it must be understood that such files will be unportable. |
1267 | 1267 | ||
1268 | The JPEG library can handle the most common colorspace conversions (namely | 1268 | The JPEG library can handle the most common colorspace conversions (namely |
1269 | RGB <=> YCbCr and CMYK <=> YCCK). It can also deal with data of an unknown | 1269 | RGB <=> YCbCr and CMYK <=> YCCK). It can also deal with data of an unknown |
1270 | color space, passing it through without conversion. If you deal extensively | 1270 | color space, passing it through without conversion. If you deal extensively |
1271 | with an unusual color space, you can easily extend the library to understand | 1271 | with an unusual color space, you can easily extend the library to understand |
1272 | additional color spaces and perform appropriate conversions. | 1272 | additional color spaces and perform appropriate conversions. |
1273 | 1273 | ||
1274 | For compression, the source data's color space is specified by field | 1274 | For compression, the source data's color space is specified by field |
1275 | in_color_space. This is transformed to the JPEG file's color space given | 1275 | in_color_space. This is transformed to the JPEG file's color space given |
1276 | by jpeg_color_space. jpeg_set_defaults() chooses a reasonable JPEG color | 1276 | by jpeg_color_space. jpeg_set_defaults() chooses a reasonable JPEG color |
1277 | space depending on in_color_space, but you can override this by calling | 1277 | space depending on in_color_space, but you can override this by calling |
1278 | jpeg_set_colorspace(). Of course you must select a supported transformation. | 1278 | jpeg_set_colorspace(). Of course you must select a supported transformation. |
1279 | jccolor.c currently supports the following transformations: | 1279 | jccolor.c currently supports the following transformations: |
1280 | RGB => YCbCr | 1280 | RGB => YCbCr |
1281 | RGB => GRAYSCALE | 1281 | RGB => GRAYSCALE |
1282 | YCbCr => GRAYSCALE | 1282 | YCbCr => GRAYSCALE |
1283 | CMYK => YCCK | 1283 | CMYK => YCCK |
1284 | plus the null transforms: GRAYSCALE => GRAYSCALE, RGB => RGB, | 1284 | plus the null transforms: GRAYSCALE => GRAYSCALE, RGB => RGB, |
1285 | YCbCr => YCbCr, CMYK => CMYK, YCCK => YCCK, and UNKNOWN => UNKNOWN. | 1285 | YCbCr => YCbCr, CMYK => CMYK, YCCK => YCCK, and UNKNOWN => UNKNOWN. |
1286 | 1286 | ||
1287 | The de-facto file format standards (JFIF and Adobe) specify APPn markers that | 1287 | The de-facto file format standards (JFIF and Adobe) specify APPn markers that |
1288 | indicate the color space of the JPEG file. It is important to ensure that | 1288 | indicate the color space of the JPEG file. It is important to ensure that |
1289 | these are written correctly, or omitted if the JPEG file's color space is not | 1289 | these are written correctly, or omitted if the JPEG file's color space is not |
1290 | one of the ones supported by the de-facto standards. jpeg_set_colorspace() | 1290 | one of the ones supported by the de-facto standards. jpeg_set_colorspace() |
1291 | will set the compression parameters to include or omit the APPn markers | 1291 | will set the compression parameters to include or omit the APPn markers |
1292 | properly, so long as it is told the truth about the JPEG color space. | 1292 | properly, so long as it is told the truth about the JPEG color space. |
1293 | For example, if you are writing some random 3-component color space without | 1293 | For example, if you are writing some random 3-component color space without |
1294 | conversion, don't try to fake out the library by setting in_color_space and | 1294 | conversion, don't try to fake out the library by setting in_color_space and |
1295 | jpeg_color_space to JCS_YCbCr; use JCS_UNKNOWN. You may want to write an | 1295 | jpeg_color_space to JCS_YCbCr; use JCS_UNKNOWN. You may want to write an |
1296 | APPn marker of your own devising to identify the colorspace --- see "Special | 1296 | APPn marker of your own devising to identify the colorspace --- see "Special |
1297 | markers", below. | 1297 | markers", below. |
1298 | 1298 | ||
1299 | When told that the color space is UNKNOWN, the library will default to using | 1299 | When told that the color space is UNKNOWN, the library will default to using |
1300 | luminance-quality compression parameters for all color components. You may | 1300 | luminance-quality compression parameters for all color components. You may |
1301 | well want to change these parameters. See the source code for | 1301 | well want to change these parameters. See the source code for |
1302 | jpeg_set_colorspace(), in jcparam.c, for details. | 1302 | jpeg_set_colorspace(), in jcparam.c, for details. |
1303 | 1303 | ||
1304 | For decompression, the JPEG file's color space is given in jpeg_color_space, | 1304 | For decompression, the JPEG file's color space is given in jpeg_color_space, |
1305 | and this is transformed to the output color space out_color_space. | 1305 | and this is transformed to the output color space out_color_space. |
1306 | jpeg_read_header's setting of jpeg_color_space can be relied on if the file | 1306 | jpeg_read_header's setting of jpeg_color_space can be relied on if the file |
1307 | conforms to JFIF or Adobe conventions, but otherwise it is no better than a | 1307 | conforms to JFIF or Adobe conventions, but otherwise it is no better than a |
1308 | guess. If you know the JPEG file's color space for certain, you can override | 1308 | guess. If you know the JPEG file's color space for certain, you can override |
1309 | jpeg_read_header's guess by setting jpeg_color_space. jpeg_read_header also | 1309 | jpeg_read_header's guess by setting jpeg_color_space. jpeg_read_header also |
1310 | selects a default output color space based on (its guess of) jpeg_color_space; | 1310 | selects a default output color space based on (its guess of) jpeg_color_space; |
1311 | set out_color_space to override this. Again, you must select a supported | 1311 | set out_color_space to override this. Again, you must select a supported |
1312 | transformation. jdcolor.c currently supports | 1312 | transformation. jdcolor.c currently supports |
1313 | YCbCr => RGB | 1313 | YCbCr => RGB |
1314 | YCbCr => GRAYSCALE | 1314 | YCbCr => GRAYSCALE |
1315 | RGB => GRAYSCALE | 1315 | RGB => GRAYSCALE |
1316 | GRAYSCALE => RGB | 1316 | GRAYSCALE => RGB |
1317 | YCCK => CMYK | 1317 | YCCK => CMYK |
1318 | as well as the null transforms. (Since GRAYSCALE=>RGB is provided, an | 1318 | as well as the null transforms. (Since GRAYSCALE=>RGB is provided, an |
1319 | application can force grayscale JPEGs to look like color JPEGs if it only | 1319 | application can force grayscale JPEGs to look like color JPEGs if it only |
1320 | wants to handle one case.) | 1320 | wants to handle one case.) |
1321 | 1321 | ||
1322 | The two-pass color quantizer, jquant2.c, is specialized to handle RGB data | 1322 | The two-pass color quantizer, jquant2.c, is specialized to handle RGB data |
1323 | (it weights distances appropriately for RGB colors). You'll need to modify | 1323 | (it weights distances appropriately for RGB colors). You'll need to modify |
1324 | the code if you want to use it for non-RGB output color spaces. Note that | 1324 | the code if you want to use it for non-RGB output color spaces. Note that |
1325 | jquant2.c is used to map to an application-supplied colormap as well as for | 1325 | jquant2.c is used to map to an application-supplied colormap as well as for |
1326 | the normal two-pass colormap selection process. | 1326 | the normal two-pass colormap selection process. |
1327 | 1327 | ||
1328 | CAUTION: it appears that Adobe Photoshop writes inverted data in CMYK JPEG | 1328 | CAUTION: it appears that Adobe Photoshop writes inverted data in CMYK JPEG |
1329 | files: 0 represents 100% ink coverage, rather than 0% ink as you'd expect. | 1329 | files: 0 represents 100% ink coverage, rather than 0% ink as you'd expect. |
1330 | This is arguably a bug in Photoshop, but if you need to work with Photoshop | 1330 | This is arguably a bug in Photoshop, but if you need to work with Photoshop |
1331 | CMYK files, you will have to deal with it in your application. We cannot | 1331 | CMYK files, you will have to deal with it in your application. We cannot |
1332 | "fix" this in the library by inverting the data during the CMYK<=>YCCK | 1332 | "fix" this in the library by inverting the data during the CMYK<=>YCCK |
1333 | transform, because that would break other applications, notably Ghostscript. | 1333 | transform, because that would break other applications, notably Ghostscript. |
1334 | Photoshop versions prior to 3.0 write EPS files containing JPEG-encoded CMYK | 1334 | Photoshop versions prior to 3.0 write EPS files containing JPEG-encoded CMYK |
1335 | data in the same inverted-YCCK representation used in bare JPEG files, but | 1335 | data in the same inverted-YCCK representation used in bare JPEG files, but |
1336 | the surrounding PostScript code performs an inversion using the PS image | 1336 | the surrounding PostScript code performs an inversion using the PS image |
1337 | operator. I am told that Photoshop 3.0 will write uninverted YCCK in | 1337 | operator. I am told that Photoshop 3.0 will write uninverted YCCK in |
1338 | EPS/JPEG files, and will omit the PS-level inversion. (But the data | 1338 | EPS/JPEG files, and will omit the PS-level inversion. (But the data |
1339 | polarity used in bare JPEG files will not change in 3.0.) In either case, | 1339 | polarity used in bare JPEG files will not change in 3.0.) In either case, |
1340 | the JPEG library must not invert the data itself, or else Ghostscript would | 1340 | the JPEG library must not invert the data itself, or else Ghostscript would |
1341 | read these EPS files incorrectly. | 1341 | read these EPS files incorrectly. |
1342 | 1342 | ||
1343 | 1343 | ||
1344 | Error handling | 1344 | Error handling |
1345 | -------------- | 1345 | -------------- |
1346 | 1346 | ||
1347 | When the default error handler is used, any error detected inside the JPEG | 1347 | When the default error handler is used, any error detected inside the JPEG |
1348 | routines will cause a message to be printed on stderr, followed by exit(). | 1348 | routines will cause a message to be printed on stderr, followed by exit(). |
1349 | You can supply your own error handling routines to override this behavior | 1349 | You can supply your own error handling routines to override this behavior |
1350 | and to control the treatment of nonfatal warnings and trace/debug messages. | 1350 | and to control the treatment of nonfatal warnings and trace/debug messages. |
1351 | The file example.c illustrates the most common case, which is to have the | 1351 | The file example.c illustrates the most common case, which is to have the |
1352 | application regain control after an error rather than exiting. | 1352 | application regain control after an error rather than exiting. |
1353 | 1353 | ||
1354 | The JPEG library never writes any message directly; it always goes through | 1354 | The JPEG library never writes any message directly; it always goes through |
1355 | the error handling routines. Three classes of messages are recognized: | 1355 | the error handling routines. Three classes of messages are recognized: |
1356 | * Fatal errors: the library cannot continue. | 1356 | * Fatal errors: the library cannot continue. |
1357 | * Warnings: the library can continue, but the data is corrupt, and a | 1357 | * Warnings: the library can continue, but the data is corrupt, and a |
1358 | damaged output image is likely to result. | 1358 | damaged output image is likely to result. |
1359 | * Trace/informational messages. These come with a trace level indicating | 1359 | * Trace/informational messages. These come with a trace level indicating |
1360 | the importance of the message; you can control the verbosity of the | 1360 | the importance of the message; you can control the verbosity of the |
1361 | program by adjusting the maximum trace level that will be displayed. | 1361 | program by adjusting the maximum trace level that will be displayed. |
1362 | 1362 | ||
1363 | You may, if you wish, simply replace the entire JPEG error handling module | 1363 | You may, if you wish, simply replace the entire JPEG error handling module |
1364 | (jerror.c) with your own code. However, you can avoid code duplication by | 1364 | (jerror.c) with your own code. However, you can avoid code duplication by |
1365 | only replacing some of the routines depending on the behavior you need. | 1365 | only replacing some of the routines depending on the behavior you need. |
1366 | This is accomplished by calling jpeg_std_error() as usual, but then overriding | 1366 | This is accomplished by calling jpeg_std_error() as usual, but then overriding |
1367 | some of the method pointers in the jpeg_error_mgr struct, as illustrated by | 1367 | some of the method pointers in the jpeg_error_mgr struct, as illustrated by |
1368 | example.c. | 1368 | example.c. |
1369 | 1369 | ||
1370 | All of the error handling routines will receive a pointer to the JPEG object | 1370 | All of the error handling routines will receive a pointer to the JPEG object |
1371 | (a j_common_ptr which points to either a jpeg_compress_struct or a | 1371 | (a j_common_ptr which points to either a jpeg_compress_struct or a |
1372 | jpeg_decompress_struct; if you need to tell which, test the is_decompressor | 1372 | jpeg_decompress_struct; if you need to tell which, test the is_decompressor |
1373 | field). This struct includes a pointer to the error manager struct in its | 1373 | field). This struct includes a pointer to the error manager struct in its |
1374 | "err" field. Frequently, custom error handler routines will need to access | 1374 | "err" field. Frequently, custom error handler routines will need to access |
1375 | additional data which is not known to the JPEG library or the standard error | 1375 | additional data which is not known to the JPEG library or the standard error |
1376 | handler. The most convenient way to do this is to embed either the JPEG | 1376 | handler. The most convenient way to do this is to embed either the JPEG |
1377 | object or the jpeg_error_mgr struct in a larger structure that contains | 1377 | object or the jpeg_error_mgr struct in a larger structure that contains |
1378 | additional fields; then casting the passed pointer provides access to the | 1378 | additional fields; then casting the passed pointer provides access to the |
1379 | additional fields. Again, see example.c for one way to do it. (Beginning | 1379 | additional fields. Again, see example.c for one way to do it. (Beginning |
1380 | with IJG version 6b, there is also a void pointer "client_data" in each | 1380 | with IJG version 6b, there is also a void pointer "client_data" in each |
1381 | JPEG object, which the application can also use to find related data. | 1381 | JPEG object, which the application can also use to find related data. |
1382 | The library does not touch client_data at all.) | 1382 | The library does not touch client_data at all.) |
1383 | 1383 | ||
1384 | The individual methods that you might wish to override are: | 1384 | The individual methods that you might wish to override are: |
1385 | 1385 | ||
1386 | error_exit (j_common_ptr cinfo) | 1386 | error_exit (j_common_ptr cinfo) |
1387 | Receives control for a fatal error. Information sufficient to | 1387 | Receives control for a fatal error. Information sufficient to |
1388 | generate the error message has been stored in cinfo->err; call | 1388 | generate the error message has been stored in cinfo->err; call |
1389 | output_message to display it. Control must NOT return to the caller; | 1389 | output_message to display it. Control must NOT return to the caller; |
1390 | generally this routine will exit() or longjmp() somewhere. | 1390 | generally this routine will exit() or longjmp() somewhere. |
1391 | Typically you would override this routine to get rid of the exit() | 1391 | Typically you would override this routine to get rid of the exit() |
1392 | default behavior. Note that if you continue processing, you should | 1392 | default behavior. Note that if you continue processing, you should |
1393 | clean up the JPEG object with jpeg_abort() or jpeg_destroy(). | 1393 | clean up the JPEG object with jpeg_abort() or jpeg_destroy(). |
1394 | 1394 | ||
1395 | output_message (j_common_ptr cinfo) | 1395 | output_message (j_common_ptr cinfo) |
1396 | Actual output of any JPEG message. Override this to send messages | 1396 | Actual output of any JPEG message. Override this to send messages |
1397 | somewhere other than stderr. Note that this method does not know | 1397 | somewhere other than stderr. Note that this method does not know |
1398 | how to generate a message, only where to send it. | 1398 | how to generate a message, only where to send it. |
1399 | 1399 | ||
1400 | format_message (j_common_ptr cinfo, char * buffer) | 1400 | format_message (j_common_ptr cinfo, char * buffer) |
1401 | Constructs a readable error message string based on the error info | 1401 | Constructs a readable error message string based on the error info |
1402 | stored in cinfo->err. This method is called by output_message. Few | 1402 | stored in cinfo->err. This method is called by output_message. Few |
1403 | applications should need to override this method. One possible | 1403 | applications should need to override this method. One possible |
1404 | reason for doing so is to implement dynamic switching of error message | 1404 | reason for doing so is to implement dynamic switching of error message |
1405 | language. | 1405 | language. |
1406 | 1406 | ||
1407 | emit_message (j_common_ptr cinfo, int msg_level) | 1407 | emit_message (j_common_ptr cinfo, int msg_level) |
1408 | Decide whether or not to emit a warning or trace message; if so, | 1408 | Decide whether or not to emit a warning or trace message; if so, |
1409 | calls output_message. The main reason for overriding this method | 1409 | calls output_message. The main reason for overriding this method |
1410 | would be to abort on warnings. msg_level is -1 for warnings, | 1410 | would be to abort on warnings. msg_level is -1 for warnings, |
1411 | 0 and up for trace messages. | 1411 | 0 and up for trace messages. |
1412 | 1412 | ||
1413 | Only error_exit() and emit_message() are called from the rest of the JPEG | 1413 | Only error_exit() and emit_message() are called from the rest of the JPEG |
1414 | library; the other two are internal to the error handler. | 1414 | library; the other two are internal to the error handler. |
1415 | 1415 | ||
1416 | The actual message texts are stored in an array of strings which is pointed to | 1416 | The actual message texts are stored in an array of strings which is pointed to |
1417 | by the field err->jpeg_message_table. The messages are numbered from 0 to | 1417 | by the field err->jpeg_message_table. The messages are numbered from 0 to |
1418 | err->last_jpeg_message, and it is these code numbers that are used in the | 1418 | err->last_jpeg_message, and it is these code numbers that are used in the |
1419 | JPEG library code. You could replace the message texts (for instance, with | 1419 | JPEG library code. You could replace the message texts (for instance, with |
1420 | messages in French or German) by changing the message table pointer. See | 1420 | messages in French or German) by changing the message table pointer. See |
1421 | jerror.h for the default texts. CAUTION: this table will almost certainly | 1421 | jerror.h for the default texts. CAUTION: this table will almost certainly |
1422 | change or grow from one library version to the next. | 1422 | change or grow from one library version to the next. |
1423 | 1423 | ||
1424 | It may be useful for an application to add its own message texts that are | 1424 | It may be useful for an application to add its own message texts that are |
1425 | handled by the same mechanism. The error handler supports a second "add-on" | 1425 | handled by the same mechanism. The error handler supports a second "add-on" |
1426 | message table for this purpose. To define an addon table, set the pointer | 1426 | message table for this purpose. To define an addon table, set the pointer |
1427 | err->addon_message_table and the message numbers err->first_addon_message and | 1427 | err->addon_message_table and the message numbers err->first_addon_message and |
1428 | err->last_addon_message. If you number the addon messages beginning at 1000 | 1428 | err->last_addon_message. If you number the addon messages beginning at 1000 |
1429 | or so, you won't have to worry about conflicts with the library's built-in | 1429 | or so, you won't have to worry about conflicts with the library's built-in |
1430 | messages. See the sample applications cjpeg/djpeg for an example of using | 1430 | messages. See the sample applications cjpeg/djpeg for an example of using |
1431 | addon messages (the addon messages are defined in cderror.h). | 1431 | addon messages (the addon messages are defined in cderror.h). |
1432 | 1432 | ||
1433 | Actual invocation of the error handler is done via macros defined in jerror.h: | 1433 | Actual invocation of the error handler is done via macros defined in jerror.h: |
1434 | ERREXITn(...) for fatal errors | 1434 | ERREXITn(...) for fatal errors |
1435 | WARNMSn(...) for corrupt-data warnings | 1435 | WARNMSn(...) for corrupt-data warnings |
1436 | TRACEMSn(...) for trace and informational messages. | 1436 | TRACEMSn(...) for trace and informational messages. |
1437 | These macros store the message code and any additional parameters into the | 1437 | These macros store the message code and any additional parameters into the |
1438 | error handler struct, then invoke the error_exit() or emit_message() method. | 1438 | error handler struct, then invoke the error_exit() or emit_message() method. |
1439 | The variants of each macro are for varying numbers of additional parameters. | 1439 | The variants of each macro are for varying numbers of additional parameters. |
1440 | The additional parameters are inserted into the generated message using | 1440 | The additional parameters are inserted into the generated message using |
1441 | standard printf() format codes. | 1441 | standard printf() format codes. |
1442 | 1442 | ||
1443 | See jerror.h and jerror.c for further details. | 1443 | See jerror.h and jerror.c for further details. |
1444 | 1444 | ||
1445 | 1445 | ||
1446 | Compressed data handling (source and destination managers) | 1446 | Compressed data handling (source and destination managers) |
1447 | ---------------------------------------------------------- | 1447 | ---------------------------------------------------------- |
1448 | 1448 | ||
1449 | The JPEG compression library sends its compressed data to a "destination | 1449 | The JPEG compression library sends its compressed data to a "destination |
1450 | manager" module. The default destination manager just writes the data to a | 1450 | manager" module. The default destination manager just writes the data to a |
1451 | memory buffer or to a stdio stream, but you can provide your own manager to | 1451 | memory buffer or to a stdio stream, but you can provide your own manager to |
1452 | do something else. Similarly, the decompression library calls a "source | 1452 | do something else. Similarly, the decompression library calls a "source |
1453 | manager" to obtain the compressed data; you can provide your own source | 1453 | manager" to obtain the compressed data; you can provide your own source |
1454 | manager if you want the data to come from somewhere other than a memory | 1454 | manager if you want the data to come from somewhere other than a memory |
1455 | buffer or a stdio stream. | 1455 | buffer or a stdio stream. |
1456 | 1456 | ||
1457 | In both cases, compressed data is processed a bufferload at a time: the | 1457 | In both cases, compressed data is processed a bufferload at a time: the |
1458 | destination or source manager provides a work buffer, and the library invokes | 1458 | destination or source manager provides a work buffer, and the library invokes |
1459 | the manager only when the buffer is filled or emptied. (You could define a | 1459 | the manager only when the buffer is filled or emptied. (You could define a |
1460 | one-character buffer to force the manager to be invoked for each byte, but | 1460 | one-character buffer to force the manager to be invoked for each byte, but |
1461 | that would be rather inefficient.) The buffer's size and location are | 1461 | that would be rather inefficient.) The buffer's size and location are |
1462 | controlled by the manager, not by the library. For example, the memory | 1462 | controlled by the manager, not by the library. For example, the memory |
1463 | source manager just makes the buffer pointer and length point to the original | 1463 | source manager just makes the buffer pointer and length point to the original |
1464 | data in memory. In this case the buffer-reload procedure will be invoked | 1464 | data in memory. In this case the buffer-reload procedure will be invoked |
1465 | only if the decompressor ran off the end of the datastream, which would | 1465 | only if the decompressor ran off the end of the datastream, which would |
1466 | indicate an erroneous datastream. | 1466 | indicate an erroneous datastream. |
1467 | 1467 | ||
1468 | The work buffer is defined as an array of datatype JOCTET, which is generally | 1468 | The work buffer is defined as an array of datatype JOCTET, which is generally |
1469 | "char" or "unsigned char". On a machine where char is not exactly 8 bits | 1469 | "char" or "unsigned char". On a machine where char is not exactly 8 bits |
1470 | wide, you must define JOCTET as a wider data type and then modify the data | 1470 | wide, you must define JOCTET as a wider data type and then modify the data |
1471 | source and destination modules to transcribe the work arrays into 8-bit units | 1471 | source and destination modules to transcribe the work arrays into 8-bit units |
1472 | on external storage. | 1472 | on external storage. |
1473 | 1473 | ||
1474 | A data destination manager struct contains a pointer and count defining the | 1474 | A data destination manager struct contains a pointer and count defining the |
1475 | next byte to write in the work buffer and the remaining free space: | 1475 | next byte to write in the work buffer and the remaining free space: |
1476 | 1476 | ||
1477 | JOCTET * next_output_byte; /* => next byte to write in buffer */ | 1477 | JOCTET * next_output_byte; /* => next byte to write in buffer */ |
1478 | size_t free_in_buffer; /* # of byte spaces remaining in buffer */ | 1478 | size_t free_in_buffer; /* # of byte spaces remaining in buffer */ |
1479 | 1479 | ||
1480 | The library increments the pointer and decrements the count until the buffer | 1480 | The library increments the pointer and decrements the count until the buffer |
1481 | is filled. The manager's empty_output_buffer method must reset the pointer | 1481 | is filled. The manager's empty_output_buffer method must reset the pointer |
1482 | and count. The manager is expected to remember the buffer's starting address | 1482 | and count. The manager is expected to remember the buffer's starting address |
1483 | and total size in private fields not visible to the library. | 1483 | and total size in private fields not visible to the library. |
1484 | 1484 | ||
1485 | A data destination manager provides three methods: | 1485 | A data destination manager provides three methods: |
1486 | 1486 | ||
1487 | init_destination (j_compress_ptr cinfo) | 1487 | init_destination (j_compress_ptr cinfo) |
1488 | Initialize destination. This is called by jpeg_start_compress() | 1488 | Initialize destination. This is called by jpeg_start_compress() |
1489 | before any data is actually written. It must initialize | 1489 | before any data is actually written. It must initialize |
1490 | next_output_byte and free_in_buffer. free_in_buffer must be | 1490 | next_output_byte and free_in_buffer. free_in_buffer must be |
1491 | initialized to a positive value. | 1491 | initialized to a positive value. |
1492 | 1492 | ||
1493 | empty_output_buffer (j_compress_ptr cinfo) | 1493 | empty_output_buffer (j_compress_ptr cinfo) |
1494 | This is called whenever the buffer has filled (free_in_buffer | 1494 | This is called whenever the buffer has filled (free_in_buffer |
1495 | reaches zero). In typical applications, it should write out the | 1495 | reaches zero). In typical applications, it should write out the |
1496 | *entire* buffer (use the saved start address and buffer length; | 1496 | *entire* buffer (use the saved start address and buffer length; |
1497 | ignore the current state of next_output_byte and free_in_buffer). | 1497 | ignore the current state of next_output_byte and free_in_buffer). |
1498 | Then reset the pointer & count to the start of the buffer, and | 1498 | Then reset the pointer & count to the start of the buffer, and |
1499 | return TRUE indicating that the buffer has been dumped. | 1499 | return TRUE indicating that the buffer has been dumped. |
1500 | free_in_buffer must be set to a positive value when TRUE is | 1500 | free_in_buffer must be set to a positive value when TRUE is |
1501 | returned. A FALSE return should only be used when I/O suspension is | 1501 | returned. A FALSE return should only be used when I/O suspension is |
1502 | desired (this operating mode is discussed in the next section). | 1502 | desired (this operating mode is discussed in the next section). |
1503 | 1503 | ||
1504 | term_destination (j_compress_ptr cinfo) | 1504 | term_destination (j_compress_ptr cinfo) |
1505 | Terminate destination --- called by jpeg_finish_compress() after all | 1505 | Terminate destination --- called by jpeg_finish_compress() after all |
1506 | data has been written. In most applications, this must flush any | 1506 | data has been written. In most applications, this must flush any |
1507 | data remaining in the buffer. Use either next_output_byte or | 1507 | data remaining in the buffer. Use either next_output_byte or |
1508 | free_in_buffer to determine how much data is in the buffer. | 1508 | free_in_buffer to determine how much data is in the buffer. |
1509 | 1509 | ||
1510 | term_destination() is NOT called by jpeg_abort() or jpeg_destroy(). If you | 1510 | term_destination() is NOT called by jpeg_abort() or jpeg_destroy(). If you |
1511 | want the destination manager to be cleaned up during an abort, you must do it | 1511 | want the destination manager to be cleaned up during an abort, you must do it |
1512 | yourself. | 1512 | yourself. |
1513 | 1513 | ||
1514 | You will also need code to create a jpeg_destination_mgr struct, fill in its | 1514 | You will also need code to create a jpeg_destination_mgr struct, fill in its |
1515 | method pointers, and insert a pointer to the struct into the "dest" field of | 1515 | method pointers, and insert a pointer to the struct into the "dest" field of |
1516 | the JPEG compression object. This can be done in-line in your setup code if | 1516 | the JPEG compression object. This can be done in-line in your setup code if |
1517 | you like, but it's probably cleaner to provide a separate routine similar to | 1517 | you like, but it's probably cleaner to provide a separate routine similar to |
1518 | the jpeg_stdio_dest() or jpeg_mem_dest() routines of the supplied destination | 1518 | the jpeg_stdio_dest() or jpeg_mem_dest() routines of the supplied destination |
1519 | managers. | 1519 | managers. |
1520 | 1520 | ||
1521 | Decompression source managers follow a parallel design, but with some | 1521 | Decompression source managers follow a parallel design, but with some |
1522 | additional frammishes. The source manager struct contains a pointer and count | 1522 | additional frammishes. The source manager struct contains a pointer and count |
1523 | defining the next byte to read from the work buffer and the number of bytes | 1523 | defining the next byte to read from the work buffer and the number of bytes |
1524 | remaining: | 1524 | remaining: |
1525 | 1525 | ||
1526 | const JOCTET * next_input_byte; /* => next byte to read from buffer */ | 1526 | const JOCTET * next_input_byte; /* => next byte to read from buffer */ |
1527 | size_t bytes_in_buffer; /* # of bytes remaining in buffer */ | 1527 | size_t bytes_in_buffer; /* # of bytes remaining in buffer */ |
1528 | 1528 | ||
1529 | The library increments the pointer and decrements the count until the buffer | 1529 | The library increments the pointer and decrements the count until the buffer |
1530 | is emptied. The manager's fill_input_buffer method must reset the pointer and | 1530 | is emptied. The manager's fill_input_buffer method must reset the pointer and |
1531 | count. In most applications, the manager must remember the buffer's starting | 1531 | count. In most applications, the manager must remember the buffer's starting |
1532 | address and total size in private fields not visible to the library. | 1532 | address and total size in private fields not visible to the library. |
1533 | 1533 | ||
1534 | A data source manager provides five methods: | 1534 | A data source manager provides five methods: |
1535 | 1535 | ||
1536 | init_source (j_decompress_ptr cinfo) | 1536 | init_source (j_decompress_ptr cinfo) |
1537 | Initialize source. This is called by jpeg_read_header() before any | 1537 | Initialize source. This is called by jpeg_read_header() before any |
1538 | data is actually read. Unlike init_destination(), it may leave | 1538 | data is actually read. Unlike init_destination(), it may leave |
1539 | bytes_in_buffer set to 0 (in which case a fill_input_buffer() call | 1539 | bytes_in_buffer set to 0 (in which case a fill_input_buffer() call |
1540 | will occur immediately). | 1540 | will occur immediately). |
1541 | 1541 | ||
1542 | fill_input_buffer (j_decompress_ptr cinfo) | 1542 | fill_input_buffer (j_decompress_ptr cinfo) |
1543 | This is called whenever bytes_in_buffer has reached zero and more | 1543 | This is called whenever bytes_in_buffer has reached zero and more |
1544 | data is wanted. In typical applications, it should read fresh data | 1544 | data is wanted. In typical applications, it should read fresh data |
1545 | into the buffer (ignoring the current state of next_input_byte and | 1545 | into the buffer (ignoring the current state of next_input_byte and |
1546 | bytes_in_buffer), reset the pointer & count to the start of the | 1546 | bytes_in_buffer), reset the pointer & count to the start of the |
1547 | buffer, and return TRUE indicating that the buffer has been reloaded. | 1547 | buffer, and return TRUE indicating that the buffer has been reloaded. |
1548 | It is not necessary to fill the buffer entirely, only to obtain at | 1548 | It is not necessary to fill the buffer entirely, only to obtain at |
1549 | least one more byte. bytes_in_buffer MUST be set to a positive value | 1549 | least one more byte. bytes_in_buffer MUST be set to a positive value |
1550 | if TRUE is returned. A FALSE return should only be used when I/O | 1550 | if TRUE is returned. A FALSE return should only be used when I/O |
1551 | suspension is desired (this mode is discussed in the next section). | 1551 | suspension is desired (this mode is discussed in the next section). |
1552 | 1552 | ||
1553 | skip_input_data (j_decompress_ptr cinfo, long num_bytes) | 1553 | skip_input_data (j_decompress_ptr cinfo, long num_bytes) |
1554 | Skip num_bytes worth of data. The buffer pointer and count should | 1554 | Skip num_bytes worth of data. The buffer pointer and count should |
1555 | be advanced over num_bytes input bytes, refilling the buffer as | 1555 | be advanced over num_bytes input bytes, refilling the buffer as |
1556 | needed. This is used to skip over a potentially large amount of | 1556 | needed. This is used to skip over a potentially large amount of |
1557 | uninteresting data (such as an APPn marker). In some applications | 1557 | uninteresting data (such as an APPn marker). In some applications |
1558 | it may be possible to optimize away the reading of the skipped data, | 1558 | it may be possible to optimize away the reading of the skipped data, |
1559 | but it's not clear that being smart is worth much trouble; large | 1559 | but it's not clear that being smart is worth much trouble; large |
1560 | skips are uncommon. bytes_in_buffer may be zero on return. | 1560 | skips are uncommon. bytes_in_buffer may be zero on return. |
1561 | A zero or negative skip count should be treated as a no-op. | 1561 | A zero or negative skip count should be treated as a no-op. |
1562 | 1562 | ||
1563 | resync_to_restart (j_decompress_ptr cinfo, int desired) | 1563 | resync_to_restart (j_decompress_ptr cinfo, int desired) |
1564 | This routine is called only when the decompressor has failed to find | 1564 | This routine is called only when the decompressor has failed to find |
1565 | a restart (RSTn) marker where one is expected. Its mission is to | 1565 | a restart (RSTn) marker where one is expected. Its mission is to |
1566 | find a suitable point for resuming decompression. For most | 1566 | find a suitable point for resuming decompression. For most |
1567 | applications, we recommend that you just use the default resync | 1567 | applications, we recommend that you just use the default resync |
1568 | procedure, jpeg_resync_to_restart(). However, if you are able to back | 1568 | procedure, jpeg_resync_to_restart(). However, if you are able to back |
1569 | up in the input data stream, or if you have a-priori knowledge about | 1569 | up in the input data stream, or if you have a-priori knowledge about |
1570 | the likely location of restart markers, you may be able to do better. | 1570 | the likely location of restart markers, you may be able to do better. |
1571 | Read the read_restart_marker() and jpeg_resync_to_restart() routines | 1571 | Read the read_restart_marker() and jpeg_resync_to_restart() routines |
1572 | in jdmarker.c if you think you'd like to implement your own resync | 1572 | in jdmarker.c if you think you'd like to implement your own resync |
1573 | procedure. | 1573 | procedure. |
1574 | 1574 | ||
1575 | term_source (j_decompress_ptr cinfo) | 1575 | term_source (j_decompress_ptr cinfo) |
1576 | Terminate source --- called by jpeg_finish_decompress() after all | 1576 | Terminate source --- called by jpeg_finish_decompress() after all |
1577 | data has been read. Often a no-op. | 1577 | data has been read. Often a no-op. |
1578 | 1578 | ||
1579 | For both fill_input_buffer() and skip_input_data(), there is no such thing | 1579 | For both fill_input_buffer() and skip_input_data(), there is no such thing |
1580 | as an EOF return. If the end of the file has been reached, the routine has | 1580 | as an EOF return. If the end of the file has been reached, the routine has |
1581 | a choice of exiting via ERREXIT() or inserting fake data into the buffer. | 1581 | a choice of exiting via ERREXIT() or inserting fake data into the buffer. |
1582 | In most cases, generating a warning message and inserting a fake EOI marker | 1582 | In most cases, generating a warning message and inserting a fake EOI marker |
1583 | is the best course of action --- this will allow the decompressor to output | 1583 | is the best course of action --- this will allow the decompressor to output |
1584 | however much of the image is there. In pathological cases, the decompressor | 1584 | however much of the image is there. In pathological cases, the decompressor |
1585 | may swallow the EOI and again demand data ... just keep feeding it fake EOIs. | 1585 | may swallow the EOI and again demand data ... just keep feeding it fake EOIs. |
1586 | jdatasrc.c illustrates the recommended error recovery behavior. | 1586 | jdatasrc.c illustrates the recommended error recovery behavior. |
1587 | 1587 | ||
1588 | term_source() is NOT called by jpeg_abort() or jpeg_destroy(). If you want | 1588 | term_source() is NOT called by jpeg_abort() or jpeg_destroy(). If you want |
1589 | the source manager to be cleaned up during an abort, you must do it yourself. | 1589 | the source manager to be cleaned up during an abort, you must do it yourself. |
1590 | 1590 | ||
1591 | You will also need code to create a jpeg_source_mgr struct, fill in its method | 1591 | You will also need code to create a jpeg_source_mgr struct, fill in its method |
1592 | pointers, and insert a pointer to the struct into the "src" field of the JPEG | 1592 | pointers, and insert a pointer to the struct into the "src" field of the JPEG |
1593 | decompression object. This can be done in-line in your setup code if you | 1593 | decompression object. This can be done in-line in your setup code if you |
1594 | like, but it's probably cleaner to provide a separate routine similar to the | 1594 | like, but it's probably cleaner to provide a separate routine similar to the |
1595 | jpeg_stdio_src() or jpeg_mem_src() routines of the supplied source managers. | 1595 | jpeg_stdio_src() or jpeg_mem_src() routines of the supplied source managers. |
1596 | 1596 | ||
1597 | For more information, consult the memory and stdio source and destination | 1597 | For more information, consult the memory and stdio source and destination |
1598 | managers in jdatasrc.c and jdatadst.c. | 1598 | managers in jdatasrc.c and jdatadst.c. |
1599 | 1599 | ||
1600 | 1600 | ||
1601 | I/O suspension | 1601 | I/O suspension |
1602 | -------------- | 1602 | -------------- |
1603 | 1603 | ||
1604 | Some applications need to use the JPEG library as an incremental memory-to- | 1604 | Some applications need to use the JPEG library as an incremental memory-to- |
1605 | memory filter: when the compressed data buffer is filled or emptied, they want | 1605 | memory filter: when the compressed data buffer is filled or emptied, they want |
1606 | control to return to the outer loop, rather than expecting that the buffer can | 1606 | control to return to the outer loop, rather than expecting that the buffer can |
1607 | be emptied or reloaded within the data source/destination manager subroutine. | 1607 | be emptied or reloaded within the data source/destination manager subroutine. |
1608 | The library supports this need by providing an "I/O suspension" mode, which we | 1608 | The library supports this need by providing an "I/O suspension" mode, which we |
1609 | describe in this section. | 1609 | describe in this section. |
1610 | 1610 | ||
1611 | The I/O suspension mode is not a panacea: nothing is guaranteed about the | 1611 | The I/O suspension mode is not a panacea: nothing is guaranteed about the |
1612 | maximum amount of time spent in any one call to the library, so it will not | 1612 | maximum amount of time spent in any one call to the library, so it will not |
1613 | eliminate response-time problems in single-threaded applications. If you | 1613 | eliminate response-time problems in single-threaded applications. If you |
1614 | need guaranteed response time, we suggest you "bite the bullet" and implement | 1614 | need guaranteed response time, we suggest you "bite the bullet" and implement |
1615 | a real multi-tasking capability. | 1615 | a real multi-tasking capability. |
1616 | 1616 | ||
1617 | To use I/O suspension, cooperation is needed between the calling application | 1617 | To use I/O suspension, cooperation is needed between the calling application |
1618 | and the data source or destination manager; you will always need a custom | 1618 | and the data source or destination manager; you will always need a custom |
1619 | source/destination manager. (Please read the previous section if you haven't | 1619 | source/destination manager. (Please read the previous section if you haven't |
1620 | already.) The basic idea is that the empty_output_buffer() or | 1620 | already.) The basic idea is that the empty_output_buffer() or |
1621 | fill_input_buffer() routine is a no-op, merely returning FALSE to indicate | 1621 | fill_input_buffer() routine is a no-op, merely returning FALSE to indicate |
1622 | that it has done nothing. Upon seeing this, the JPEG library suspends | 1622 | that it has done nothing. Upon seeing this, the JPEG library suspends |
1623 | operation and returns to its caller. The surrounding application is | 1623 | operation and returns to its caller. The surrounding application is |
1624 | responsible for emptying or refilling the work buffer before calling the | 1624 | responsible for emptying or refilling the work buffer before calling the |
1625 | JPEG library again. | 1625 | JPEG library again. |
1626 | 1626 | ||
1627 | Compression suspension: | 1627 | Compression suspension: |
1628 | 1628 | ||
1629 | For compression suspension, use an empty_output_buffer() routine that returns | 1629 | For compression suspension, use an empty_output_buffer() routine that returns |
1630 | FALSE; typically it will not do anything else. This will cause the | 1630 | FALSE; typically it will not do anything else. This will cause the |
1631 | compressor to return to the caller of jpeg_write_scanlines(), with the return | 1631 | compressor to return to the caller of jpeg_write_scanlines(), with the return |
1632 | value indicating that not all the supplied scanlines have been accepted. | 1632 | value indicating that not all the supplied scanlines have been accepted. |
1633 | The application must make more room in the output buffer, adjust the output | 1633 | The application must make more room in the output buffer, adjust the output |
1634 | buffer pointer/count appropriately, and then call jpeg_write_scanlines() | 1634 | buffer pointer/count appropriately, and then call jpeg_write_scanlines() |
1635 | again, pointing to the first unconsumed scanline. | 1635 | again, pointing to the first unconsumed scanline. |
1636 | 1636 | ||
1637 | When forced to suspend, the compressor will backtrack to a convenient stopping | 1637 | When forced to suspend, the compressor will backtrack to a convenient stopping |
1638 | point (usually the start of the current MCU); it will regenerate some output | 1638 | point (usually the start of the current MCU); it will regenerate some output |
1639 | data when restarted. Therefore, although empty_output_buffer() is only | 1639 | data when restarted. Therefore, although empty_output_buffer() is only |
1640 | called when the buffer is filled, you should NOT write out the entire buffer | 1640 | called when the buffer is filled, you should NOT write out the entire buffer |
1641 | after a suspension. Write only the data up to the current position of | 1641 | after a suspension. Write only the data up to the current position of |
1642 | next_output_byte/free_in_buffer. The data beyond that point will be | 1642 | next_output_byte/free_in_buffer. The data beyond that point will be |
1643 | regenerated after resumption. | 1643 | regenerated after resumption. |
1644 | 1644 | ||
1645 | Because of the backtracking behavior, a good-size output buffer is essential | 1645 | Because of the backtracking behavior, a good-size output buffer is essential |
1646 | for efficiency; you don't want the compressor to suspend often. (In fact, an | 1646 | for efficiency; you don't want the compressor to suspend often. (In fact, an |
1647 | overly small buffer could lead to infinite looping, if a single MCU required | 1647 | overly small buffer could lead to infinite looping, if a single MCU required |
1648 | more data than would fit in the buffer.) We recommend a buffer of at least | 1648 | more data than would fit in the buffer.) We recommend a buffer of at least |
1649 | several Kbytes. You may want to insert explicit code to ensure that you don't | 1649 | several Kbytes. You may want to insert explicit code to ensure that you don't |
1650 | call jpeg_write_scanlines() unless there is a reasonable amount of space in | 1650 | call jpeg_write_scanlines() unless there is a reasonable amount of space in |
1651 | the output buffer; in other words, flush the buffer before trying to compress | 1651 | the output buffer; in other words, flush the buffer before trying to compress |
1652 | more data. | 1652 | more data. |
1653 | 1653 | ||
1654 | The compressor does not allow suspension while it is trying to write JPEG | 1654 | The compressor does not allow suspension while it is trying to write JPEG |
1655 | markers at the beginning and end of the file. This means that: | 1655 | markers at the beginning and end of the file. This means that: |
1656 | * At the beginning of a compression operation, there must be enough free | 1656 | * At the beginning of a compression operation, there must be enough free |
1657 | space in the output buffer to hold the header markers (typically 600 or | 1657 | space in the output buffer to hold the header markers (typically 600 or |
1658 | so bytes). The recommended buffer size is bigger than this anyway, so | 1658 | so bytes). The recommended buffer size is bigger than this anyway, so |
1659 | this is not a problem as long as you start with an empty buffer. However, | 1659 | this is not a problem as long as you start with an empty buffer. However, |
1660 | this restriction might catch you if you insert large special markers, such | 1660 | this restriction might catch you if you insert large special markers, such |
1661 | as a JFIF thumbnail image, without flushing the buffer afterwards. | 1661 | as a JFIF thumbnail image, without flushing the buffer afterwards. |
1662 | * When you call jpeg_finish_compress(), there must be enough space in the | 1662 | * When you call jpeg_finish_compress(), there must be enough space in the |
1663 | output buffer to emit any buffered data and the final EOI marker. In the | 1663 | output buffer to emit any buffered data and the final EOI marker. In the |
1664 | current implementation, half a dozen bytes should suffice for this, but | 1664 | current implementation, half a dozen bytes should suffice for this, but |
1665 | for safety's sake we recommend ensuring that at least 100 bytes are free | 1665 | for safety's sake we recommend ensuring that at least 100 bytes are free |
1666 | before calling jpeg_finish_compress(). | 1666 | before calling jpeg_finish_compress(). |
1667 | 1667 | ||
1668 | A more significant restriction is that jpeg_finish_compress() cannot suspend. | 1668 | A more significant restriction is that jpeg_finish_compress() cannot suspend. |
1669 | This means you cannot use suspension with multi-pass operating modes, namely | 1669 | This means you cannot use suspension with multi-pass operating modes, namely |
1670 | Huffman code optimization and multiple-scan output. Those modes write the | 1670 | Huffman code optimization and multiple-scan output. Those modes write the |
1671 | whole file during jpeg_finish_compress(), which will certainly result in | 1671 | whole file during jpeg_finish_compress(), which will certainly result in |
1672 | buffer overrun. (Note that this restriction applies only to compression, | 1672 | buffer overrun. (Note that this restriction applies only to compression, |
1673 | not decompression. The decompressor supports input suspension in all of its | 1673 | not decompression. The decompressor supports input suspension in all of its |
1674 | operating modes.) | 1674 | operating modes.) |
1675 | 1675 | ||
1676 | Decompression suspension: | 1676 | Decompression suspension: |
1677 | 1677 | ||
1678 | For decompression suspension, use a fill_input_buffer() routine that simply | 1678 | For decompression suspension, use a fill_input_buffer() routine that simply |
1679 | returns FALSE (except perhaps during error recovery, as discussed below). | 1679 | returns FALSE (except perhaps during error recovery, as discussed below). |
1680 | This will cause the decompressor to return to its caller with an indication | 1680 | This will cause the decompressor to return to its caller with an indication |
1681 | that suspension has occurred. This can happen at four places: | 1681 | that suspension has occurred. This can happen at four places: |
1682 | * jpeg_read_header(): will return JPEG_SUSPENDED. | 1682 | * jpeg_read_header(): will return JPEG_SUSPENDED. |
1683 | * jpeg_start_decompress(): will return FALSE, rather than its usual TRUE. | 1683 | * jpeg_start_decompress(): will return FALSE, rather than its usual TRUE. |
1684 | * jpeg_read_scanlines(): will return the number of scanlines already | 1684 | * jpeg_read_scanlines(): will return the number of scanlines already |
1685 | completed (possibly 0). | 1685 | completed (possibly 0). |
1686 | * jpeg_finish_decompress(): will return FALSE, rather than its usual TRUE. | 1686 | * jpeg_finish_decompress(): will return FALSE, rather than its usual TRUE. |
1687 | The surrounding application must recognize these cases, load more data into | 1687 | The surrounding application must recognize these cases, load more data into |
1688 | the input buffer, and repeat the call. In the case of jpeg_read_scanlines(), | 1688 | the input buffer, and repeat the call. In the case of jpeg_read_scanlines(), |
1689 | increment the passed pointers past any scanlines successfully read. | 1689 | increment the passed pointers past any scanlines successfully read. |
1690 | 1690 | ||
1691 | Just as with compression, the decompressor will typically backtrack to a | 1691 | Just as with compression, the decompressor will typically backtrack to a |
1692 | convenient restart point before suspending. When fill_input_buffer() is | 1692 | convenient restart point before suspending. When fill_input_buffer() is |
1693 | called, next_input_byte/bytes_in_buffer point to the current restart point, | 1693 | called, next_input_byte/bytes_in_buffer point to the current restart point, |
1694 | which is where the decompressor will backtrack to if FALSE is returned. | 1694 | which is where the decompressor will backtrack to if FALSE is returned. |
1695 | The data beyond that position must NOT be discarded if you suspend; it needs | 1695 | The data beyond that position must NOT be discarded if you suspend; it needs |
1696 | to be re-read upon resumption. In most implementations, you'll need to shift | 1696 | to be re-read upon resumption. In most implementations, you'll need to shift |
1697 | this data down to the start of your work buffer and then load more data after | 1697 | this data down to the start of your work buffer and then load more data after |
1698 | it. Again, this behavior means that a several-Kbyte work buffer is essential | 1698 | it. Again, this behavior means that a several-Kbyte work buffer is essential |
1699 | for decent performance; furthermore, you should load a reasonable amount of | 1699 | for decent performance; furthermore, you should load a reasonable amount of |
1700 | new data before resuming decompression. (If you loaded, say, only one new | 1700 | new data before resuming decompression. (If you loaded, say, only one new |
1701 | byte each time around, you could waste a LOT of cycles.) | 1701 | byte each time around, you could waste a LOT of cycles.) |
1702 | 1702 | ||
1703 | The skip_input_data() source manager routine requires special care in a | 1703 | The skip_input_data() source manager routine requires special care in a |
1704 | suspension scenario. This routine is NOT granted the ability to suspend the | 1704 | suspension scenario. This routine is NOT granted the ability to suspend the |
1705 | decompressor; it can decrement bytes_in_buffer to zero, but no more. If the | 1705 | decompressor; it can decrement bytes_in_buffer to zero, but no more. If the |
1706 | requested skip distance exceeds the amount of data currently in the input | 1706 | requested skip distance exceeds the amount of data currently in the input |
1707 | buffer, then skip_input_data() must set bytes_in_buffer to zero and record the | 1707 | buffer, then skip_input_data() must set bytes_in_buffer to zero and record the |
1708 | additional skip distance somewhere else. The decompressor will immediately | 1708 | additional skip distance somewhere else. The decompressor will immediately |
1709 | call fill_input_buffer(), which should return FALSE, which will cause a | 1709 | call fill_input_buffer(), which should return FALSE, which will cause a |
1710 | suspension return. The surrounding application must then arrange to discard | 1710 | suspension return. The surrounding application must then arrange to discard |
1711 | the recorded number of bytes before it resumes loading the input buffer. | 1711 | the recorded number of bytes before it resumes loading the input buffer. |
1712 | (Yes, this design is rather baroque, but it avoids complexity in the far more | 1712 | (Yes, this design is rather baroque, but it avoids complexity in the far more |
1713 | common case where a non-suspending source manager is used.) | 1713 | common case where a non-suspending source manager is used.) |
1714 | 1714 | ||
1715 | If the input data has been exhausted, we recommend that you emit a warning | 1715 | If the input data has been exhausted, we recommend that you emit a warning |
1716 | and insert dummy EOI markers just as a non-suspending data source manager | 1716 | and insert dummy EOI markers just as a non-suspending data source manager |
1717 | would do. This can be handled either in the surrounding application logic or | 1717 | would do. This can be handled either in the surrounding application logic or |
1718 | within fill_input_buffer(); the latter is probably more efficient. If | 1718 | within fill_input_buffer(); the latter is probably more efficient. If |
1719 | fill_input_buffer() knows that no more data is available, it can set the | 1719 | fill_input_buffer() knows that no more data is available, it can set the |
1720 | pointer/count to point to a dummy EOI marker and then return TRUE just as | 1720 | pointer/count to point to a dummy EOI marker and then return TRUE just as |
1721 | though it had read more data in a non-suspending situation. | 1721 | though it had read more data in a non-suspending situation. |
1722 | 1722 | ||
1723 | The decompressor does not attempt to suspend within standard JPEG markers; | 1723 | The decompressor does not attempt to suspend within standard JPEG markers; |
1724 | instead it will backtrack to the start of the marker and reprocess the whole | 1724 | instead it will backtrack to the start of the marker and reprocess the whole |
1725 | marker next time. Hence the input buffer must be large enough to hold the | 1725 | marker next time. Hence the input buffer must be large enough to hold the |
1726 | longest standard marker in the file. Standard JPEG markers should normally | 1726 | longest standard marker in the file. Standard JPEG markers should normally |
1727 | not exceed a few hundred bytes each (DHT tables are typically the longest). | 1727 | not exceed a few hundred bytes each (DHT tables are typically the longest). |
1728 | We recommend at least a 2K buffer for performance reasons, which is much | 1728 | We recommend at least a 2K buffer for performance reasons, which is much |
1729 | larger than any correct marker is likely to be. For robustness against | 1729 | larger than any correct marker is likely to be. For robustness against |
1730 | damaged marker length counts, you may wish to insert a test in your | 1730 | damaged marker length counts, you may wish to insert a test in your |
1731 | application for the case that the input buffer is completely full and yet | 1731 | application for the case that the input buffer is completely full and yet |
1732 | the decoder has suspended without consuming any data --- otherwise, if this | 1732 | the decoder has suspended without consuming any data --- otherwise, if this |
1733 | situation did occur, it would lead to an endless loop. (The library can't | 1733 | situation did occur, it would lead to an endless loop. (The library can't |
1734 | provide this test since it has no idea whether "the buffer is full", or | 1734 | provide this test since it has no idea whether "the buffer is full", or |
1735 | even whether there is a fixed-size input buffer.) | 1735 | even whether there is a fixed-size input buffer.) |
1736 | 1736 | ||
1737 | The input buffer would need to be 64K to allow for arbitrary COM or APPn | 1737 | The input buffer would need to be 64K to allow for arbitrary COM or APPn |
1738 | markers, but these are handled specially: they are either saved into allocated | 1738 | markers, but these are handled specially: they are either saved into allocated |
1739 | memory, or skipped over by calling skip_input_data(). In the former case, | 1739 | memory, or skipped over by calling skip_input_data(). In the former case, |
1740 | suspension is handled correctly, and in the latter case, the problem of | 1740 | suspension is handled correctly, and in the latter case, the problem of |
1741 | buffer overrun is placed on skip_input_data's shoulders, as explained above. | 1741 | buffer overrun is placed on skip_input_data's shoulders, as explained above. |
1742 | Note that if you provide your own marker handling routine for large markers, | 1742 | Note that if you provide your own marker handling routine for large markers, |
1743 | you should consider how to deal with buffer overflow. | 1743 | you should consider how to deal with buffer overflow. |
1744 | 1744 | ||
1745 | Multiple-buffer management: | 1745 | Multiple-buffer management: |
1746 | 1746 | ||
1747 | In some applications it is desirable to store the compressed data in a linked | 1747 | In some applications it is desirable to store the compressed data in a linked |
1748 | list of buffer areas, so as to avoid data copying. This can be handled by | 1748 | list of buffer areas, so as to avoid data copying. This can be handled by |
1749 | having empty_output_buffer() or fill_input_buffer() set the pointer and count | 1749 | having empty_output_buffer() or fill_input_buffer() set the pointer and count |
1750 | to reference the next available buffer; FALSE is returned only if no more | 1750 | to reference the next available buffer; FALSE is returned only if no more |
1751 | buffers are available. Although seemingly straightforward, there is a | 1751 | buffers are available. Although seemingly straightforward, there is a |
1752 | pitfall in this approach: the backtrack that occurs when FALSE is returned | 1752 | pitfall in this approach: the backtrack that occurs when FALSE is returned |
1753 | could back up into an earlier buffer. For example, when fill_input_buffer() | 1753 | could back up into an earlier buffer. For example, when fill_input_buffer() |
1754 | is called, the current pointer & count indicate the backtrack restart point. | 1754 | is called, the current pointer & count indicate the backtrack restart point. |
1755 | Since fill_input_buffer() will set the pointer and count to refer to a new | 1755 | Since fill_input_buffer() will set the pointer and count to refer to a new |
1756 | buffer, the restart position must be saved somewhere else. Suppose a second | 1756 | buffer, the restart position must be saved somewhere else. Suppose a second |
1757 | call to fill_input_buffer() occurs in the same library call, and no | 1757 | call to fill_input_buffer() occurs in the same library call, and no |
1758 | additional input data is available, so fill_input_buffer must return FALSE. | 1758 | additional input data is available, so fill_input_buffer must return FALSE. |
1759 | If the JPEG library has not moved the pointer/count forward in the current | 1759 | If the JPEG library has not moved the pointer/count forward in the current |
1760 | buffer, then *the correct restart point is the saved position in the prior | 1760 | buffer, then *the correct restart point is the saved position in the prior |
1761 | buffer*. Prior buffers may be discarded only after the library establishes | 1761 | buffer*. Prior buffers may be discarded only after the library establishes |
1762 | a restart point within a later buffer. Similar remarks apply for output into | 1762 | a restart point within a later buffer. Similar remarks apply for output into |
1763 | a chain of buffers. | 1763 | a chain of buffers. |
1764 | 1764 | ||
1765 | The library will never attempt to backtrack over a skip_input_data() call, | 1765 | The library will never attempt to backtrack over a skip_input_data() call, |
1766 | so any skipped data can be permanently discarded. You still have to deal | 1766 | so any skipped data can be permanently discarded. You still have to deal |
1767 | with the case of skipping not-yet-received data, however. | 1767 | with the case of skipping not-yet-received data, however. |
1768 | 1768 | ||
1769 | It's much simpler to use only a single buffer; when fill_input_buffer() is | 1769 | It's much simpler to use only a single buffer; when fill_input_buffer() is |
1770 | called, move any unconsumed data (beyond the current pointer/count) down to | 1770 | called, move any unconsumed data (beyond the current pointer/count) down to |
1771 | the beginning of this buffer and then load new data into the remaining buffer | 1771 | the beginning of this buffer and then load new data into the remaining buffer |
1772 | space. This approach requires a little more data copying but is far easier | 1772 | space. This approach requires a little more data copying but is far easier |
1773 | to get right. | 1773 | to get right. |
1774 | 1774 | ||
1775 | 1775 | ||
1776 | Progressive JPEG support | 1776 | Progressive JPEG support |
1777 | ------------------------ | 1777 | ------------------------ |
1778 | 1778 | ||
1779 | Progressive JPEG rearranges the stored data into a series of scans of | 1779 | Progressive JPEG rearranges the stored data into a series of scans of |
1780 | increasing quality. In situations where a JPEG file is transmitted across a | 1780 | increasing quality. In situations where a JPEG file is transmitted across a |
1781 | slow communications link, a decoder can generate a low-quality image very | 1781 | slow communications link, a decoder can generate a low-quality image very |
1782 | quickly from the first scan, then gradually improve the displayed quality as | 1782 | quickly from the first scan, then gradually improve the displayed quality as |
1783 | more scans are received. The final image after all scans are complete is | 1783 | more scans are received. The final image after all scans are complete is |
1784 | identical to that of a regular (sequential) JPEG file of the same quality | 1784 | identical to that of a regular (sequential) JPEG file of the same quality |
1785 | setting. Progressive JPEG files are often slightly smaller than equivalent | 1785 | setting. Progressive JPEG files are often slightly smaller than equivalent |
1786 | sequential JPEG files, but the possibility of incremental display is the main | 1786 | sequential JPEG files, but the possibility of incremental display is the main |
1787 | reason for using progressive JPEG. | 1787 | reason for using progressive JPEG. |
1788 | 1788 | ||
1789 | The IJG encoder library generates progressive JPEG files when given a | 1789 | The IJG encoder library generates progressive JPEG files when given a |
1790 | suitable "scan script" defining how to divide the data into scans. | 1790 | suitable "scan script" defining how to divide the data into scans. |
1791 | Creation of progressive JPEG files is otherwise transparent to the encoder. | 1791 | Creation of progressive JPEG files is otherwise transparent to the encoder. |
1792 | Progressive JPEG files can also be read transparently by the decoder library. | 1792 | Progressive JPEG files can also be read transparently by the decoder library. |
1793 | If the decoding application simply uses the library as defined above, it | 1793 | If the decoding application simply uses the library as defined above, it |
1794 | will receive a final decoded image without any indication that the file was | 1794 | will receive a final decoded image without any indication that the file was |
1795 | progressive. Of course, this approach does not allow incremental display. | 1795 | progressive. Of course, this approach does not allow incremental display. |
1796 | To perform incremental display, an application needs to use the decoder | 1796 | To perform incremental display, an application needs to use the decoder |
1797 | library's "buffered-image" mode, in which it receives a decoded image | 1797 | library's "buffered-image" mode, in which it receives a decoded image |
1798 | multiple times. | 1798 | multiple times. |
1799 | 1799 | ||
1800 | Each displayed scan requires about as much work to decode as a full JPEG | 1800 | Each displayed scan requires about as much work to decode as a full JPEG |
1801 | image of the same size, so the decoder must be fairly fast in relation to the | 1801 | image of the same size, so the decoder must be fairly fast in relation to the |
1802 | data transmission rate in order to make incremental display useful. However, | 1802 | data transmission rate in order to make incremental display useful. However, |
1803 | it is possible to skip displaying the image and simply add the incoming bits | 1803 | it is possible to skip displaying the image and simply add the incoming bits |
1804 | to the decoder's coefficient buffer. This is fast because only Huffman | 1804 | to the decoder's coefficient buffer. This is fast because only Huffman |
1805 | decoding need be done, not IDCT, upsampling, colorspace conversion, etc. | 1805 | decoding need be done, not IDCT, upsampling, colorspace conversion, etc. |
1806 | The IJG decoder library allows the application to switch dynamically between | 1806 | The IJG decoder library allows the application to switch dynamically between |
1807 | displaying the image and simply absorbing the incoming bits. A properly | 1807 | displaying the image and simply absorbing the incoming bits. A properly |
1808 | coded application can automatically adapt the number of display passes to | 1808 | coded application can automatically adapt the number of display passes to |
1809 | suit the time available as the image is received. Also, a final | 1809 | suit the time available as the image is received. Also, a final |
1810 | higher-quality display cycle can be performed from the buffered data after | 1810 | higher-quality display cycle can be performed from the buffered data after |
1811 | the end of the file is reached. | 1811 | the end of the file is reached. |
1812 | 1812 | ||
1813 | Progressive compression: | 1813 | Progressive compression: |
1814 | 1814 | ||
1815 | To create a progressive JPEG file (or a multiple-scan sequential JPEG file), | 1815 | To create a progressive JPEG file (or a multiple-scan sequential JPEG file), |
1816 | set the scan_info cinfo field to point to an array of scan descriptors, and | 1816 | set the scan_info cinfo field to point to an array of scan descriptors, and |
1817 | perform compression as usual. Instead of constructing your own scan list, | 1817 | perform compression as usual. Instead of constructing your own scan list, |
1818 | you can call the jpeg_simple_progression() helper routine to create a | 1818 | you can call the jpeg_simple_progression() helper routine to create a |
1819 | recommended progression sequence; this method should be used by all | 1819 | recommended progression sequence; this method should be used by all |
1820 | applications that don't want to get involved in the nitty-gritty of | 1820 | applications that don't want to get involved in the nitty-gritty of |
1821 | progressive scan sequence design. (If you want to provide user control of | 1821 | progressive scan sequence design. (If you want to provide user control of |
1822 | scan sequences, you may wish to borrow the scan script reading code found | 1822 | scan sequences, you may wish to borrow the scan script reading code found |
1823 | in rdswitch.c, so that you can read scan script files just like cjpeg's.) | 1823 | in rdswitch.c, so that you can read scan script files just like cjpeg's.) |
1824 | When scan_info is not NULL, the compression library will store DCT'd data | 1824 | When scan_info is not NULL, the compression library will store DCT'd data |
1825 | into a buffer array as jpeg_write_scanlines() is called, and will emit all | 1825 | into a buffer array as jpeg_write_scanlines() is called, and will emit all |
1826 | the requested scans during jpeg_finish_compress(). This implies that | 1826 | the requested scans during jpeg_finish_compress(). This implies that |
1827 | multiple-scan output cannot be created with a suspending data destination | 1827 | multiple-scan output cannot be created with a suspending data destination |
1828 | manager, since jpeg_finish_compress() does not support suspension. We | 1828 | manager, since jpeg_finish_compress() does not support suspension. We |
1829 | should also note that the compressor currently forces Huffman optimization | 1829 | should also note that the compressor currently forces Huffman optimization |
1830 | mode when creating a progressive JPEG file, because the default Huffman | 1830 | mode when creating a progressive JPEG file, because the default Huffman |
1831 | tables are unsuitable for progressive files. | 1831 | tables are unsuitable for progressive files. |
1832 | 1832 | ||
1833 | Progressive decompression: | 1833 | Progressive decompression: |
1834 | 1834 | ||
1835 | When buffered-image mode is not used, the decoder library will read all of | 1835 | When buffered-image mode is not used, the decoder library will read all of |
1836 | a multi-scan file during jpeg_start_decompress(), so that it can provide a | 1836 | a multi-scan file during jpeg_start_decompress(), so that it can provide a |
1837 | final decoded image. (Here "multi-scan" means either progressive or | 1837 | final decoded image. (Here "multi-scan" means either progressive or |
1838 | multi-scan sequential.) This makes multi-scan files transparent to the | 1838 | multi-scan sequential.) This makes multi-scan files transparent to the |
1839 | decoding application. However, existing applications that used suspending | 1839 | decoding application. However, existing applications that used suspending |
1840 | input with version 5 of the IJG library will need to be modified to check | 1840 | input with version 5 of the IJG library will need to be modified to check |
1841 | for a suspension return from jpeg_start_decompress(). | 1841 | for a suspension return from jpeg_start_decompress(). |
1842 | 1842 | ||
1843 | To perform incremental display, an application must use the library's | 1843 | To perform incremental display, an application must use the library's |
1844 | buffered-image mode. This is described in the next section. | 1844 | buffered-image mode. This is described in the next section. |
1845 | 1845 | ||
1846 | 1846 | ||
1847 | Buffered-image mode | 1847 | Buffered-image mode |
1848 | ------------------- | 1848 | ------------------- |
1849 | 1849 | ||
1850 | In buffered-image mode, the library stores the partially decoded image in a | 1850 | In buffered-image mode, the library stores the partially decoded image in a |
1851 | coefficient buffer, from which it can be read out as many times as desired. | 1851 | coefficient buffer, from which it can be read out as many times as desired. |
1852 | This mode is typically used for incremental display of progressive JPEG files, | 1852 | This mode is typically used for incremental display of progressive JPEG files, |
1853 | but it can be used with any JPEG file. Each scan of a progressive JPEG file | 1853 | but it can be used with any JPEG file. Each scan of a progressive JPEG file |
1854 | adds more data (more detail) to the buffered image. The application can | 1854 | adds more data (more detail) to the buffered image. The application can |
1855 | display in lockstep with the source file (one display pass per input scan), | 1855 | display in lockstep with the source file (one display pass per input scan), |
1856 | or it can allow input processing to outrun display processing. By making | 1856 | or it can allow input processing to outrun display processing. By making |
1857 | input and display processing run independently, it is possible for the | 1857 | input and display processing run independently, it is possible for the |
1858 | application to adapt progressive display to a wide range of data transmission | 1858 | application to adapt progressive display to a wide range of data transmission |
1859 | rates. | 1859 | rates. |
1860 | 1860 | ||
1861 | The basic control flow for buffered-image decoding is | 1861 | The basic control flow for buffered-image decoding is |
1862 | 1862 | ||
1863 | jpeg_create_decompress() | 1863 | jpeg_create_decompress() |
1864 | set data source | 1864 | set data source |
1865 | jpeg_read_header() | 1865 | jpeg_read_header() |
1866 | set overall decompression parameters | 1866 | set overall decompression parameters |
1867 | cinfo.buffered_image = TRUE; /* select buffered-image mode */ | 1867 | cinfo.buffered_image = TRUE; /* select buffered-image mode */ |
1868 | jpeg_start_decompress() | 1868 | jpeg_start_decompress() |
1869 | for (each output pass) { | 1869 | for (each output pass) { |
1870 | adjust output decompression parameters if required | 1870 | adjust output decompression parameters if required |
1871 | jpeg_start_output() /* start a new output pass */ | 1871 | jpeg_start_output() /* start a new output pass */ |
1872 | for (all scanlines in image) { | 1872 | for (all scanlines in image) { |
1873 | jpeg_read_scanlines() | 1873 | jpeg_read_scanlines() |
1874 | display scanlines | 1874 | display scanlines |
1875 | } | 1875 | } |
1876 | jpeg_finish_output() /* terminate output pass */ | 1876 | jpeg_finish_output() /* terminate output pass */ |
1877 | } | 1877 | } |
1878 | jpeg_finish_decompress() | 1878 | jpeg_finish_decompress() |
1879 | jpeg_destroy_decompress() | 1879 | jpeg_destroy_decompress() |
1880 | 1880 | ||
1881 | This differs from ordinary unbuffered decoding in that there is an additional | 1881 | This differs from ordinary unbuffered decoding in that there is an additional |
1882 | level of looping. The application can choose how many output passes to make | 1882 | level of looping. The application can choose how many output passes to make |
1883 | and how to display each pass. | 1883 | and how to display each pass. |
1884 | 1884 | ||
1885 | The simplest approach to displaying progressive images is to do one display | 1885 | The simplest approach to displaying progressive images is to do one display |
1886 | pass for each scan appearing in the input file. In this case the outer loop | 1886 | pass for each scan appearing in the input file. In this case the outer loop |
1887 | condition is typically | 1887 | condition is typically |
1888 | while (! jpeg_input_complete(&cinfo)) | 1888 | while (! jpeg_input_complete(&cinfo)) |
1889 | and the start-output call should read | 1889 | and the start-output call should read |
1890 | jpeg_start_output(&cinfo, cinfo.input_scan_number); | 1890 | jpeg_start_output(&cinfo, cinfo.input_scan_number); |
1891 | The second parameter to jpeg_start_output() indicates which scan of the input | 1891 | The second parameter to jpeg_start_output() indicates which scan of the input |
1892 | file is to be displayed; the scans are numbered starting at 1 for this | 1892 | file is to be displayed; the scans are numbered starting at 1 for this |
1893 | purpose. (You can use a loop counter starting at 1 if you like, but using | 1893 | purpose. (You can use a loop counter starting at 1 if you like, but using |
1894 | the library's input scan counter is easier.) The library automatically reads | 1894 | the library's input scan counter is easier.) The library automatically reads |
1895 | data as necessary to complete each requested scan, and jpeg_finish_output() | 1895 | data as necessary to complete each requested scan, and jpeg_finish_output() |
1896 | advances to the next scan or end-of-image marker (hence input_scan_number | 1896 | advances to the next scan or end-of-image marker (hence input_scan_number |
1897 | will be incremented by the time control arrives back at jpeg_start_output()). | 1897 | will be incremented by the time control arrives back at jpeg_start_output()). |
1898 | With this technique, data is read from the input file only as needed, and | 1898 | With this technique, data is read from the input file only as needed, and |
1899 | input and output processing run in lockstep. | 1899 | input and output processing run in lockstep. |
1900 | 1900 | ||
1901 | After reading the final scan and reaching the end of the input file, the | 1901 | After reading the final scan and reaching the end of the input file, the |
1902 | buffered image remains available; it can be read additional times by | 1902 | buffered image remains available; it can be read additional times by |
1903 | repeating the jpeg_start_output()/jpeg_read_scanlines()/jpeg_finish_output() | 1903 | repeating the jpeg_start_output()/jpeg_read_scanlines()/jpeg_finish_output() |
1904 | sequence. For example, a useful technique is to use fast one-pass color | 1904 | sequence. For example, a useful technique is to use fast one-pass color |
1905 | quantization for display passes made while the image is arriving, followed by | 1905 | quantization for display passes made while the image is arriving, followed by |
1906 | a final display pass using two-pass quantization for highest quality. This | 1906 | a final display pass using two-pass quantization for highest quality. This |
1907 | is done by changing the library parameters before the final output pass. | 1907 | is done by changing the library parameters before the final output pass. |
1908 | Changing parameters between passes is discussed in detail below. | 1908 | Changing parameters between passes is discussed in detail below. |
1909 | 1909 | ||
1910 | In general the last scan of a progressive file cannot be recognized as such | 1910 | In general the last scan of a progressive file cannot be recognized as such |
1911 | until after it is read, so a post-input display pass is the best approach if | 1911 | until after it is read, so a post-input display pass is the best approach if |
1912 | you want special processing in the final pass. | 1912 | you want special processing in the final pass. |
1913 | 1913 | ||
1914 | When done with the image, be sure to call jpeg_finish_decompress() to release | 1914 | When done with the image, be sure to call jpeg_finish_decompress() to release |
1915 | the buffered image (or just use jpeg_destroy_decompress()). | 1915 | the buffered image (or just use jpeg_destroy_decompress()). |
1916 | 1916 | ||
1917 | If input data arrives faster than it can be displayed, the application can | 1917 | If input data arrives faster than it can be displayed, the application can |
1918 | cause the library to decode input data in advance of what's needed to produce | 1918 | cause the library to decode input data in advance of what's needed to produce |
1919 | output. This is done by calling the routine jpeg_consume_input(). | 1919 | output. This is done by calling the routine jpeg_consume_input(). |
1920 | The return value is one of the following: | 1920 | The return value is one of the following: |
1921 | JPEG_REACHED_SOS: reached an SOS marker (the start of a new scan) | 1921 | JPEG_REACHED_SOS: reached an SOS marker (the start of a new scan) |
1922 | JPEG_REACHED_EOI: reached the EOI marker (end of image) | 1922 | JPEG_REACHED_EOI: reached the EOI marker (end of image) |
1923 | JPEG_ROW_COMPLETED: completed reading one MCU row of compressed data | 1923 | JPEG_ROW_COMPLETED: completed reading one MCU row of compressed data |
1924 | JPEG_SCAN_COMPLETED: completed reading last MCU row of current scan | 1924 | JPEG_SCAN_COMPLETED: completed reading last MCU row of current scan |
1925 | JPEG_SUSPENDED: suspended before completing any of the above | 1925 | JPEG_SUSPENDED: suspended before completing any of the above |
1926 | (JPEG_SUSPENDED can occur only if a suspending data source is used.) This | 1926 | (JPEG_SUSPENDED can occur only if a suspending data source is used.) This |
1927 | routine can be called at any time after initializing the JPEG object. It | 1927 | routine can be called at any time after initializing the JPEG object. It |
1928 | reads some additional data and returns when one of the indicated significant | 1928 | reads some additional data and returns when one of the indicated significant |
1929 | events occurs. (If called after the EOI marker is reached, it will | 1929 | events occurs. (If called after the EOI marker is reached, it will |
1930 | immediately return JPEG_REACHED_EOI without attempting to read more data.) | 1930 | immediately return JPEG_REACHED_EOI without attempting to read more data.) |
1931 | 1931 | ||
1932 | The library's output processing will automatically call jpeg_consume_input() | 1932 | The library's output processing will automatically call jpeg_consume_input() |
1933 | whenever the output processing overtakes the input; thus, simple lockstep | 1933 | whenever the output processing overtakes the input; thus, simple lockstep |
1934 | display requires no direct calls to jpeg_consume_input(). But by adding | 1934 | display requires no direct calls to jpeg_consume_input(). But by adding |
1935 | calls to jpeg_consume_input(), you can absorb data in advance of what is | 1935 | calls to jpeg_consume_input(), you can absorb data in advance of what is |
1936 | being displayed. This has two benefits: | 1936 | being displayed. This has two benefits: |
1937 | * You can limit buildup of unprocessed data in your input buffer. | 1937 | * You can limit buildup of unprocessed data in your input buffer. |
1938 | * You can eliminate extra display passes by paying attention to the | 1938 | * You can eliminate extra display passes by paying attention to the |
1939 | state of the library's input processing. | 1939 | state of the library's input processing. |
1940 | 1940 | ||
1941 | The first of these benefits only requires interspersing calls to | 1941 | The first of these benefits only requires interspersing calls to |
1942 | jpeg_consume_input() with your display operations and any other processing | 1942 | jpeg_consume_input() with your display operations and any other processing |
1943 | you may be doing. To avoid wasting cycles due to backtracking, it's best to | 1943 | you may be doing. To avoid wasting cycles due to backtracking, it's best to |
1944 | call jpeg_consume_input() only after a hundred or so new bytes have arrived. | 1944 | call jpeg_consume_input() only after a hundred or so new bytes have arrived. |
1945 | This is discussed further under "I/O suspension", above. (Note: the JPEG | 1945 | This is discussed further under "I/O suspension", above. (Note: the JPEG |
1946 | library currently is not thread-safe. You must not call jpeg_consume_input() | 1946 | library currently is not thread-safe. You must not call jpeg_consume_input() |
1947 | from one thread of control if a different library routine is working on the | 1947 | from one thread of control if a different library routine is working on the |
1948 | same JPEG object in another thread.) | 1948 | same JPEG object in another thread.) |
1949 | 1949 | ||
1950 | When input arrives fast enough that more than one new scan is available | 1950 | When input arrives fast enough that more than one new scan is available |
1951 | before you start a new output pass, you may as well skip the output pass | 1951 | before you start a new output pass, you may as well skip the output pass |
1952 | corresponding to the completed scan. This occurs for free if you pass | 1952 | corresponding to the completed scan. This occurs for free if you pass |
1953 | cinfo.input_scan_number as the target scan number to jpeg_start_output(). | 1953 | cinfo.input_scan_number as the target scan number to jpeg_start_output(). |
1954 | The input_scan_number field is simply the index of the scan currently being | 1954 | The input_scan_number field is simply the index of the scan currently being |
1955 | consumed by the input processor. You can ensure that this is up-to-date by | 1955 | consumed by the input processor. You can ensure that this is up-to-date by |
1956 | emptying the input buffer just before calling jpeg_start_output(): call | 1956 | emptying the input buffer just before calling jpeg_start_output(): call |
1957 | jpeg_consume_input() repeatedly until it returns JPEG_SUSPENDED or | 1957 | jpeg_consume_input() repeatedly until it returns JPEG_SUSPENDED or |
1958 | JPEG_REACHED_EOI. | 1958 | JPEG_REACHED_EOI. |
1959 | 1959 | ||
1960 | The target scan number passed to jpeg_start_output() is saved in the | 1960 | The target scan number passed to jpeg_start_output() is saved in the |
1961 | cinfo.output_scan_number field. The library's output processing calls | 1961 | cinfo.output_scan_number field. The library's output processing calls |
1962 | jpeg_consume_input() whenever the current input scan number and row within | 1962 | jpeg_consume_input() whenever the current input scan number and row within |
1963 | that scan is less than or equal to the current output scan number and row. | 1963 | that scan is less than or equal to the current output scan number and row. |
1964 | Thus, input processing can "get ahead" of the output processing but is not | 1964 | Thus, input processing can "get ahead" of the output processing but is not |
1965 | allowed to "fall behind". You can achieve several different effects by | 1965 | allowed to "fall behind". You can achieve several different effects by |
1966 | manipulating this interlock rule. For example, if you pass a target scan | 1966 | manipulating this interlock rule. For example, if you pass a target scan |
1967 | number greater than the current input scan number, the output processor will | 1967 | number greater than the current input scan number, the output processor will |
1968 | wait until that scan starts to arrive before producing any output. (To avoid | 1968 | wait until that scan starts to arrive before producing any output. (To avoid |
1969 | an infinite loop, the target scan number is automatically reset to the last | 1969 | an infinite loop, the target scan number is automatically reset to the last |
1970 | scan number when the end of image is reached. Thus, if you specify a large | 1970 | scan number when the end of image is reached. Thus, if you specify a large |
1971 | target scan number, the library will just absorb the entire input file and | 1971 | target scan number, the library will just absorb the entire input file and |
1972 | then perform an output pass. This is effectively the same as what | 1972 | then perform an output pass. This is effectively the same as what |
1973 | jpeg_start_decompress() does when you don't select buffered-image mode.) | 1973 | jpeg_start_decompress() does when you don't select buffered-image mode.) |
1974 | When you pass a target scan number equal to the current input scan number, | 1974 | When you pass a target scan number equal to the current input scan number, |
1975 | the image is displayed no faster than the current input scan arrives. The | 1975 | the image is displayed no faster than the current input scan arrives. The |
1976 | final possibility is to pass a target scan number less than the current input | 1976 | final possibility is to pass a target scan number less than the current input |
1977 | scan number; this disables the input/output interlock and causes the output | 1977 | scan number; this disables the input/output interlock and causes the output |
1978 | processor to simply display whatever it finds in the image buffer, without | 1978 | processor to simply display whatever it finds in the image buffer, without |
1979 | waiting for input. (However, the library will not accept a target scan | 1979 | waiting for input. (However, the library will not accept a target scan |
1980 | number less than one, so you can't avoid waiting for the first scan.) | 1980 | number less than one, so you can't avoid waiting for the first scan.) |
1981 | 1981 | ||
1982 | When data is arriving faster than the output display processing can advance | 1982 | When data is arriving faster than the output display processing can advance |
1983 | through the image, jpeg_consume_input() will store data into the buffered | 1983 | through the image, jpeg_consume_input() will store data into the buffered |
1984 | image beyond the point at which the output processing is reading data out | 1984 | image beyond the point at which the output processing is reading data out |
1985 | again. If the input arrives fast enough, it may "wrap around" the buffer to | 1985 | again. If the input arrives fast enough, it may "wrap around" the buffer to |
1986 | the point where the input is more than one whole scan ahead of the output. | 1986 | the point where the input is more than one whole scan ahead of the output. |
1987 | If the output processing simply proceeds through its display pass without | 1987 | If the output processing simply proceeds through its display pass without |
1988 | paying attention to the input, the effect seen on-screen is that the lower | 1988 | paying attention to the input, the effect seen on-screen is that the lower |
1989 | part of the image is one or more scans better in quality than the upper part. | 1989 | part of the image is one or more scans better in quality than the upper part. |
1990 | Then, when the next output scan is started, you have a choice of what target | 1990 | Then, when the next output scan is started, you have a choice of what target |
1991 | scan number to use. The recommended choice is to use the current input scan | 1991 | scan number to use. The recommended choice is to use the current input scan |
1992 | number at that time, which implies that you've skipped the output scans | 1992 | number at that time, which implies that you've skipped the output scans |
1993 | corresponding to the input scans that were completed while you processed the | 1993 | corresponding to the input scans that were completed while you processed the |
1994 | previous output scan. In this way, the decoder automatically adapts its | 1994 | previous output scan. In this way, the decoder automatically adapts its |
1995 | speed to the arriving data, by skipping output scans as necessary to keep up | 1995 | speed to the arriving data, by skipping output scans as necessary to keep up |
1996 | with the arriving data. | 1996 | with the arriving data. |
1997 | 1997 | ||
1998 | When using this strategy, you'll want to be sure that you perform a final | 1998 | When using this strategy, you'll want to be sure that you perform a final |
1999 | output pass after receiving all the data; otherwise your last display may not | 1999 | output pass after receiving all the data; otherwise your last display may not |
2000 | be full quality across the whole screen. So the right outer loop logic is | 2000 | be full quality across the whole screen. So the right outer loop logic is |
2001 | something like this: | 2001 | something like this: |
2002 | do { | 2002 | do { |
2003 | absorb any waiting input by calling jpeg_consume_input() | 2003 | absorb any waiting input by calling jpeg_consume_input() |
2004 | final_pass = jpeg_input_complete(&cinfo); | 2004 | final_pass = jpeg_input_complete(&cinfo); |
2005 | adjust output decompression parameters if required | 2005 | adjust output decompression parameters if required |
2006 | jpeg_start_output(&cinfo, cinfo.input_scan_number); | 2006 | jpeg_start_output(&cinfo, cinfo.input_scan_number); |
2007 | ... | 2007 | ... |
2008 | jpeg_finish_output() | 2008 | jpeg_finish_output() |
2009 | } while (! final_pass); | 2009 | } while (! final_pass); |
2010 | rather than quitting as soon as jpeg_input_complete() returns TRUE. This | 2010 | rather than quitting as soon as jpeg_input_complete() returns TRUE. This |
2011 | arrangement makes it simple to use higher-quality decoding parameters | 2011 | arrangement makes it simple to use higher-quality decoding parameters |
2012 | for the final pass. But if you don't want to use special parameters for | 2012 | for the final pass. But if you don't want to use special parameters for |
2013 | the final pass, the right loop logic is like this: | 2013 | the final pass, the right loop logic is like this: |
2014 | for (;;) { | 2014 | for (;;) { |
2015 | absorb any waiting input by calling jpeg_consume_input() | 2015 | absorb any waiting input by calling jpeg_consume_input() |
2016 | jpeg_start_output(&cinfo, cinfo.input_scan_number); | 2016 | jpeg_start_output(&cinfo, cinfo.input_scan_number); |
2017 | ... | 2017 | ... |
2018 | jpeg_finish_output() | 2018 | jpeg_finish_output() |
2019 | if (jpeg_input_complete(&cinfo) && | 2019 | if (jpeg_input_complete(&cinfo) && |
2020 | cinfo.input_scan_number == cinfo.output_scan_number) | 2020 | cinfo.input_scan_number == cinfo.output_scan_number) |
2021 | break; | 2021 | break; |
2022 | } | 2022 | } |
2023 | In this case you don't need to know in advance whether an output pass is to | 2023 | In this case you don't need to know in advance whether an output pass is to |
2024 | be the last one, so it's not necessary to have reached EOF before starting | 2024 | be the last one, so it's not necessary to have reached EOF before starting |
2025 | the final output pass; rather, what you want to test is whether the output | 2025 | the final output pass; rather, what you want to test is whether the output |
2026 | pass was performed in sync with the final input scan. This form of the loop | 2026 | pass was performed in sync with the final input scan. This form of the loop |
2027 | will avoid an extra output pass whenever the decoder is able (or nearly able) | 2027 | will avoid an extra output pass whenever the decoder is able (or nearly able) |
2028 | to keep up with the incoming data. | 2028 | to keep up with the incoming data. |
2029 | 2029 | ||
2030 | When the data transmission speed is high, you might begin a display pass, | 2030 | When the data transmission speed is high, you might begin a display pass, |
2031 | then find that much or all of the file has arrived before you can complete | 2031 | then find that much or all of the file has arrived before you can complete |
2032 | the pass. (You can detect this by noting the JPEG_REACHED_EOI return code | 2032 | the pass. (You can detect this by noting the JPEG_REACHED_EOI return code |
2033 | from jpeg_consume_input(), or equivalently by testing jpeg_input_complete().) | 2033 | from jpeg_consume_input(), or equivalently by testing jpeg_input_complete().) |
2034 | In this situation you may wish to abort the current display pass and start a | 2034 | In this situation you may wish to abort the current display pass and start a |
2035 | new one using the newly arrived information. To do so, just call | 2035 | new one using the newly arrived information. To do so, just call |
2036 | jpeg_finish_output() and then start a new pass with jpeg_start_output(). | 2036 | jpeg_finish_output() and then start a new pass with jpeg_start_output(). |
2037 | 2037 | ||
2038 | A variant strategy is to abort and restart display if more than one complete | 2038 | A variant strategy is to abort and restart display if more than one complete |
2039 | scan arrives during an output pass; this can be detected by noting | 2039 | scan arrives during an output pass; this can be detected by noting |
2040 | JPEG_REACHED_SOS returns and/or examining cinfo.input_scan_number. This | 2040 | JPEG_REACHED_SOS returns and/or examining cinfo.input_scan_number. This |
2041 | idea should be employed with caution, however, since the display process | 2041 | idea should be employed with caution, however, since the display process |
2042 | might never get to the bottom of the image before being aborted, resulting | 2042 | might never get to the bottom of the image before being aborted, resulting |
2043 | in the lower part of the screen being several passes worse than the upper. | 2043 | in the lower part of the screen being several passes worse than the upper. |
2044 | In most cases it's probably best to abort an output pass only if the whole | 2044 | In most cases it's probably best to abort an output pass only if the whole |
2045 | file has arrived and you want to begin the final output pass immediately. | 2045 | file has arrived and you want to begin the final output pass immediately. |
2046 | 2046 | ||
2047 | When receiving data across a communication link, we recommend always using | 2047 | When receiving data across a communication link, we recommend always using |
2048 | the current input scan number for the output target scan number; if a | 2048 | the current input scan number for the output target scan number; if a |
2049 | higher-quality final pass is to be done, it should be started (aborting any | 2049 | higher-quality final pass is to be done, it should be started (aborting any |
2050 | incomplete output pass) as soon as the end of file is received. However, | 2050 | incomplete output pass) as soon as the end of file is received. However, |
2051 | many other strategies are possible. For example, the application can examine | 2051 | many other strategies are possible. For example, the application can examine |
2052 | the parameters of the current input scan and decide whether to display it or | 2052 | the parameters of the current input scan and decide whether to display it or |
2053 | not. If the scan contains only chroma data, one might choose not to use it | 2053 | not. If the scan contains only chroma data, one might choose not to use it |
2054 | as the target scan, expecting that the scan will be small and will arrive | 2054 | as the target scan, expecting that the scan will be small and will arrive |
2055 | quickly. To skip to the next scan, call jpeg_consume_input() until it | 2055 | quickly. To skip to the next scan, call jpeg_consume_input() until it |
2056 | returns JPEG_REACHED_SOS or JPEG_REACHED_EOI. Or just use the next higher | 2056 | returns JPEG_REACHED_SOS or JPEG_REACHED_EOI. Or just use the next higher |
2057 | number as the target scan for jpeg_start_output(); but that method doesn't | 2057 | number as the target scan for jpeg_start_output(); but that method doesn't |
2058 | let you inspect the next scan's parameters before deciding to display it. | 2058 | let you inspect the next scan's parameters before deciding to display it. |
2059 | 2059 | ||
2060 | 2060 | ||
2061 | In buffered-image mode, jpeg_start_decompress() never performs input and | 2061 | In buffered-image mode, jpeg_start_decompress() never performs input and |
2062 | thus never suspends. An application that uses input suspension with | 2062 | thus never suspends. An application that uses input suspension with |
2063 | buffered-image mode must be prepared for suspension returns from these | 2063 | buffered-image mode must be prepared for suspension returns from these |
2064 | routines: | 2064 | routines: |
2065 | * jpeg_start_output() performs input only if you request 2-pass quantization | 2065 | * jpeg_start_output() performs input only if you request 2-pass quantization |
2066 | and the target scan isn't fully read yet. (This is discussed below.) | 2066 | and the target scan isn't fully read yet. (This is discussed below.) |
2067 | * jpeg_read_scanlines(), as always, returns the number of scanlines that it | 2067 | * jpeg_read_scanlines(), as always, returns the number of scanlines that it |
2068 | was able to produce before suspending. | 2068 | was able to produce before suspending. |
2069 | * jpeg_finish_output() will read any markers following the target scan, | 2069 | * jpeg_finish_output() will read any markers following the target scan, |
2070 | up to the end of the file or the SOS marker that begins another scan. | 2070 | up to the end of the file or the SOS marker that begins another scan. |
2071 | (But it reads no input if jpeg_consume_input() has already reached the | 2071 | (But it reads no input if jpeg_consume_input() has already reached the |
2072 | end of the file or a SOS marker beyond the target output scan.) | 2072 | end of the file or a SOS marker beyond the target output scan.) |
2073 | * jpeg_finish_decompress() will read until the end of file, and thus can | 2073 | * jpeg_finish_decompress() will read until the end of file, and thus can |
2074 | suspend if the end hasn't already been reached (as can be tested by | 2074 | suspend if the end hasn't already been reached (as can be tested by |
2075 | calling jpeg_input_complete()). | 2075 | calling jpeg_input_complete()). |
2076 | jpeg_start_output(), jpeg_finish_output(), and jpeg_finish_decompress() | 2076 | jpeg_start_output(), jpeg_finish_output(), and jpeg_finish_decompress() |
2077 | all return TRUE if they completed their tasks, FALSE if they had to suspend. | 2077 | all return TRUE if they completed their tasks, FALSE if they had to suspend. |
2078 | In the event of a FALSE return, the application must load more input data | 2078 | In the event of a FALSE return, the application must load more input data |
2079 | and repeat the call. Applications that use non-suspending data sources need | 2079 | and repeat the call. Applications that use non-suspending data sources need |
2080 | not check the return values of these three routines. | 2080 | not check the return values of these three routines. |
2081 | 2081 | ||
2082 | 2082 | ||
2083 | It is possible to change decoding parameters between output passes in the | 2083 | It is possible to change decoding parameters between output passes in the |
2084 | buffered-image mode. The decoder library currently supports only very | 2084 | buffered-image mode. The decoder library currently supports only very |
2085 | limited changes of parameters. ONLY THE FOLLOWING parameter changes are | 2085 | limited changes of parameters. ONLY THE FOLLOWING parameter changes are |
2086 | allowed after jpeg_start_decompress() is called: | 2086 | allowed after jpeg_start_decompress() is called: |
2087 | * dct_method can be changed before each call to jpeg_start_output(). | 2087 | * dct_method can be changed before each call to jpeg_start_output(). |
2088 | For example, one could use a fast DCT method for early scans, changing | 2088 | For example, one could use a fast DCT method for early scans, changing |
2089 | to a higher quality method for the final scan. | 2089 | to a higher quality method for the final scan. |
2090 | * dither_mode can be changed before each call to jpeg_start_output(); | 2090 | * dither_mode can be changed before each call to jpeg_start_output(); |
2091 | of course this has no impact if not using color quantization. Typically | 2091 | of course this has no impact if not using color quantization. Typically |
2092 | one would use ordered dither for initial passes, then switch to | 2092 | one would use ordered dither for initial passes, then switch to |
2093 | Floyd-Steinberg dither for the final pass. Caution: changing dither mode | 2093 | Floyd-Steinberg dither for the final pass. Caution: changing dither mode |
2094 | can cause more memory to be allocated by the library. Although the amount | 2094 | can cause more memory to be allocated by the library. Although the amount |
2095 | of memory involved is not large (a scanline or so), it may cause the | 2095 | of memory involved is not large (a scanline or so), it may cause the |
2096 | initial max_memory_to_use specification to be exceeded, which in the worst | 2096 | initial max_memory_to_use specification to be exceeded, which in the worst |
2097 | case would result in an out-of-memory failure. | 2097 | case would result in an out-of-memory failure. |
2098 | * do_block_smoothing can be changed before each call to jpeg_start_output(). | 2098 | * do_block_smoothing can be changed before each call to jpeg_start_output(). |
2099 | This setting is relevant only when decoding a progressive JPEG image. | 2099 | This setting is relevant only when decoding a progressive JPEG image. |
2100 | During the first DC-only scan, block smoothing provides a very "fuzzy" look | 2100 | During the first DC-only scan, block smoothing provides a very "fuzzy" look |
2101 | instead of the very "blocky" look seen without it; which is better seems a | 2101 | instead of the very "blocky" look seen without it; which is better seems a |
2102 | matter of personal taste. But block smoothing is nearly always a win | 2102 | matter of personal taste. But block smoothing is nearly always a win |
2103 | during later stages, especially when decoding a successive-approximation | 2103 | during later stages, especially when decoding a successive-approximation |
2104 | image: smoothing helps to hide the slight blockiness that otherwise shows | 2104 | image: smoothing helps to hide the slight blockiness that otherwise shows |
2105 | up on smooth gradients until the lowest coefficient bits are sent. | 2105 | up on smooth gradients until the lowest coefficient bits are sent. |
2106 | * Color quantization mode can be changed under the rules described below. | 2106 | * Color quantization mode can be changed under the rules described below. |
2107 | You *cannot* change between full-color and quantized output (because that | 2107 | You *cannot* change between full-color and quantized output (because that |
2108 | would alter the required I/O buffer sizes), but you can change which | 2108 | would alter the required I/O buffer sizes), but you can change which |
2109 | quantization method is used. | 2109 | quantization method is used. |
2110 | 2110 | ||
2111 | When generating color-quantized output, changing quantization method is a | 2111 | When generating color-quantized output, changing quantization method is a |
2112 | very useful way of switching between high-speed and high-quality display. | 2112 | very useful way of switching between high-speed and high-quality display. |
2113 | The library allows you to change among its three quantization methods: | 2113 | The library allows you to change among its three quantization methods: |
2114 | 1. Single-pass quantization to a fixed color cube. | 2114 | 1. Single-pass quantization to a fixed color cube. |
2115 | Selected by cinfo.two_pass_quantize = FALSE and cinfo.colormap = NULL. | 2115 | Selected by cinfo.two_pass_quantize = FALSE and cinfo.colormap = NULL. |
2116 | 2. Single-pass quantization to an application-supplied colormap. | 2116 | 2. Single-pass quantization to an application-supplied colormap. |
2117 | Selected by setting cinfo.colormap to point to the colormap (the value of | 2117 | Selected by setting cinfo.colormap to point to the colormap (the value of |
2118 | two_pass_quantize is ignored); also set cinfo.actual_number_of_colors. | 2118 | two_pass_quantize is ignored); also set cinfo.actual_number_of_colors. |
2119 | 3. Two-pass quantization to a colormap chosen specifically for the image. | 2119 | 3. Two-pass quantization to a colormap chosen specifically for the image. |
2120 | Selected by cinfo.two_pass_quantize = TRUE and cinfo.colormap = NULL. | 2120 | Selected by cinfo.two_pass_quantize = TRUE and cinfo.colormap = NULL. |
2121 | (This is the default setting selected by jpeg_read_header, but it is | 2121 | (This is the default setting selected by jpeg_read_header, but it is |
2122 | probably NOT what you want for the first pass of progressive display!) | 2122 | probably NOT what you want for the first pass of progressive display!) |
2123 | These methods offer successively better quality and lesser speed. However, | 2123 | These methods offer successively better quality and lesser speed. However, |
2124 | only the first method is available for quantizing in non-RGB color spaces. | 2124 | only the first method is available for quantizing in non-RGB color spaces. |
2125 | 2125 | ||
2126 | IMPORTANT: because the different quantizer methods have very different | 2126 | IMPORTANT: because the different quantizer methods have very different |
2127 | working-storage requirements, the library requires you to indicate which | 2127 | working-storage requirements, the library requires you to indicate which |
2128 | one(s) you intend to use before you call jpeg_start_decompress(). (If we did | 2128 | one(s) you intend to use before you call jpeg_start_decompress(). (If we did |
2129 | not require this, the max_memory_to_use setting would be a complete fiction.) | 2129 | not require this, the max_memory_to_use setting would be a complete fiction.) |
2130 | You do this by setting one or more of these three cinfo fields to TRUE: | 2130 | You do this by setting one or more of these three cinfo fields to TRUE: |
2131 | enable_1pass_quant Fixed color cube colormap | 2131 | enable_1pass_quant Fixed color cube colormap |
2132 | enable_external_quant Externally-supplied colormap | 2132 | enable_external_quant Externally-supplied colormap |
2133 | enable_2pass_quant Two-pass custom colormap | 2133 | enable_2pass_quant Two-pass custom colormap |
2134 | All three are initialized FALSE by jpeg_read_header(). But | 2134 | All three are initialized FALSE by jpeg_read_header(). But |
2135 | jpeg_start_decompress() automatically sets TRUE the one selected by the | 2135 | jpeg_start_decompress() automatically sets TRUE the one selected by the |
2136 | current two_pass_quantize and colormap settings, so you only need to set the | 2136 | current two_pass_quantize and colormap settings, so you only need to set the |
2137 | enable flags for any other quantization methods you plan to change to later. | 2137 | enable flags for any other quantization methods you plan to change to later. |
2138 | 2138 | ||
2139 | After setting the enable flags correctly at jpeg_start_decompress() time, you | 2139 | After setting the enable flags correctly at jpeg_start_decompress() time, you |
2140 | can change to any enabled quantization method by setting two_pass_quantize | 2140 | can change to any enabled quantization method by setting two_pass_quantize |
2141 | and colormap properly just before calling jpeg_start_output(). The following | 2141 | and colormap properly just before calling jpeg_start_output(). The following |
2142 | special rules apply: | 2142 | special rules apply: |
2143 | 1. You must explicitly set cinfo.colormap to NULL when switching to 1-pass | 2143 | 1. You must explicitly set cinfo.colormap to NULL when switching to 1-pass |
2144 | or 2-pass mode from a different mode, or when you want the 2-pass | 2144 | or 2-pass mode from a different mode, or when you want the 2-pass |
2145 | quantizer to be re-run to generate a new colormap. | 2145 | quantizer to be re-run to generate a new colormap. |
2146 | 2. To switch to an external colormap, or to change to a different external | 2146 | 2. To switch to an external colormap, or to change to a different external |
2147 | colormap than was used on the prior pass, you must call | 2147 | colormap than was used on the prior pass, you must call |
2148 | jpeg_new_colormap() after setting cinfo.colormap. | 2148 | jpeg_new_colormap() after setting cinfo.colormap. |
2149 | NOTE: if you want to use the same colormap as was used in the prior pass, | 2149 | NOTE: if you want to use the same colormap as was used in the prior pass, |
2150 | you should not do either of these things. This will save some nontrivial | 2150 | you should not do either of these things. This will save some nontrivial |
2151 | switchover costs. | 2151 | switchover costs. |
2152 | (These requirements exist because cinfo.colormap will always be non-NULL | 2152 | (These requirements exist because cinfo.colormap will always be non-NULL |
2153 | after completing a prior output pass, since both the 1-pass and 2-pass | 2153 | after completing a prior output pass, since both the 1-pass and 2-pass |
2154 | quantizers set it to point to their output colormaps. Thus you have to | 2154 | quantizers set it to point to their output colormaps. Thus you have to |
2155 | do one of these two things to notify the library that something has changed. | 2155 | do one of these two things to notify the library that something has changed. |
2156 | Yup, it's a bit klugy, but it's necessary to do it this way for backwards | 2156 | Yup, it's a bit klugy, but it's necessary to do it this way for backwards |
2157 | compatibility.) | 2157 | compatibility.) |
2158 | 2158 | ||
2159 | Note that in buffered-image mode, the library generates any requested colormap | 2159 | Note that in buffered-image mode, the library generates any requested colormap |
2160 | during jpeg_start_output(), not during jpeg_start_decompress(). | 2160 | during jpeg_start_output(), not during jpeg_start_decompress(). |
2161 | 2161 | ||
2162 | When using two-pass quantization, jpeg_start_output() makes a pass over the | 2162 | When using two-pass quantization, jpeg_start_output() makes a pass over the |
2163 | buffered image to determine the optimum color map; it therefore may take a | 2163 | buffered image to determine the optimum color map; it therefore may take a |
2164 | significant amount of time, whereas ordinarily it does little work. The | 2164 | significant amount of time, whereas ordinarily it does little work. The |
2165 | progress monitor hook is called during this pass, if defined. It is also | 2165 | progress monitor hook is called during this pass, if defined. It is also |
2166 | important to realize that if the specified target scan number is greater than | 2166 | important to realize that if the specified target scan number is greater than |
2167 | or equal to the current input scan number, jpeg_start_output() will attempt | 2167 | or equal to the current input scan number, jpeg_start_output() will attempt |
2168 | to consume input as it makes this pass. If you use a suspending data source, | 2168 | to consume input as it makes this pass. If you use a suspending data source, |
2169 | you need to check for a FALSE return from jpeg_start_output() under these | 2169 | you need to check for a FALSE return from jpeg_start_output() under these |
2170 | conditions. The combination of 2-pass quantization and a not-yet-fully-read | 2170 | conditions. The combination of 2-pass quantization and a not-yet-fully-read |
2171 | target scan is the only case in which jpeg_start_output() will consume input. | 2171 | target scan is the only case in which jpeg_start_output() will consume input. |
2172 | 2172 | ||
2173 | 2173 | ||
2174 | Application authors who support buffered-image mode may be tempted to use it | 2174 | Application authors who support buffered-image mode may be tempted to use it |
2175 | for all JPEG images, even single-scan ones. This will work, but it is | 2175 | for all JPEG images, even single-scan ones. This will work, but it is |
2176 | inefficient: there is no need to create an image-sized coefficient buffer for | 2176 | inefficient: there is no need to create an image-sized coefficient buffer for |
2177 | single-scan images. Requesting buffered-image mode for such an image wastes | 2177 | single-scan images. Requesting buffered-image mode for such an image wastes |
2178 | memory. Worse, it can cost time on large images, since the buffered data has | 2178 | memory. Worse, it can cost time on large images, since the buffered data has |
2179 | to be swapped out or written to a temporary file. If you are concerned about | 2179 | to be swapped out or written to a temporary file. If you are concerned about |
2180 | maximum performance on baseline JPEG files, you should use buffered-image | 2180 | maximum performance on baseline JPEG files, you should use buffered-image |
2181 | mode only when the incoming file actually has multiple scans. This can be | 2181 | mode only when the incoming file actually has multiple scans. This can be |
2182 | tested by calling jpeg_has_multiple_scans(), which will return a correct | 2182 | tested by calling jpeg_has_multiple_scans(), which will return a correct |
2183 | result at any time after jpeg_read_header() completes. | 2183 | result at any time after jpeg_read_header() completes. |
2184 | 2184 | ||
2185 | It is also worth noting that when you use jpeg_consume_input() to let input | 2185 | It is also worth noting that when you use jpeg_consume_input() to let input |
2186 | processing get ahead of output processing, the resulting pattern of access to | 2186 | processing get ahead of output processing, the resulting pattern of access to |
2187 | the coefficient buffer is quite nonsequential. It's best to use the memory | 2187 | the coefficient buffer is quite nonsequential. It's best to use the memory |
2188 | manager jmemnobs.c if you can (ie, if you have enough real or virtual main | 2188 | manager jmemnobs.c if you can (ie, if you have enough real or virtual main |
2189 | memory). If not, at least make sure that max_memory_to_use is set as high as | 2189 | memory). If not, at least make sure that max_memory_to_use is set as high as |
2190 | possible. If the JPEG memory manager has to use a temporary file, you will | 2190 | possible. If the JPEG memory manager has to use a temporary file, you will |
2191 | probably see a lot of disk traffic and poor performance. (This could be | 2191 | probably see a lot of disk traffic and poor performance. (This could be |
2192 | improved with additional work on the memory manager, but we haven't gotten | 2192 | improved with additional work on the memory manager, but we haven't gotten |
2193 | around to it yet.) | 2193 | around to it yet.) |
2194 | 2194 | ||
2195 | In some applications it may be convenient to use jpeg_consume_input() for all | 2195 | In some applications it may be convenient to use jpeg_consume_input() for all |
2196 | input processing, including reading the initial markers; that is, you may | 2196 | input processing, including reading the initial markers; that is, you may |
2197 | wish to call jpeg_consume_input() instead of jpeg_read_header() during | 2197 | wish to call jpeg_consume_input() instead of jpeg_read_header() during |
2198 | startup. This works, but note that you must check for JPEG_REACHED_SOS and | 2198 | startup. This works, but note that you must check for JPEG_REACHED_SOS and |
2199 | JPEG_REACHED_EOI return codes as the equivalent of jpeg_read_header's codes. | 2199 | JPEG_REACHED_EOI return codes as the equivalent of jpeg_read_header's codes. |
2200 | Once the first SOS marker has been reached, you must call | 2200 | Once the first SOS marker has been reached, you must call |
2201 | jpeg_start_decompress() before jpeg_consume_input() will consume more input; | 2201 | jpeg_start_decompress() before jpeg_consume_input() will consume more input; |
2202 | it'll just keep returning JPEG_REACHED_SOS until you do. If you read a | 2202 | it'll just keep returning JPEG_REACHED_SOS until you do. If you read a |
2203 | tables-only file this way, jpeg_consume_input() will return JPEG_REACHED_EOI | 2203 | tables-only file this way, jpeg_consume_input() will return JPEG_REACHED_EOI |
2204 | without ever returning JPEG_REACHED_SOS; be sure to check for this case. | 2204 | without ever returning JPEG_REACHED_SOS; be sure to check for this case. |
2205 | If this happens, the decompressor will not read any more input until you call | 2205 | If this happens, the decompressor will not read any more input until you call |
2206 | jpeg_abort() to reset it. It is OK to call jpeg_consume_input() even when not | 2206 | jpeg_abort() to reset it. It is OK to call jpeg_consume_input() even when not |
2207 | using buffered-image mode, but in that case it's basically a no-op after the | 2207 | using buffered-image mode, but in that case it's basically a no-op after the |
2208 | initial markers have been read: it will just return JPEG_SUSPENDED. | 2208 | initial markers have been read: it will just return JPEG_SUSPENDED. |
2209 | 2209 | ||
2210 | 2210 | ||
2211 | Abbreviated datastreams and multiple images | 2211 | Abbreviated datastreams and multiple images |
2212 | ------------------------------------------- | 2212 | ------------------------------------------- |
2213 | 2213 | ||
2214 | A JPEG compression or decompression object can be reused to process multiple | 2214 | A JPEG compression or decompression object can be reused to process multiple |
2215 | images. This saves a small amount of time per image by eliminating the | 2215 | images. This saves a small amount of time per image by eliminating the |
2216 | "create" and "destroy" operations, but that isn't the real purpose of the | 2216 | "create" and "destroy" operations, but that isn't the real purpose of the |
2217 | feature. Rather, reuse of an object provides support for abbreviated JPEG | 2217 | feature. Rather, reuse of an object provides support for abbreviated JPEG |
2218 | datastreams. Object reuse can also simplify processing a series of images in | 2218 | datastreams. Object reuse can also simplify processing a series of images in |
2219 | a single input or output file. This section explains these features. | 2219 | a single input or output file. This section explains these features. |
2220 | 2220 | ||
2221 | A JPEG file normally contains several hundred bytes worth of quantization | 2221 | A JPEG file normally contains several hundred bytes worth of quantization |
2222 | and Huffman tables. In a situation where many images will be stored or | 2222 | and Huffman tables. In a situation where many images will be stored or |
2223 | transmitted with identical tables, this may represent an annoying overhead. | 2223 | transmitted with identical tables, this may represent an annoying overhead. |
2224 | The JPEG standard therefore permits tables to be omitted. The standard | 2224 | The JPEG standard therefore permits tables to be omitted. The standard |
2225 | defines three classes of JPEG datastreams: | 2225 | defines three classes of JPEG datastreams: |
2226 | * "Interchange" datastreams contain an image and all tables needed to decode | 2226 | * "Interchange" datastreams contain an image and all tables needed to decode |
2227 | the image. These are the usual kind of JPEG file. | 2227 | the image. These are the usual kind of JPEG file. |
2228 | * "Abbreviated image" datastreams contain an image, but are missing some or | 2228 | * "Abbreviated image" datastreams contain an image, but are missing some or |
2229 | all of the tables needed to decode that image. | 2229 | all of the tables needed to decode that image. |
2230 | * "Abbreviated table specification" (henceforth "tables-only") datastreams | 2230 | * "Abbreviated table specification" (henceforth "tables-only") datastreams |
2231 | contain only table specifications. | 2231 | contain only table specifications. |
2232 | To decode an abbreviated image, it is necessary to load the missing table(s) | 2232 | To decode an abbreviated image, it is necessary to load the missing table(s) |
2233 | into the decoder beforehand. This can be accomplished by reading a separate | 2233 | into the decoder beforehand. This can be accomplished by reading a separate |
2234 | tables-only file. A variant scheme uses a series of images in which the first | 2234 | tables-only file. A variant scheme uses a series of images in which the first |
2235 | image is an interchange (complete) datastream, while subsequent ones are | 2235 | image is an interchange (complete) datastream, while subsequent ones are |
2236 | abbreviated and rely on the tables loaded by the first image. It is assumed | 2236 | abbreviated and rely on the tables loaded by the first image. It is assumed |
2237 | that once the decoder has read a table, it will remember that table until a | 2237 | that once the decoder has read a table, it will remember that table until a |
2238 | new definition for the same table number is encountered. | 2238 | new definition for the same table number is encountered. |
2239 | 2239 | ||
2240 | It is the application designer's responsibility to figure out how to associate | 2240 | It is the application designer's responsibility to figure out how to associate |
2241 | the correct tables with an abbreviated image. While abbreviated datastreams | 2241 | the correct tables with an abbreviated image. While abbreviated datastreams |
2242 | can be useful in a closed environment, their use is strongly discouraged in | 2242 | can be useful in a closed environment, their use is strongly discouraged in |
2243 | any situation where data exchange with other applications might be needed. | 2243 | any situation where data exchange with other applications might be needed. |
2244 | Caveat designer. | 2244 | Caveat designer. |
2245 | 2245 | ||
2246 | The JPEG library provides support for reading and writing any combination of | 2246 | The JPEG library provides support for reading and writing any combination of |
2247 | tables-only datastreams and abbreviated images. In both compression and | 2247 | tables-only datastreams and abbreviated images. In both compression and |
2248 | decompression objects, a quantization or Huffman table will be retained for | 2248 | decompression objects, a quantization or Huffman table will be retained for |
2249 | the lifetime of the object, unless it is overwritten by a new table definition. | 2249 | the lifetime of the object, unless it is overwritten by a new table definition. |
2250 | 2250 | ||
2251 | 2251 | ||
2252 | To create abbreviated image datastreams, it is only necessary to tell the | 2252 | To create abbreviated image datastreams, it is only necessary to tell the |
2253 | compressor not to emit some or all of the tables it is using. Each | 2253 | compressor not to emit some or all of the tables it is using. Each |
2254 | quantization and Huffman table struct contains a boolean field "sent_table", | 2254 | quantization and Huffman table struct contains a boolean field "sent_table", |
2255 | which normally is initialized to FALSE. For each table used by the image, the | 2255 | which normally is initialized to FALSE. For each table used by the image, the |
2256 | header-writing process emits the table and sets sent_table = TRUE unless it is | 2256 | header-writing process emits the table and sets sent_table = TRUE unless it is |
2257 | already TRUE. (In normal usage, this prevents outputting the same table | 2257 | already TRUE. (In normal usage, this prevents outputting the same table |
2258 | definition multiple times, as would otherwise occur because the chroma | 2258 | definition multiple times, as would otherwise occur because the chroma |
2259 | components typically share tables.) Thus, setting this field to TRUE before | 2259 | components typically share tables.) Thus, setting this field to TRUE before |
2260 | calling jpeg_start_compress() will prevent the table from being written at | 2260 | calling jpeg_start_compress() will prevent the table from being written at |
2261 | all. | 2261 | all. |
2262 | 2262 | ||
2263 | If you want to create a "pure" abbreviated image file containing no tables, | 2263 | If you want to create a "pure" abbreviated image file containing no tables, |
2264 | just call "jpeg_suppress_tables(&cinfo, TRUE)" after constructing all the | 2264 | just call "jpeg_suppress_tables(&cinfo, TRUE)" after constructing all the |
2265 | tables. If you want to emit some but not all tables, you'll need to set the | 2265 | tables. If you want to emit some but not all tables, you'll need to set the |
2266 | individual sent_table fields directly. | 2266 | individual sent_table fields directly. |
2267 | 2267 | ||
2268 | To create an abbreviated image, you must also call jpeg_start_compress() | 2268 | To create an abbreviated image, you must also call jpeg_start_compress() |
2269 | with a second parameter of FALSE, not TRUE. Otherwise jpeg_start_compress() | 2269 | with a second parameter of FALSE, not TRUE. Otherwise jpeg_start_compress() |
2270 | will force all the sent_table fields to FALSE. (This is a safety feature to | 2270 | will force all the sent_table fields to FALSE. (This is a safety feature to |
2271 | prevent abbreviated images from being created accidentally.) | 2271 | prevent abbreviated images from being created accidentally.) |
2272 | 2272 | ||
2273 | To create a tables-only file, perform the same parameter setup that you | 2273 | To create a tables-only file, perform the same parameter setup that you |
2274 | normally would, but instead of calling jpeg_start_compress() and so on, call | 2274 | normally would, but instead of calling jpeg_start_compress() and so on, call |
2275 | jpeg_write_tables(&cinfo). This will write an abbreviated datastream | 2275 | jpeg_write_tables(&cinfo). This will write an abbreviated datastream |
2276 | containing only SOI, DQT and/or DHT markers, and EOI. All the quantization | 2276 | containing only SOI, DQT and/or DHT markers, and EOI. All the quantization |
2277 | and Huffman tables that are currently defined in the compression object will | 2277 | and Huffman tables that are currently defined in the compression object will |
2278 | be emitted unless their sent_tables flag is already TRUE, and then all the | 2278 | be emitted unless their sent_tables flag is already TRUE, and then all the |
2279 | sent_tables flags will be set TRUE. | 2279 | sent_tables flags will be set TRUE. |
2280 | 2280 | ||
2281 | A sure-fire way to create matching tables-only and abbreviated image files | 2281 | A sure-fire way to create matching tables-only and abbreviated image files |
2282 | is to proceed as follows: | 2282 | is to proceed as follows: |
2283 | 2283 | ||
2284 | create JPEG compression object | 2284 | create JPEG compression object |
2285 | set JPEG parameters | 2285 | set JPEG parameters |
2286 | set destination to tables-only file | 2286 | set destination to tables-only file |
2287 | jpeg_write_tables(&cinfo); | 2287 | jpeg_write_tables(&cinfo); |
2288 | set destination to image file | 2288 | set destination to image file |
2289 | jpeg_start_compress(&cinfo, FALSE); | 2289 | jpeg_start_compress(&cinfo, FALSE); |
2290 | write data... | 2290 | write data... |
2291 | jpeg_finish_compress(&cinfo); | 2291 | jpeg_finish_compress(&cinfo); |
2292 | 2292 | ||
2293 | Since the JPEG parameters are not altered between writing the table file and | 2293 | Since the JPEG parameters are not altered between writing the table file and |
2294 | the abbreviated image file, the same tables are sure to be used. Of course, | 2294 | the abbreviated image file, the same tables are sure to be used. Of course, |
2295 | you can repeat the jpeg_start_compress() ... jpeg_finish_compress() sequence | 2295 | you can repeat the jpeg_start_compress() ... jpeg_finish_compress() sequence |
2296 | many times to produce many abbreviated image files matching the table file. | 2296 | many times to produce many abbreviated image files matching the table file. |
2297 | 2297 | ||
2298 | You cannot suppress output of the computed Huffman tables when Huffman | 2298 | You cannot suppress output of the computed Huffman tables when Huffman |
2299 | optimization is selected. (If you could, there'd be no way to decode the | 2299 | optimization is selected. (If you could, there'd be no way to decode the |
2300 | image...) Generally, you don't want to set optimize_coding = TRUE when | 2300 | image...) Generally, you don't want to set optimize_coding = TRUE when |
2301 | you are trying to produce abbreviated files. | 2301 | you are trying to produce abbreviated files. |
2302 | 2302 | ||
2303 | In some cases you might want to compress an image using tables which are | 2303 | In some cases you might want to compress an image using tables which are |
2304 | not stored in the application, but are defined in an interchange or | 2304 | not stored in the application, but are defined in an interchange or |
2305 | tables-only file readable by the application. This can be done by setting up | 2305 | tables-only file readable by the application. This can be done by setting up |
2306 | a JPEG decompression object to read the specification file, then copying the | 2306 | a JPEG decompression object to read the specification file, then copying the |
2307 | tables into your compression object. See jpeg_copy_critical_parameters() | 2307 | tables into your compression object. See jpeg_copy_critical_parameters() |
2308 | for an example of copying quantization tables. | 2308 | for an example of copying quantization tables. |
2309 | 2309 | ||
2310 | 2310 | ||
2311 | To read abbreviated image files, you simply need to load the proper tables | 2311 | To read abbreviated image files, you simply need to load the proper tables |
2312 | into the decompression object before trying to read the abbreviated image. | 2312 | into the decompression object before trying to read the abbreviated image. |
2313 | If the proper tables are stored in the application program, you can just | 2313 | If the proper tables are stored in the application program, you can just |
2314 | allocate the table structs and fill in their contents directly. For example, | 2314 | allocate the table structs and fill in their contents directly. For example, |
2315 | to load a fixed quantization table into table slot "n": | 2315 | to load a fixed quantization table into table slot "n": |
2316 | 2316 | ||
2317 | if (cinfo.quant_tbl_ptrs[n] == NULL) | 2317 | if (cinfo.quant_tbl_ptrs[n] == NULL) |
2318 | cinfo.quant_tbl_ptrs[n] = jpeg_alloc_quant_table((j_common_ptr) &cinfo); | 2318 | cinfo.quant_tbl_ptrs[n] = jpeg_alloc_quant_table((j_common_ptr) &cinfo); |
2319 | quant_ptr = cinfo.quant_tbl_ptrs[n]; /* quant_ptr is JQUANT_TBL* */ | 2319 | quant_ptr = cinfo.quant_tbl_ptrs[n]; /* quant_ptr is JQUANT_TBL* */ |
2320 | for (i = 0; i < 64; i++) { | 2320 | for (i = 0; i < 64; i++) { |
2321 | /* Qtable[] is desired quantization table, in natural array order */ | 2321 | /* Qtable[] is desired quantization table, in natural array order */ |
2322 | quant_ptr->quantval[i] = Qtable[i]; | 2322 | quant_ptr->quantval[i] = Qtable[i]; |
2323 | } | 2323 | } |
2324 | 2324 | ||
2325 | Code to load a fixed Huffman table is typically (for AC table "n"): | 2325 | Code to load a fixed Huffman table is typically (for AC table "n"): |
2326 | 2326 | ||
2327 | if (cinfo.ac_huff_tbl_ptrs[n] == NULL) | 2327 | if (cinfo.ac_huff_tbl_ptrs[n] == NULL) |
2328 | cinfo.ac_huff_tbl_ptrs[n] = jpeg_alloc_huff_table((j_common_ptr) &cinfo); | 2328 | cinfo.ac_huff_tbl_ptrs[n] = jpeg_alloc_huff_table((j_common_ptr) &cinfo); |
2329 | huff_ptr = cinfo.ac_huff_tbl_ptrs[n]; /* huff_ptr is JHUFF_TBL* */ | 2329 | huff_ptr = cinfo.ac_huff_tbl_ptrs[n]; /* huff_ptr is JHUFF_TBL* */ |
2330 | for (i = 1; i <= 16; i++) { | 2330 | for (i = 1; i <= 16; i++) { |
2331 | /* counts[i] is number of Huffman codes of length i bits, i=1..16 */ | 2331 | /* counts[i] is number of Huffman codes of length i bits, i=1..16 */ |
2332 | huff_ptr->bits[i] = counts[i]; | 2332 | huff_ptr->bits[i] = counts[i]; |
2333 | } | 2333 | } |
2334 | for (i = 0; i < 256; i++) { | 2334 | for (i = 0; i < 256; i++) { |
2335 | /* symbols[] is the list of Huffman symbols, in code-length order */ | 2335 | /* symbols[] is the list of Huffman symbols, in code-length order */ |
2336 | huff_ptr->huffval[i] = symbols[i]; | 2336 | huff_ptr->huffval[i] = symbols[i]; |
2337 | } | 2337 | } |
2338 | 2338 | ||
2339 | (Note that trying to set cinfo.quant_tbl_ptrs[n] to point directly at a | 2339 | (Note that trying to set cinfo.quant_tbl_ptrs[n] to point directly at a |
2340 | constant JQUANT_TBL object is not safe. If the incoming file happened to | 2340 | constant JQUANT_TBL object is not safe. If the incoming file happened to |
2341 | contain a quantization table definition, your master table would get | 2341 | contain a quantization table definition, your master table would get |
2342 | overwritten! Instead allocate a working table copy and copy the master table | 2342 | overwritten! Instead allocate a working table copy and copy the master table |
2343 | into it, as illustrated above. Ditto for Huffman tables, of course.) | 2343 | into it, as illustrated above. Ditto for Huffman tables, of course.) |
2344 | 2344 | ||
2345 | You might want to read the tables from a tables-only file, rather than | 2345 | You might want to read the tables from a tables-only file, rather than |
2346 | hard-wiring them into your application. The jpeg_read_header() call is | 2346 | hard-wiring them into your application. The jpeg_read_header() call is |
2347 | sufficient to read a tables-only file. You must pass a second parameter of | 2347 | sufficient to read a tables-only file. You must pass a second parameter of |
2348 | FALSE to indicate that you do not require an image to be present. Thus, the | 2348 | FALSE to indicate that you do not require an image to be present. Thus, the |
2349 | typical scenario is | 2349 | typical scenario is |
2350 | 2350 | ||
2351 | create JPEG decompression object | 2351 | create JPEG decompression object |
2352 | set source to tables-only file | 2352 | set source to tables-only file |
2353 | jpeg_read_header(&cinfo, FALSE); | 2353 | jpeg_read_header(&cinfo, FALSE); |
2354 | set source to abbreviated image file | 2354 | set source to abbreviated image file |
2355 | jpeg_read_header(&cinfo, TRUE); | 2355 | jpeg_read_header(&cinfo, TRUE); |
2356 | set decompression parameters | 2356 | set decompression parameters |
2357 | jpeg_start_decompress(&cinfo); | 2357 | jpeg_start_decompress(&cinfo); |
2358 | read data... | 2358 | read data... |
2359 | jpeg_finish_decompress(&cinfo); | 2359 | jpeg_finish_decompress(&cinfo); |
2360 | 2360 | ||
2361 | In some cases, you may want to read a file without knowing whether it contains | 2361 | In some cases, you may want to read a file without knowing whether it contains |
2362 | an image or just tables. In that case, pass FALSE and check the return value | 2362 | an image or just tables. In that case, pass FALSE and check the return value |
2363 | from jpeg_read_header(): it will be JPEG_HEADER_OK if an image was found, | 2363 | from jpeg_read_header(): it will be JPEG_HEADER_OK if an image was found, |
2364 | JPEG_HEADER_TABLES_ONLY if only tables were found. (A third return value, | 2364 | JPEG_HEADER_TABLES_ONLY if only tables were found. (A third return value, |
2365 | JPEG_SUSPENDED, is possible when using a suspending data source manager.) | 2365 | JPEG_SUSPENDED, is possible when using a suspending data source manager.) |
2366 | Note that jpeg_read_header() will not complain if you read an abbreviated | 2366 | Note that jpeg_read_header() will not complain if you read an abbreviated |
2367 | image for which you haven't loaded the missing tables; the missing-table check | 2367 | image for which you haven't loaded the missing tables; the missing-table check |
2368 | occurs later, in jpeg_start_decompress(). | 2368 | occurs later, in jpeg_start_decompress(). |
2369 | 2369 | ||
2370 | 2370 | ||
2371 | It is possible to read a series of images from a single source file by | 2371 | It is possible to read a series of images from a single source file by |
2372 | repeating the jpeg_read_header() ... jpeg_finish_decompress() sequence, | 2372 | repeating the jpeg_read_header() ... jpeg_finish_decompress() sequence, |
2373 | without releasing/recreating the JPEG object or the data source module. | 2373 | without releasing/recreating the JPEG object or the data source module. |
2374 | (If you did reinitialize, any partial bufferload left in the data source | 2374 | (If you did reinitialize, any partial bufferload left in the data source |
2375 | buffer at the end of one image would be discarded, causing you to lose the | 2375 | buffer at the end of one image would be discarded, causing you to lose the |
2376 | start of the next image.) When you use this method, stored tables are | 2376 | start of the next image.) When you use this method, stored tables are |
2377 | automatically carried forward, so some of the images can be abbreviated images | 2377 | automatically carried forward, so some of the images can be abbreviated images |
2378 | that depend on tables from earlier images. | 2378 | that depend on tables from earlier images. |
2379 | 2379 | ||
2380 | If you intend to write a series of images into a single destination file, | 2380 | If you intend to write a series of images into a single destination file, |
2381 | you might want to make a specialized data destination module that doesn't | 2381 | you might want to make a specialized data destination module that doesn't |
2382 | flush the output buffer at term_destination() time. This would speed things | 2382 | flush the output buffer at term_destination() time. This would speed things |
2383 | up by some trifling amount. Of course, you'd need to remember to flush the | 2383 | up by some trifling amount. Of course, you'd need to remember to flush the |
2384 | buffer after the last image. You can make the later images be abbreviated | 2384 | buffer after the last image. You can make the later images be abbreviated |
2385 | ones by passing FALSE to jpeg_start_compress(). | 2385 | ones by passing FALSE to jpeg_start_compress(). |
2386 | 2386 | ||
2387 | 2387 | ||
2388 | Special markers | 2388 | Special markers |
2389 | --------------- | 2389 | --------------- |
2390 | 2390 | ||
2391 | Some applications may need to insert or extract special data in the JPEG | 2391 | Some applications may need to insert or extract special data in the JPEG |
2392 | datastream. The JPEG standard provides marker types "COM" (comment) and | 2392 | datastream. The JPEG standard provides marker types "COM" (comment) and |
2393 | "APP0" through "APP15" (application) to hold application-specific data. | 2393 | "APP0" through "APP15" (application) to hold application-specific data. |
2394 | Unfortunately, the use of these markers is not specified by the standard. | 2394 | Unfortunately, the use of these markers is not specified by the standard. |
2395 | COM markers are fairly widely used to hold user-supplied text. The JFIF file | 2395 | COM markers are fairly widely used to hold user-supplied text. The JFIF file |
2396 | format spec uses APP0 markers with specified initial strings to hold certain | 2396 | format spec uses APP0 markers with specified initial strings to hold certain |
2397 | data. Adobe applications use APP14 markers beginning with the string "Adobe" | 2397 | data. Adobe applications use APP14 markers beginning with the string "Adobe" |
2398 | for miscellaneous data. Other APPn markers are rarely seen, but might | 2398 | for miscellaneous data. Other APPn markers are rarely seen, but might |
2399 | contain almost anything. | 2399 | contain almost anything. |
2400 | 2400 | ||
2401 | If you wish to store user-supplied text, we recommend you use COM markers | 2401 | If you wish to store user-supplied text, we recommend you use COM markers |
2402 | and place readable 7-bit ASCII text in them. Newline conventions are not | 2402 | and place readable 7-bit ASCII text in them. Newline conventions are not |
2403 | standardized --- expect to find LF (Unix style), CR/LF (DOS style), or CR | 2403 | standardized --- expect to find LF (Unix style), CR/LF (DOS style), or CR |
2404 | (Mac style). A robust COM reader should be able to cope with random binary | 2404 | (Mac style). A robust COM reader should be able to cope with random binary |
2405 | garbage, including nulls, since some applications generate COM markers | 2405 | garbage, including nulls, since some applications generate COM markers |
2406 | containing non-ASCII junk. (But yours should not be one of them.) | 2406 | containing non-ASCII junk. (But yours should not be one of them.) |
2407 | 2407 | ||
2408 | For program-supplied data, use an APPn marker, and be sure to begin it with an | 2408 | For program-supplied data, use an APPn marker, and be sure to begin it with an |
2409 | identifying string so that you can tell whether the marker is actually yours. | 2409 | identifying string so that you can tell whether the marker is actually yours. |
2410 | It's probably best to avoid using APP0 or APP14 for any private markers. | 2410 | It's probably best to avoid using APP0 or APP14 for any private markers. |
2411 | (NOTE: the upcoming SPIFF standard will use APP8 markers; we recommend you | 2411 | (NOTE: the upcoming SPIFF standard will use APP8 markers; we recommend you |
2412 | not use APP8 markers for any private purposes, either.) | 2412 | not use APP8 markers for any private purposes, either.) |
2413 | 2413 | ||
2414 | Keep in mind that at most 65533 bytes can be put into one marker, but you | 2414 | Keep in mind that at most 65533 bytes can be put into one marker, but you |
2415 | can have as many markers as you like. | 2415 | can have as many markers as you like. |
2416 | 2416 | ||
2417 | By default, the IJG compression library will write a JFIF APP0 marker if the | 2417 | By default, the IJG compression library will write a JFIF APP0 marker if the |
2418 | selected JPEG colorspace is grayscale or YCbCr, or an Adobe APP14 marker if | 2418 | selected JPEG colorspace is grayscale or YCbCr, or an Adobe APP14 marker if |
2419 | the selected colorspace is RGB, CMYK, or YCCK. You can disable this, but | 2419 | the selected colorspace is RGB, CMYK, or YCCK. You can disable this, but |
2420 | we don't recommend it. The decompression library will recognize JFIF and | 2420 | we don't recommend it. The decompression library will recognize JFIF and |
2421 | Adobe markers and will set the JPEG colorspace properly when one is found. | 2421 | Adobe markers and will set the JPEG colorspace properly when one is found. |
2422 | 2422 | ||
2423 | 2423 | ||
2424 | You can write special markers immediately following the datastream header by | 2424 | You can write special markers immediately following the datastream header by |
2425 | calling jpeg_write_marker() after jpeg_start_compress() and before the first | 2425 | calling jpeg_write_marker() after jpeg_start_compress() and before the first |
2426 | call to jpeg_write_scanlines(). When you do this, the markers appear after | 2426 | call to jpeg_write_scanlines(). When you do this, the markers appear after |
2427 | the SOI and the JFIF APP0 and Adobe APP14 markers (if written), but before | 2427 | the SOI and the JFIF APP0 and Adobe APP14 markers (if written), but before |
2428 | all else. Specify the marker type parameter as "JPEG_COM" for COM or | 2428 | all else. Specify the marker type parameter as "JPEG_COM" for COM or |
2429 | "JPEG_APP0 + n" for APPn. (Actually, jpeg_write_marker will let you write | 2429 | "JPEG_APP0 + n" for APPn. (Actually, jpeg_write_marker will let you write |
2430 | any marker type, but we don't recommend writing any other kinds of marker.) | 2430 | any marker type, but we don't recommend writing any other kinds of marker.) |
2431 | For example, to write a user comment string pointed to by comment_text: | 2431 | For example, to write a user comment string pointed to by comment_text: |
2432 | jpeg_write_marker(cinfo, JPEG_COM, comment_text, strlen(comment_text)); | 2432 | jpeg_write_marker(cinfo, JPEG_COM, comment_text, strlen(comment_text)); |
2433 | 2433 | ||
2434 | If it's not convenient to store all the marker data in memory at once, | 2434 | If it's not convenient to store all the marker data in memory at once, |
2435 | you can instead call jpeg_write_m_header() followed by multiple calls to | 2435 | you can instead call jpeg_write_m_header() followed by multiple calls to |
2436 | jpeg_write_m_byte(). If you do it this way, it's your responsibility to | 2436 | jpeg_write_m_byte(). If you do it this way, it's your responsibility to |
2437 | call jpeg_write_m_byte() exactly the number of times given in the length | 2437 | call jpeg_write_m_byte() exactly the number of times given in the length |
2438 | parameter to jpeg_write_m_header(). (This method lets you empty the | 2438 | parameter to jpeg_write_m_header(). (This method lets you empty the |
2439 | output buffer partway through a marker, which might be important when | 2439 | output buffer partway through a marker, which might be important when |
2440 | using a suspending data destination module. In any case, if you are using | 2440 | using a suspending data destination module. In any case, if you are using |
2441 | a suspending destination, you should flush its buffer after inserting | 2441 | a suspending destination, you should flush its buffer after inserting |
2442 | any special markers. See "I/O suspension".) | 2442 | any special markers. See "I/O suspension".) |
2443 | 2443 | ||
2444 | Or, if you prefer to synthesize the marker byte sequence yourself, | 2444 | Or, if you prefer to synthesize the marker byte sequence yourself, |
2445 | you can just cram it straight into the data destination module. | 2445 | you can just cram it straight into the data destination module. |
2446 | 2446 | ||
2447 | If you are writing JFIF 1.02 extension markers (thumbnail images), don't | 2447 | If you are writing JFIF 1.02 extension markers (thumbnail images), don't |
2448 | forget to set cinfo.JFIF_minor_version = 2 so that the encoder will write the | 2448 | forget to set cinfo.JFIF_minor_version = 2 so that the encoder will write the |
2449 | correct JFIF version number in the JFIF header marker. The library's default | 2449 | correct JFIF version number in the JFIF header marker. The library's default |
2450 | is to write version 1.01, but that's wrong if you insert any 1.02 extension | 2450 | is to write version 1.01, but that's wrong if you insert any 1.02 extension |
2451 | markers. (We could probably get away with just defaulting to 1.02, but there | 2451 | markers. (We could probably get away with just defaulting to 1.02, but there |
2452 | used to be broken decoders that would complain about unknown minor version | 2452 | used to be broken decoders that would complain about unknown minor version |
2453 | numbers. To reduce compatibility risks it's safest not to write 1.02 unless | 2453 | numbers. To reduce compatibility risks it's safest not to write 1.02 unless |
2454 | you are actually using 1.02 extensions.) | 2454 | you are actually using 1.02 extensions.) |
2455 | 2455 | ||
2456 | 2456 | ||
2457 | When reading, two methods of handling special markers are available: | 2457 | When reading, two methods of handling special markers are available: |
2458 | 1. You can ask the library to save the contents of COM and/or APPn markers | 2458 | 1. You can ask the library to save the contents of COM and/or APPn markers |
2459 | into memory, and then examine them at your leisure afterwards. | 2459 | into memory, and then examine them at your leisure afterwards. |
2460 | 2. You can supply your own routine to process COM and/or APPn markers | 2460 | 2. You can supply your own routine to process COM and/or APPn markers |
2461 | on-the-fly as they are read. | 2461 | on-the-fly as they are read. |
2462 | The first method is simpler to use, especially if you are using a suspending | 2462 | The first method is simpler to use, especially if you are using a suspending |
2463 | data source; writing a marker processor that copes with input suspension is | 2463 | data source; writing a marker processor that copes with input suspension is |
2464 | not easy (consider what happens if the marker is longer than your available | 2464 | not easy (consider what happens if the marker is longer than your available |
2465 | input buffer). However, the second method conserves memory since the marker | 2465 | input buffer). However, the second method conserves memory since the marker |
2466 | data need not be kept around after it's been processed. | 2466 | data need not be kept around after it's been processed. |
2467 | 2467 | ||
2468 | For either method, you'd normally set up marker handling after creating a | 2468 | For either method, you'd normally set up marker handling after creating a |
2469 | decompression object and before calling jpeg_read_header(), because the | 2469 | decompression object and before calling jpeg_read_header(), because the |
2470 | markers of interest will typically be near the head of the file and so will | 2470 | markers of interest will typically be near the head of the file and so will |
2471 | be scanned by jpeg_read_header. Once you've established a marker handling | 2471 | be scanned by jpeg_read_header. Once you've established a marker handling |
2472 | method, it will be used for the life of that decompression object | 2472 | method, it will be used for the life of that decompression object |
2473 | (potentially many datastreams), unless you change it. Marker handling is | 2473 | (potentially many datastreams), unless you change it. Marker handling is |
2474 | determined separately for COM markers and for each APPn marker code. | 2474 | determined separately for COM markers and for each APPn marker code. |
2475 | 2475 | ||
2476 | 2476 | ||
2477 | To save the contents of special markers in memory, call | 2477 | To save the contents of special markers in memory, call |
2478 | jpeg_save_markers(cinfo, marker_code, length_limit) | 2478 | jpeg_save_markers(cinfo, marker_code, length_limit) |
2479 | where marker_code is the marker type to save, JPEG_COM or JPEG_APP0+n. | 2479 | where marker_code is the marker type to save, JPEG_COM or JPEG_APP0+n. |
2480 | (To arrange to save all the special marker types, you need to call this | 2480 | (To arrange to save all the special marker types, you need to call this |
2481 | routine 17 times, for COM and APP0-APP15.) If the incoming marker is longer | 2481 | routine 17 times, for COM and APP0-APP15.) If the incoming marker is longer |
2482 | than length_limit data bytes, only length_limit bytes will be saved; this | 2482 | than length_limit data bytes, only length_limit bytes will be saved; this |
2483 | parameter allows you to avoid chewing up memory when you only need to see the | 2483 | parameter allows you to avoid chewing up memory when you only need to see the |
2484 | first few bytes of a potentially large marker. If you want to save all the | 2484 | first few bytes of a potentially large marker. If you want to save all the |
2485 | data, set length_limit to 0xFFFF; that is enough since marker lengths are only | 2485 | data, set length_limit to 0xFFFF; that is enough since marker lengths are only |
2486 | 16 bits. As a special case, setting length_limit to 0 prevents that marker | 2486 | 16 bits. As a special case, setting length_limit to 0 prevents that marker |
2487 | type from being saved at all. (That is the default behavior, in fact.) | 2487 | type from being saved at all. (That is the default behavior, in fact.) |
2488 | 2488 | ||
2489 | After jpeg_read_header() completes, you can examine the special markers by | 2489 | After jpeg_read_header() completes, you can examine the special markers by |
2490 | following the cinfo->marker_list pointer chain. All the special markers in | 2490 | following the cinfo->marker_list pointer chain. All the special markers in |
2491 | the file appear in this list, in order of their occurrence in the file (but | 2491 | the file appear in this list, in order of their occurrence in the file (but |
2492 | omitting any markers of types you didn't ask for). Both the original data | 2492 | omitting any markers of types you didn't ask for). Both the original data |
2493 | length and the saved data length are recorded for each list entry; the latter | 2493 | length and the saved data length are recorded for each list entry; the latter |
2494 | will not exceed length_limit for the particular marker type. Note that these | 2494 | will not exceed length_limit for the particular marker type. Note that these |
2495 | lengths exclude the marker length word, whereas the stored representation | 2495 | lengths exclude the marker length word, whereas the stored representation |
2496 | within the JPEG file includes it. (Hence the maximum data length is really | 2496 | within the JPEG file includes it. (Hence the maximum data length is really |
2497 | only 65533.) | 2497 | only 65533.) |
2498 | 2498 | ||
2499 | It is possible that additional special markers appear in the file beyond the | 2499 | It is possible that additional special markers appear in the file beyond the |
2500 | SOS marker at which jpeg_read_header stops; if so, the marker list will be | 2500 | SOS marker at which jpeg_read_header stops; if so, the marker list will be |
2501 | extended during reading of the rest of the file. This is not expected to be | 2501 | extended during reading of the rest of the file. This is not expected to be |
2502 | common, however. If you are short on memory you may want to reset the length | 2502 | common, however. If you are short on memory you may want to reset the length |
2503 | limit to zero for all marker types after finishing jpeg_read_header, to | 2503 | limit to zero for all marker types after finishing jpeg_read_header, to |
2504 | ensure that the max_memory_to_use setting cannot be exceeded due to addition | 2504 | ensure that the max_memory_to_use setting cannot be exceeded due to addition |
2505 | of later markers. | 2505 | of later markers. |
2506 | 2506 | ||
2507 | The marker list remains stored until you call jpeg_finish_decompress or | 2507 | The marker list remains stored until you call jpeg_finish_decompress or |
2508 | jpeg_abort, at which point the memory is freed and the list is set to empty. | 2508 | jpeg_abort, at which point the memory is freed and the list is set to empty. |
2509 | (jpeg_destroy also releases the storage, of course.) | 2509 | (jpeg_destroy also releases the storage, of course.) |
2510 | 2510 | ||
2511 | Note that the library is internally interested in APP0 and APP14 markers; | 2511 | Note that the library is internally interested in APP0 and APP14 markers; |
2512 | if you try to set a small nonzero length limit on these types, the library | 2512 | if you try to set a small nonzero length limit on these types, the library |
2513 | will silently force the length up to the minimum it wants. (But you can set | 2513 | will silently force the length up to the minimum it wants. (But you can set |
2514 | a zero length limit to prevent them from being saved at all.) Also, in a | 2514 | a zero length limit to prevent them from being saved at all.) Also, in a |
2515 | 16-bit environment, the maximum length limit may be constrained to less than | 2515 | 16-bit environment, the maximum length limit may be constrained to less than |
2516 | 65533 by malloc() limitations. It is therefore best not to assume that the | 2516 | 65533 by malloc() limitations. It is therefore best not to assume that the |
2517 | effective length limit is exactly what you set it to be. | 2517 | effective length limit is exactly what you set it to be. |
2518 | 2518 | ||
2519 | 2519 | ||
2520 | If you want to supply your own marker-reading routine, you do it by calling | 2520 | If you want to supply your own marker-reading routine, you do it by calling |
2521 | jpeg_set_marker_processor(). A marker processor routine must have the | 2521 | jpeg_set_marker_processor(). A marker processor routine must have the |
2522 | signature | 2522 | signature |
2523 | boolean jpeg_marker_parser_method (j_decompress_ptr cinfo) | 2523 | boolean jpeg_marker_parser_method (j_decompress_ptr cinfo) |
2524 | Although the marker code is not explicitly passed, the routine can find it | 2524 | Although the marker code is not explicitly passed, the routine can find it |
2525 | in cinfo->unread_marker. At the time of call, the marker proper has been | 2525 | in cinfo->unread_marker. At the time of call, the marker proper has been |
2526 | read from the data source module. The processor routine is responsible for | 2526 | read from the data source module. The processor routine is responsible for |
2527 | reading the marker length word and the remaining parameter bytes, if any. | 2527 | reading the marker length word and the remaining parameter bytes, if any. |
2528 | Return TRUE to indicate success. (FALSE should be returned only if you are | 2528 | Return TRUE to indicate success. (FALSE should be returned only if you are |
2529 | using a suspending data source and it tells you to suspend. See the standard | 2529 | using a suspending data source and it tells you to suspend. See the standard |
2530 | marker processors in jdmarker.c for appropriate coding methods if you need to | 2530 | marker processors in jdmarker.c for appropriate coding methods if you need to |
2531 | use a suspending data source.) | 2531 | use a suspending data source.) |
2532 | 2532 | ||
2533 | If you override the default APP0 or APP14 processors, it is up to you to | 2533 | If you override the default APP0 or APP14 processors, it is up to you to |
2534 | recognize JFIF and Adobe markers if you want colorspace recognition to occur | 2534 | recognize JFIF and Adobe markers if you want colorspace recognition to occur |
2535 | properly. We recommend copying and extending the default processors if you | 2535 | properly. We recommend copying and extending the default processors if you |
2536 | want to do that. (A better idea is to save these marker types for later | 2536 | want to do that. (A better idea is to save these marker types for later |
2537 | examination by calling jpeg_save_markers(); that method doesn't interfere | 2537 | examination by calling jpeg_save_markers(); that method doesn't interfere |
2538 | with the library's own processing of these markers.) | 2538 | with the library's own processing of these markers.) |
2539 | 2539 | ||
2540 | jpeg_set_marker_processor() and jpeg_save_markers() are mutually exclusive | 2540 | jpeg_set_marker_processor() and jpeg_save_markers() are mutually exclusive |
2541 | --- if you call one it overrides any previous call to the other, for the | 2541 | --- if you call one it overrides any previous call to the other, for the |
2542 | particular marker type specified. | 2542 | particular marker type specified. |
2543 | 2543 | ||
2544 | A simple example of an external COM processor can be found in djpeg.c. | 2544 | A simple example of an external COM processor can be found in djpeg.c. |
2545 | Also, see jpegtran.c for an example of using jpeg_save_markers. | 2545 | Also, see jpegtran.c for an example of using jpeg_save_markers. |
2546 | 2546 | ||
2547 | 2547 | ||
2548 | Raw (downsampled) image data | 2548 | Raw (downsampled) image data |
2549 | ---------------------------- | 2549 | ---------------------------- |
2550 | 2550 | ||
2551 | Some applications need to supply already-downsampled image data to the JPEG | 2551 | Some applications need to supply already-downsampled image data to the JPEG |
2552 | compressor, or to receive raw downsampled data from the decompressor. The | 2552 | compressor, or to receive raw downsampled data from the decompressor. The |
2553 | library supports this requirement by allowing the application to write or | 2553 | library supports this requirement by allowing the application to write or |
2554 | read raw data, bypassing the normal preprocessing or postprocessing steps. | 2554 | read raw data, bypassing the normal preprocessing or postprocessing steps. |
2555 | The interface is different from the standard one and is somewhat harder to | 2555 | The interface is different from the standard one and is somewhat harder to |
2556 | use. If your interest is merely in bypassing color conversion, we recommend | 2556 | use. If your interest is merely in bypassing color conversion, we recommend |
2557 | that you use the standard interface and simply set jpeg_color_space = | 2557 | that you use the standard interface and simply set jpeg_color_space = |
2558 | in_color_space (or jpeg_color_space = out_color_space for decompression). | 2558 | in_color_space (or jpeg_color_space = out_color_space for decompression). |
2559 | The mechanism described in this section is necessary only to supply or | 2559 | The mechanism described in this section is necessary only to supply or |
2560 | receive downsampled image data, in which not all components have the same | 2560 | receive downsampled image data, in which not all components have the same |
2561 | dimensions. | 2561 | dimensions. |
2562 | 2562 | ||
2563 | 2563 | ||
2564 | To compress raw data, you must supply the data in the colorspace to be used | 2564 | To compress raw data, you must supply the data in the colorspace to be used |
2565 | in the JPEG file (please read the earlier section on Special color spaces) | 2565 | in the JPEG file (please read the earlier section on Special color spaces) |
2566 | and downsampled to the sampling factors specified in the JPEG parameters. | 2566 | and downsampled to the sampling factors specified in the JPEG parameters. |
2567 | You must supply the data in the format used internally by the JPEG library, | 2567 | You must supply the data in the format used internally by the JPEG library, |
2568 | namely a JSAMPIMAGE array. This is an array of pointers to two-dimensional | 2568 | namely a JSAMPIMAGE array. This is an array of pointers to two-dimensional |
2569 | arrays, each of type JSAMPARRAY. Each 2-D array holds the values for one | 2569 | arrays, each of type JSAMPARRAY. Each 2-D array holds the values for one |
2570 | color component. This structure is necessary since the components are of | 2570 | color component. This structure is necessary since the components are of |
2571 | different sizes. If the image dimensions are not a multiple of the MCU size, | 2571 | different sizes. If the image dimensions are not a multiple of the MCU size, |
2572 | you must also pad the data correctly (usually, this is done by replicating | 2572 | you must also pad the data correctly (usually, this is done by replicating |
2573 | the last column and/or row). The data must be padded to a multiple of a DCT | 2573 | the last column and/or row). The data must be padded to a multiple of a DCT |
2574 | block in each component: that is, each downsampled row must contain a | 2574 | block in each component: that is, each downsampled row must contain a |
2575 | multiple of 8 valid samples, and there must be a multiple of 8 sample rows | 2575 | multiple of 8 valid samples, and there must be a multiple of 8 sample rows |
2576 | for each component. (For applications such as conversion of digital TV | 2576 | for each component. (For applications such as conversion of digital TV |
2577 | images, the standard image size is usually a multiple of the DCT block size, | 2577 | images, the standard image size is usually a multiple of the DCT block size, |
2578 | so that no padding need actually be done.) | 2578 | so that no padding need actually be done.) |
2579 | 2579 | ||
2580 | The procedure for compression of raw data is basically the same as normal | 2580 | The procedure for compression of raw data is basically the same as normal |
2581 | compression, except that you call jpeg_write_raw_data() in place of | 2581 | compression, except that you call jpeg_write_raw_data() in place of |
2582 | jpeg_write_scanlines(). Before calling jpeg_start_compress(), you must do | 2582 | jpeg_write_scanlines(). Before calling jpeg_start_compress(), you must do |
2583 | the following: | 2583 | the following: |
2584 | * Set cinfo->raw_data_in to TRUE. (It is set FALSE by jpeg_set_defaults().) | 2584 | * Set cinfo->raw_data_in to TRUE. (It is set FALSE by jpeg_set_defaults().) |
2585 | This notifies the library that you will be supplying raw data. | 2585 | This notifies the library that you will be supplying raw data. |
2586 | Furthermore, set cinfo->do_fancy_downsampling to FALSE if you want to use | 2586 | Furthermore, set cinfo->do_fancy_downsampling to FALSE if you want to use |
2587 | real downsampled data. (It is set TRUE by jpeg_set_defaults().) | 2587 | real downsampled data. (It is set TRUE by jpeg_set_defaults().) |
2588 | * Ensure jpeg_color_space is correct --- an explicit jpeg_set_colorspace() | 2588 | * Ensure jpeg_color_space is correct --- an explicit jpeg_set_colorspace() |
2589 | call is a good idea. Note that since color conversion is bypassed, | 2589 | call is a good idea. Note that since color conversion is bypassed, |
2590 | in_color_space is ignored, except that jpeg_set_defaults() uses it to | 2590 | in_color_space is ignored, except that jpeg_set_defaults() uses it to |
2591 | choose the default jpeg_color_space setting. | 2591 | choose the default jpeg_color_space setting. |
2592 | * Ensure the sampling factors, cinfo->comp_info[i].h_samp_factor and | 2592 | * Ensure the sampling factors, cinfo->comp_info[i].h_samp_factor and |
2593 | cinfo->comp_info[i].v_samp_factor, are correct. Since these indicate the | 2593 | cinfo->comp_info[i].v_samp_factor, are correct. Since these indicate the |
2594 | dimensions of the data you are supplying, it's wise to set them | 2594 | dimensions of the data you are supplying, it's wise to set them |
2595 | explicitly, rather than assuming the library's defaults are what you want. | 2595 | explicitly, rather than assuming the library's defaults are what you want. |
2596 | 2596 | ||
2597 | To pass raw data to the library, call jpeg_write_raw_data() in place of | 2597 | To pass raw data to the library, call jpeg_write_raw_data() in place of |
2598 | jpeg_write_scanlines(). The two routines work similarly except that | 2598 | jpeg_write_scanlines(). The two routines work similarly except that |
2599 | jpeg_write_raw_data takes a JSAMPIMAGE data array rather than JSAMPARRAY. | 2599 | jpeg_write_raw_data takes a JSAMPIMAGE data array rather than JSAMPARRAY. |
2600 | The scanlines count passed to and returned from jpeg_write_raw_data is | 2600 | The scanlines count passed to and returned from jpeg_write_raw_data is |
2601 | measured in terms of the component with the largest v_samp_factor. | 2601 | measured in terms of the component with the largest v_samp_factor. |
2602 | 2602 | ||
2603 | jpeg_write_raw_data() processes one MCU row per call, which is to say | 2603 | jpeg_write_raw_data() processes one MCU row per call, which is to say |
2604 | v_samp_factor*DCTSIZE sample rows of each component. The passed num_lines | 2604 | v_samp_factor*DCTSIZE sample rows of each component. The passed num_lines |
2605 | value must be at least max_v_samp_factor*DCTSIZE, and the return value will | 2605 | value must be at least max_v_samp_factor*DCTSIZE, and the return value will |
2606 | be exactly that amount (or possibly some multiple of that amount, in future | 2606 | be exactly that amount (or possibly some multiple of that amount, in future |
2607 | library versions). This is true even on the last call at the bottom of the | 2607 | library versions). This is true even on the last call at the bottom of the |
2608 | image; don't forget to pad your data as necessary. | 2608 | image; don't forget to pad your data as necessary. |
2609 | 2609 | ||
2610 | The required dimensions of the supplied data can be computed for each | 2610 | The required dimensions of the supplied data can be computed for each |
2611 | component as | 2611 | component as |
2612 | cinfo->comp_info[i].width_in_blocks*DCTSIZE samples per row | 2612 | cinfo->comp_info[i].width_in_blocks*DCTSIZE samples per row |
2613 | cinfo->comp_info[i].height_in_blocks*DCTSIZE rows in image | 2613 | cinfo->comp_info[i].height_in_blocks*DCTSIZE rows in image |
2614 | after jpeg_start_compress() has initialized those fields. If the valid data | 2614 | after jpeg_start_compress() has initialized those fields. If the valid data |
2615 | is smaller than this, it must be padded appropriately. For some sampling | 2615 | is smaller than this, it must be padded appropriately. For some sampling |
2616 | factors and image sizes, additional dummy DCT blocks are inserted to make | 2616 | factors and image sizes, additional dummy DCT blocks are inserted to make |
2617 | the image a multiple of the MCU dimensions. The library creates such dummy | 2617 | the image a multiple of the MCU dimensions. The library creates such dummy |
2618 | blocks itself; it does not read them from your supplied data. Therefore you | 2618 | blocks itself; it does not read them from your supplied data. Therefore you |
2619 | need never pad by more than DCTSIZE samples. An example may help here. | 2619 | need never pad by more than DCTSIZE samples. An example may help here. |
2620 | Assume 2h2v downsampling of YCbCr data, that is | 2620 | Assume 2h2v downsampling of YCbCr data, that is |
2621 | cinfo->comp_info[0].h_samp_factor = 2 for Y | 2621 | cinfo->comp_info[0].h_samp_factor = 2 for Y |
2622 | cinfo->comp_info[0].v_samp_factor = 2 | 2622 | cinfo->comp_info[0].v_samp_factor = 2 |
2623 | cinfo->comp_info[1].h_samp_factor = 1 for Cb | 2623 | cinfo->comp_info[1].h_samp_factor = 1 for Cb |
2624 | cinfo->comp_info[1].v_samp_factor = 1 | 2624 | cinfo->comp_info[1].v_samp_factor = 1 |
2625 | cinfo->comp_info[2].h_samp_factor = 1 for Cr | 2625 | cinfo->comp_info[2].h_samp_factor = 1 for Cr |
2626 | cinfo->comp_info[2].v_samp_factor = 1 | 2626 | cinfo->comp_info[2].v_samp_factor = 1 |
2627 | and suppose that the nominal image dimensions (cinfo->image_width and | 2627 | and suppose that the nominal image dimensions (cinfo->image_width and |
2628 | cinfo->image_height) are 101x101 pixels. Then jpeg_start_compress() will | 2628 | cinfo->image_height) are 101x101 pixels. Then jpeg_start_compress() will |
2629 | compute downsampled_width = 101 and width_in_blocks = 13 for Y, | 2629 | compute downsampled_width = 101 and width_in_blocks = 13 for Y, |
2630 | downsampled_width = 51 and width_in_blocks = 7 for Cb and Cr (and the same | 2630 | downsampled_width = 51 and width_in_blocks = 7 for Cb and Cr (and the same |
2631 | for the height fields). You must pad the Y data to at least 13*8 = 104 | 2631 | for the height fields). You must pad the Y data to at least 13*8 = 104 |
2632 | columns and rows, the Cb/Cr data to at least 7*8 = 56 columns and rows. The | 2632 | columns and rows, the Cb/Cr data to at least 7*8 = 56 columns and rows. The |
2633 | MCU height is max_v_samp_factor = 2 DCT rows so you must pass at least 16 | 2633 | MCU height is max_v_samp_factor = 2 DCT rows so you must pass at least 16 |
2634 | scanlines on each call to jpeg_write_raw_data(), which is to say 16 actual | 2634 | scanlines on each call to jpeg_write_raw_data(), which is to say 16 actual |
2635 | sample rows of Y and 8 each of Cb and Cr. A total of 7 MCU rows are needed, | 2635 | sample rows of Y and 8 each of Cb and Cr. A total of 7 MCU rows are needed, |
2636 | so you must pass a total of 7*16 = 112 "scanlines". The last DCT block row | 2636 | so you must pass a total of 7*16 = 112 "scanlines". The last DCT block row |
2637 | of Y data is dummy, so it doesn't matter what you pass for it in the data | 2637 | of Y data is dummy, so it doesn't matter what you pass for it in the data |
2638 | arrays, but the scanlines count must total up to 112 so that all of the Cb | 2638 | arrays, but the scanlines count must total up to 112 so that all of the Cb |
2639 | and Cr data gets passed. | 2639 | and Cr data gets passed. |
2640 | 2640 | ||
2641 | Output suspension is supported with raw-data compression: if the data | 2641 | Output suspension is supported with raw-data compression: if the data |
2642 | destination module suspends, jpeg_write_raw_data() will return 0. | 2642 | destination module suspends, jpeg_write_raw_data() will return 0. |
2643 | In this case the same data rows must be passed again on the next call. | 2643 | In this case the same data rows must be passed again on the next call. |
2644 | 2644 | ||
2645 | 2645 | ||
2646 | Decompression with raw data output implies bypassing all postprocessing. | 2646 | Decompression with raw data output implies bypassing all postprocessing. |
2647 | You must deal with the color space and sampling factors present in the | 2647 | You must deal with the color space and sampling factors present in the |
2648 | incoming file. If your application only handles, say, 2h1v YCbCr data, | 2648 | incoming file. If your application only handles, say, 2h1v YCbCr data, |
2649 | you must check for and fail on other color spaces or other sampling factors. | 2649 | you must check for and fail on other color spaces or other sampling factors. |
2650 | The library will not convert to a different color space for you. | 2650 | The library will not convert to a different color space for you. |
2651 | 2651 | ||
2652 | To obtain raw data output, set cinfo->raw_data_out = TRUE before | 2652 | To obtain raw data output, set cinfo->raw_data_out = TRUE before |
2653 | jpeg_start_decompress() (it is set FALSE by jpeg_read_header()). Be sure to | 2653 | jpeg_start_decompress() (it is set FALSE by jpeg_read_header()). Be sure to |
2654 | verify that the color space and sampling factors are ones you can handle. | 2654 | verify that the color space and sampling factors are ones you can handle. |
2655 | Furthermore, set cinfo->do_fancy_upsampling = FALSE if you want to get real | 2655 | Furthermore, set cinfo->do_fancy_upsampling = FALSE if you want to get real |
2656 | downsampled data (it is set TRUE by jpeg_read_header()). | 2656 | downsampled data (it is set TRUE by jpeg_read_header()). |
2657 | Then call jpeg_read_raw_data() in place of jpeg_read_scanlines(). The | 2657 | Then call jpeg_read_raw_data() in place of jpeg_read_scanlines(). The |
2658 | decompression process is otherwise the same as usual. | 2658 | decompression process is otherwise the same as usual. |
2659 | 2659 | ||
2660 | jpeg_read_raw_data() returns one MCU row per call, and thus you must pass a | 2660 | jpeg_read_raw_data() returns one MCU row per call, and thus you must pass a |
2661 | buffer of at least max_v_samp_factor*DCTSIZE scanlines (scanline counting is | 2661 | buffer of at least max_v_samp_factor*DCTSIZE scanlines (scanline counting is |
2662 | the same as for raw-data compression). The buffer you pass must be large | 2662 | the same as for raw-data compression). The buffer you pass must be large |
2663 | enough to hold the actual data plus padding to DCT-block boundaries. As with | 2663 | enough to hold the actual data plus padding to DCT-block boundaries. As with |
2664 | compression, any entirely dummy DCT blocks are not processed so you need not | 2664 | compression, any entirely dummy DCT blocks are not processed so you need not |
2665 | allocate space for them, but the total scanline count includes them. The | 2665 | allocate space for them, but the total scanline count includes them. The |
2666 | above example of computing buffer dimensions for raw-data compression is | 2666 | above example of computing buffer dimensions for raw-data compression is |
2667 | equally valid for decompression. | 2667 | equally valid for decompression. |
2668 | 2668 | ||
2669 | Input suspension is supported with raw-data decompression: if the data source | 2669 | Input suspension is supported with raw-data decompression: if the data source |
2670 | module suspends, jpeg_read_raw_data() will return 0. You can also use | 2670 | module suspends, jpeg_read_raw_data() will return 0. You can also use |
2671 | buffered-image mode to read raw data in multiple passes. | 2671 | buffered-image mode to read raw data in multiple passes. |
2672 | 2672 | ||
2673 | 2673 | ||
2674 | Really raw data: DCT coefficients | 2674 | Really raw data: DCT coefficients |
2675 | --------------------------------- | 2675 | --------------------------------- |
2676 | 2676 | ||
2677 | It is possible to read or write the contents of a JPEG file as raw DCT | 2677 | It is possible to read or write the contents of a JPEG file as raw DCT |
2678 | coefficients. This facility is mainly intended for use in lossless | 2678 | coefficients. This facility is mainly intended for use in lossless |
2679 | transcoding between different JPEG file formats. Other possible applications | 2679 | transcoding between different JPEG file formats. Other possible applications |
2680 | include lossless cropping of a JPEG image, lossless reassembly of a | 2680 | include lossless cropping of a JPEG image, lossless reassembly of a |
2681 | multi-strip or multi-tile TIFF/JPEG file into a single JPEG datastream, etc. | 2681 | multi-strip or multi-tile TIFF/JPEG file into a single JPEG datastream, etc. |
2682 | 2682 | ||
2683 | To read the contents of a JPEG file as DCT coefficients, open the file and do | 2683 | To read the contents of a JPEG file as DCT coefficients, open the file and do |
2684 | jpeg_read_header() as usual. But instead of calling jpeg_start_decompress() | 2684 | jpeg_read_header() as usual. But instead of calling jpeg_start_decompress() |
2685 | and jpeg_read_scanlines(), call jpeg_read_coefficients(). This will read the | 2685 | and jpeg_read_scanlines(), call jpeg_read_coefficients(). This will read the |
2686 | entire image into a set of virtual coefficient-block arrays, one array per | 2686 | entire image into a set of virtual coefficient-block arrays, one array per |
2687 | component. The return value is a pointer to an array of virtual-array | 2687 | component. The return value is a pointer to an array of virtual-array |
2688 | descriptors. Each virtual array can be accessed directly using the JPEG | 2688 | descriptors. Each virtual array can be accessed directly using the JPEG |
2689 | memory manager's access_virt_barray method (see Memory management, below, | 2689 | memory manager's access_virt_barray method (see Memory management, below, |
2690 | and also read structure.txt's discussion of virtual array handling). Or, | 2690 | and also read structure.txt's discussion of virtual array handling). Or, |
2691 | for simple transcoding to a different JPEG file format, the array list can | 2691 | for simple transcoding to a different JPEG file format, the array list can |
2692 | just be handed directly to jpeg_write_coefficients(). | 2692 | just be handed directly to jpeg_write_coefficients(). |
2693 | 2693 | ||
2694 | Each block in the block arrays contains quantized coefficient values in | 2694 | Each block in the block arrays contains quantized coefficient values in |
2695 | normal array order (not JPEG zigzag order). The block arrays contain only | 2695 | normal array order (not JPEG zigzag order). The block arrays contain only |
2696 | DCT blocks containing real data; any entirely-dummy blocks added to fill out | 2696 | DCT blocks containing real data; any entirely-dummy blocks added to fill out |
2697 | interleaved MCUs at the right or bottom edges of the image are discarded | 2697 | interleaved MCUs at the right or bottom edges of the image are discarded |
2698 | during reading and are not stored in the block arrays. (The size of each | 2698 | during reading and are not stored in the block arrays. (The size of each |
2699 | block array can be determined from the width_in_blocks and height_in_blocks | 2699 | block array can be determined from the width_in_blocks and height_in_blocks |
2700 | fields of the component's comp_info entry.) This is also the data format | 2700 | fields of the component's comp_info entry.) This is also the data format |
2701 | expected by jpeg_write_coefficients(). | 2701 | expected by jpeg_write_coefficients(). |
2702 | 2702 | ||
2703 | When you are done using the virtual arrays, call jpeg_finish_decompress() | 2703 | When you are done using the virtual arrays, call jpeg_finish_decompress() |
2704 | to release the array storage and return the decompression object to an idle | 2704 | to release the array storage and return the decompression object to an idle |
2705 | state; or just call jpeg_destroy() if you don't need to reuse the object. | 2705 | state; or just call jpeg_destroy() if you don't need to reuse the object. |
2706 | 2706 | ||
2707 | If you use a suspending data source, jpeg_read_coefficients() will return | 2707 | If you use a suspending data source, jpeg_read_coefficients() will return |
2708 | NULL if it is forced to suspend; a non-NULL return value indicates successful | 2708 | NULL if it is forced to suspend; a non-NULL return value indicates successful |
2709 | completion. You need not test for a NULL return value when using a | 2709 | completion. You need not test for a NULL return value when using a |
2710 | non-suspending data source. | 2710 | non-suspending data source. |
2711 | 2711 | ||
2712 | It is also possible to call jpeg_read_coefficients() to obtain access to the | 2712 | It is also possible to call jpeg_read_coefficients() to obtain access to the |
2713 | decoder's coefficient arrays during a normal decode cycle in buffered-image | 2713 | decoder's coefficient arrays during a normal decode cycle in buffered-image |
2714 | mode. This frammish might be useful for progressively displaying an incoming | 2714 | mode. This frammish might be useful for progressively displaying an incoming |
2715 | image and then re-encoding it without loss. To do this, decode in buffered- | 2715 | image and then re-encoding it without loss. To do this, decode in buffered- |
2716 | image mode as discussed previously, then call jpeg_read_coefficients() after | 2716 | image mode as discussed previously, then call jpeg_read_coefficients() after |
2717 | the last jpeg_finish_output() call. The arrays will be available for your use | 2717 | the last jpeg_finish_output() call. The arrays will be available for your use |
2718 | until you call jpeg_finish_decompress(). | 2718 | until you call jpeg_finish_decompress(). |
2719 | 2719 | ||
2720 | 2720 | ||
2721 | To write the contents of a JPEG file as DCT coefficients, you must provide | 2721 | To write the contents of a JPEG file as DCT coefficients, you must provide |
2722 | the DCT coefficients stored in virtual block arrays. You can either pass | 2722 | the DCT coefficients stored in virtual block arrays. You can either pass |
2723 | block arrays read from an input JPEG file by jpeg_read_coefficients(), or | 2723 | block arrays read from an input JPEG file by jpeg_read_coefficients(), or |
2724 | allocate virtual arrays from the JPEG compression object and fill them | 2724 | allocate virtual arrays from the JPEG compression object and fill them |
2725 | yourself. In either case, jpeg_write_coefficients() is substituted for | 2725 | yourself. In either case, jpeg_write_coefficients() is substituted for |
2726 | jpeg_start_compress() and jpeg_write_scanlines(). Thus the sequence is | 2726 | jpeg_start_compress() and jpeg_write_scanlines(). Thus the sequence is |
2727 | * Create compression object | 2727 | * Create compression object |
2728 | * Set all compression parameters as necessary | 2728 | * Set all compression parameters as necessary |
2729 | * Request virtual arrays if needed | 2729 | * Request virtual arrays if needed |
2730 | * jpeg_write_coefficients() | 2730 | * jpeg_write_coefficients() |
2731 | * jpeg_finish_compress() | 2731 | * jpeg_finish_compress() |
2732 | * Destroy or re-use compression object | 2732 | * Destroy or re-use compression object |
2733 | jpeg_write_coefficients() is passed a pointer to an array of virtual block | 2733 | jpeg_write_coefficients() is passed a pointer to an array of virtual block |
2734 | array descriptors; the number of arrays is equal to cinfo.num_components. | 2734 | array descriptors; the number of arrays is equal to cinfo.num_components. |
2735 | 2735 | ||
2736 | The virtual arrays need only have been requested, not realized, before | 2736 | The virtual arrays need only have been requested, not realized, before |
2737 | jpeg_write_coefficients() is called. A side-effect of | 2737 | jpeg_write_coefficients() is called. A side-effect of |
2738 | jpeg_write_coefficients() is to realize any virtual arrays that have been | 2738 | jpeg_write_coefficients() is to realize any virtual arrays that have been |
2739 | requested from the compression object's memory manager. Thus, when obtaining | 2739 | requested from the compression object's memory manager. Thus, when obtaining |
2740 | the virtual arrays from the compression object, you should fill the arrays | 2740 | the virtual arrays from the compression object, you should fill the arrays |
2741 | after calling jpeg_write_coefficients(). The data is actually written out | 2741 | after calling jpeg_write_coefficients(). The data is actually written out |
2742 | when you call jpeg_finish_compress(); jpeg_write_coefficients() only writes | 2742 | when you call jpeg_finish_compress(); jpeg_write_coefficients() only writes |
2743 | the file header. | 2743 | the file header. |
2744 | 2744 | ||
2745 | When writing raw DCT coefficients, it is crucial that the JPEG quantization | 2745 | When writing raw DCT coefficients, it is crucial that the JPEG quantization |
2746 | tables and sampling factors match the way the data was encoded, or the | 2746 | tables and sampling factors match the way the data was encoded, or the |
2747 | resulting file will be invalid. For transcoding from an existing JPEG file, | 2747 | resulting file will be invalid. For transcoding from an existing JPEG file, |
2748 | we recommend using jpeg_copy_critical_parameters(). This routine initializes | 2748 | we recommend using jpeg_copy_critical_parameters(). This routine initializes |
2749 | all the compression parameters to default values (like jpeg_set_defaults()), | 2749 | all the compression parameters to default values (like jpeg_set_defaults()), |
2750 | then copies the critical information from a source decompression object. | 2750 | then copies the critical information from a source decompression object. |
2751 | The decompression object should have just been used to read the entire | 2751 | The decompression object should have just been used to read the entire |
2752 | JPEG input file --- that is, it should be awaiting jpeg_finish_decompress(). | 2752 | JPEG input file --- that is, it should be awaiting jpeg_finish_decompress(). |
2753 | 2753 | ||
2754 | jpeg_write_coefficients() marks all tables stored in the compression object | 2754 | jpeg_write_coefficients() marks all tables stored in the compression object |
2755 | as needing to be written to the output file (thus, it acts like | 2755 | as needing to be written to the output file (thus, it acts like |
2756 | jpeg_start_compress(cinfo, TRUE)). This is for safety's sake, to avoid | 2756 | jpeg_start_compress(cinfo, TRUE)). This is for safety's sake, to avoid |
2757 | emitting abbreviated JPEG files by accident. If you really want to emit an | 2757 | emitting abbreviated JPEG files by accident. If you really want to emit an |
2758 | abbreviated JPEG file, call jpeg_suppress_tables(), or set the tables' | 2758 | abbreviated JPEG file, call jpeg_suppress_tables(), or set the tables' |
2759 | individual sent_table flags, between calling jpeg_write_coefficients() and | 2759 | individual sent_table flags, between calling jpeg_write_coefficients() and |
2760 | jpeg_finish_compress(). | 2760 | jpeg_finish_compress(). |
2761 | 2761 | ||
2762 | 2762 | ||
2763 | Progress monitoring | 2763 | Progress monitoring |
2764 | ------------------- | 2764 | ------------------- |
2765 | 2765 | ||
2766 | Some applications may need to regain control from the JPEG library every so | 2766 | Some applications may need to regain control from the JPEG library every so |
2767 | often. The typical use of this feature is to produce a percent-done bar or | 2767 | often. The typical use of this feature is to produce a percent-done bar or |
2768 | other progress display. (For a simple example, see cjpeg.c or djpeg.c.) | 2768 | other progress display. (For a simple example, see cjpeg.c or djpeg.c.) |
2769 | Although you do get control back frequently during the data-transferring pass | 2769 | Although you do get control back frequently during the data-transferring pass |
2770 | (the jpeg_read_scanlines or jpeg_write_scanlines loop), any additional passes | 2770 | (the jpeg_read_scanlines or jpeg_write_scanlines loop), any additional passes |
2771 | will occur inside jpeg_finish_compress or jpeg_start_decompress; those | 2771 | will occur inside jpeg_finish_compress or jpeg_start_decompress; those |
2772 | routines may take a long time to execute, and you don't get control back | 2772 | routines may take a long time to execute, and you don't get control back |
2773 | until they are done. | 2773 | until they are done. |
2774 | 2774 | ||
2775 | You can define a progress-monitor routine which will be called periodically | 2775 | You can define a progress-monitor routine which will be called periodically |
2776 | by the library. No guarantees are made about how often this call will occur, | 2776 | by the library. No guarantees are made about how often this call will occur, |
2777 | so we don't recommend you use it for mouse tracking or anything like that. | 2777 | so we don't recommend you use it for mouse tracking or anything like that. |
2778 | At present, a call will occur once per MCU row, scanline, or sample row | 2778 | At present, a call will occur once per MCU row, scanline, or sample row |
2779 | group, whichever unit is convenient for the current processing mode; so the | 2779 | group, whichever unit is convenient for the current processing mode; so the |
2780 | wider the image, the longer the time between calls. During the data | 2780 | wider the image, the longer the time between calls. During the data |
2781 | transferring pass, only one call occurs per call of jpeg_read_scanlines or | 2781 | transferring pass, only one call occurs per call of jpeg_read_scanlines or |
2782 | jpeg_write_scanlines, so don't pass a large number of scanlines at once if | 2782 | jpeg_write_scanlines, so don't pass a large number of scanlines at once if |
2783 | you want fine resolution in the progress count. (If you really need to use | 2783 | you want fine resolution in the progress count. (If you really need to use |
2784 | the callback mechanism for time-critical tasks like mouse tracking, you could | 2784 | the callback mechanism for time-critical tasks like mouse tracking, you could |
2785 | insert additional calls inside some of the library's inner loops.) | 2785 | insert additional calls inside some of the library's inner loops.) |
2786 | 2786 | ||
2787 | To establish a progress-monitor callback, create a struct jpeg_progress_mgr, | 2787 | To establish a progress-monitor callback, create a struct jpeg_progress_mgr, |
2788 | fill in its progress_monitor field with a pointer to your callback routine, | 2788 | fill in its progress_monitor field with a pointer to your callback routine, |
2789 | and set cinfo->progress to point to the struct. The callback will be called | 2789 | and set cinfo->progress to point to the struct. The callback will be called |
2790 | whenever cinfo->progress is non-NULL. (This pointer is set to NULL by | 2790 | whenever cinfo->progress is non-NULL. (This pointer is set to NULL by |
2791 | jpeg_create_compress or jpeg_create_decompress; the library will not change | 2791 | jpeg_create_compress or jpeg_create_decompress; the library will not change |
2792 | it thereafter. So if you allocate dynamic storage for the progress struct, | 2792 | it thereafter. So if you allocate dynamic storage for the progress struct, |
2793 | make sure it will live as long as the JPEG object does. Allocating from the | 2793 | make sure it will live as long as the JPEG object does. Allocating from the |
2794 | JPEG memory manager with lifetime JPOOL_PERMANENT will work nicely.) You | 2794 | JPEG memory manager with lifetime JPOOL_PERMANENT will work nicely.) You |
2795 | can use the same callback routine for both compression and decompression. | 2795 | can use the same callback routine for both compression and decompression. |
2796 | 2796 | ||
2797 | The jpeg_progress_mgr struct contains four fields which are set by the library: | 2797 | The jpeg_progress_mgr struct contains four fields which are set by the library: |
2798 | long pass_counter; /* work units completed in this pass */ | 2798 | long pass_counter; /* work units completed in this pass */ |
2799 | long pass_limit; /* total number of work units in this pass */ | 2799 | long pass_limit; /* total number of work units in this pass */ |
2800 | int completed_passes; /* passes completed so far */ | 2800 | int completed_passes; /* passes completed so far */ |
2801 | int total_passes; /* total number of passes expected */ | 2801 | int total_passes; /* total number of passes expected */ |
2802 | During any one pass, pass_counter increases from 0 up to (not including) | 2802 | During any one pass, pass_counter increases from 0 up to (not including) |
2803 | pass_limit; the step size is usually but not necessarily 1. The pass_limit | 2803 | pass_limit; the step size is usually but not necessarily 1. The pass_limit |
2804 | value may change from one pass to another. The expected total number of | 2804 | value may change from one pass to another. The expected total number of |
2805 | passes is in total_passes, and the number of passes already completed is in | 2805 | passes is in total_passes, and the number of passes already completed is in |
2806 | completed_passes. Thus the fraction of work completed may be estimated as | 2806 | completed_passes. Thus the fraction of work completed may be estimated as |
2807 | completed_passes + (pass_counter/pass_limit) | 2807 | completed_passes + (pass_counter/pass_limit) |
2808 | -------------------------------------------- | 2808 | -------------------------------------------- |
2809 | total_passes | 2809 | total_passes |
2810 | ignoring the fact that the passes may not be equal amounts of work. | 2810 | ignoring the fact that the passes may not be equal amounts of work. |
2811 | 2811 | ||
2812 | When decompressing, pass_limit can even change within a pass, because it | 2812 | When decompressing, pass_limit can even change within a pass, because it |
2813 | depends on the number of scans in the JPEG file, which isn't always known in | 2813 | depends on the number of scans in the JPEG file, which isn't always known in |
2814 | advance. The computed fraction-of-work-done may jump suddenly (if the library | 2814 | advance. The computed fraction-of-work-done may jump suddenly (if the library |
2815 | discovers it has overestimated the number of scans) or even decrease (in the | 2815 | discovers it has overestimated the number of scans) or even decrease (in the |
2816 | opposite case). It is not wise to put great faith in the work estimate. | 2816 | opposite case). It is not wise to put great faith in the work estimate. |
2817 | 2817 | ||
2818 | When using the decompressor's buffered-image mode, the progress monitor work | 2818 | When using the decompressor's buffered-image mode, the progress monitor work |
2819 | estimate is likely to be completely unhelpful, because the library has no way | 2819 | estimate is likely to be completely unhelpful, because the library has no way |
2820 | to know how many output passes will be demanded of it. Currently, the library | 2820 | to know how many output passes will be demanded of it. Currently, the library |
2821 | sets total_passes based on the assumption that there will be one more output | 2821 | sets total_passes based on the assumption that there will be one more output |
2822 | pass if the input file end hasn't yet been read (jpeg_input_complete() isn't | 2822 | pass if the input file end hasn't yet been read (jpeg_input_complete() isn't |
2823 | TRUE), but no more output passes if the file end has been reached when the | 2823 | TRUE), but no more output passes if the file end has been reached when the |
2824 | output pass is started. This means that total_passes will rise as additional | 2824 | output pass is started. This means that total_passes will rise as additional |
2825 | output passes are requested. If you have a way of determining the input file | 2825 | output passes are requested. If you have a way of determining the input file |
2826 | size, estimating progress based on the fraction of the file that's been read | 2826 | size, estimating progress based on the fraction of the file that's been read |
2827 | will probably be more useful than using the library's value. | 2827 | will probably be more useful than using the library's value. |
2828 | 2828 | ||
2829 | 2829 | ||
2830 | Memory management | 2830 | Memory management |
2831 | ----------------- | 2831 | ----------------- |
2832 | 2832 | ||
2833 | This section covers some key facts about the JPEG library's built-in memory | 2833 | This section covers some key facts about the JPEG library's built-in memory |
2834 | manager. For more info, please read structure.txt's section about the memory | 2834 | manager. For more info, please read structure.txt's section about the memory |
2835 | manager, and consult the source code if necessary. | 2835 | manager, and consult the source code if necessary. |
2836 | 2836 | ||
2837 | All memory and temporary file allocation within the library is done via the | 2837 | All memory and temporary file allocation within the library is done via the |
2838 | memory manager. If necessary, you can replace the "back end" of the memory | 2838 | memory manager. If necessary, you can replace the "back end" of the memory |
2839 | manager to control allocation yourself (for example, if you don't want the | 2839 | manager to control allocation yourself (for example, if you don't want the |
2840 | library to use malloc() and free() for some reason). | 2840 | library to use malloc() and free() for some reason). |
2841 | 2841 | ||
2842 | Some data is allocated "permanently" and will not be freed until the JPEG | 2842 | Some data is allocated "permanently" and will not be freed until the JPEG |
2843 | object is destroyed. Most data is allocated "per image" and is freed by | 2843 | object is destroyed. Most data is allocated "per image" and is freed by |
2844 | jpeg_finish_compress, jpeg_finish_decompress, or jpeg_abort. You can call the | 2844 | jpeg_finish_compress, jpeg_finish_decompress, or jpeg_abort. You can call the |
2845 | memory manager yourself to allocate structures that will automatically be | 2845 | memory manager yourself to allocate structures that will automatically be |
2846 | freed at these times. Typical code for this is | 2846 | freed at these times. Typical code for this is |
2847 | ptr = (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, size); | 2847 | ptr = (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, size); |
2848 | Use JPOOL_PERMANENT to get storage that lasts as long as the JPEG object. | 2848 | Use JPOOL_PERMANENT to get storage that lasts as long as the JPEG object. |
2849 | Use alloc_large instead of alloc_small for anything bigger than a few Kbytes. | 2849 | Use alloc_large instead of alloc_small for anything bigger than a few Kbytes. |
2850 | There are also alloc_sarray and alloc_barray routines that automatically | 2850 | There are also alloc_sarray and alloc_barray routines that automatically |
2851 | build 2-D sample or block arrays. | 2851 | build 2-D sample or block arrays. |
2852 | 2852 | ||
2853 | The library's minimum space requirements to process an image depend on the | 2853 | The library's minimum space requirements to process an image depend on the |
2854 | image's width, but not on its height, because the library ordinarily works | 2854 | image's width, but not on its height, because the library ordinarily works |
2855 | with "strip" buffers that are as wide as the image but just a few rows high. | 2855 | with "strip" buffers that are as wide as the image but just a few rows high. |
2856 | Some operating modes (eg, two-pass color quantization) require full-image | 2856 | Some operating modes (eg, two-pass color quantization) require full-image |
2857 | buffers. Such buffers are treated as "virtual arrays": only the current strip | 2857 | buffers. Such buffers are treated as "virtual arrays": only the current strip |
2858 | need be in memory, and the rest can be swapped out to a temporary file. | 2858 | need be in memory, and the rest can be swapped out to a temporary file. |
2859 | 2859 | ||
2860 | If you use the simplest memory manager back end (jmemnobs.c), then no | 2860 | If you use the simplest memory manager back end (jmemnobs.c), then no |
2861 | temporary files are used; virtual arrays are simply malloc()'d. Images bigger | 2861 | temporary files are used; virtual arrays are simply malloc()'d. Images bigger |
2862 | than memory can be processed only if your system supports virtual memory. | 2862 | than memory can be processed only if your system supports virtual memory. |
2863 | The other memory manager back ends support temporary files of various flavors | 2863 | The other memory manager back ends support temporary files of various flavors |
2864 | and thus work in machines without virtual memory. They may also be useful on | 2864 | and thus work in machines without virtual memory. They may also be useful on |
2865 | Unix machines if you need to process images that exceed available swap space. | 2865 | Unix machines if you need to process images that exceed available swap space. |
2866 | 2866 | ||
2867 | When using temporary files, the library will make the in-memory buffers for | 2867 | When using temporary files, the library will make the in-memory buffers for |
2868 | its virtual arrays just big enough to stay within a "maximum memory" setting. | 2868 | its virtual arrays just big enough to stay within a "maximum memory" setting. |
2869 | Your application can set this limit by setting cinfo->mem->max_memory_to_use | 2869 | Your application can set this limit by setting cinfo->mem->max_memory_to_use |
2870 | after creating the JPEG object. (Of course, there is still a minimum size for | 2870 | after creating the JPEG object. (Of course, there is still a minimum size for |
2871 | the buffers, so the max-memory setting is effective only if it is bigger than | 2871 | the buffers, so the max-memory setting is effective only if it is bigger than |
2872 | the minimum space needed.) If you allocate any large structures yourself, you | 2872 | the minimum space needed.) If you allocate any large structures yourself, you |
2873 | must allocate them before jpeg_start_compress() or jpeg_start_decompress() in | 2873 | must allocate them before jpeg_start_compress() or jpeg_start_decompress() in |
2874 | order to have them counted against the max memory limit. Also keep in mind | 2874 | order to have them counted against the max memory limit. Also keep in mind |
2875 | that space allocated with alloc_small() is ignored, on the assumption that | 2875 | that space allocated with alloc_small() is ignored, on the assumption that |
2876 | it's too small to be worth worrying about; so a reasonable safety margin | 2876 | it's too small to be worth worrying about; so a reasonable safety margin |
2877 | should be left when setting max_memory_to_use. | 2877 | should be left when setting max_memory_to_use. |
2878 | 2878 | ||
2879 | If you use the jmemname.c or jmemdos.c memory manager back end, it is | 2879 | If you use the jmemname.c or jmemdos.c memory manager back end, it is |
2880 | important to clean up the JPEG object properly to ensure that the temporary | 2880 | important to clean up the JPEG object properly to ensure that the temporary |
2881 | files get deleted. (This is especially crucial with jmemdos.c, where the | 2881 | files get deleted. (This is especially crucial with jmemdos.c, where the |
2882 | "temporary files" may be extended-memory segments; if they are not freed, | 2882 | "temporary files" may be extended-memory segments; if they are not freed, |
2883 | DOS will require a reboot to recover the memory.) Thus, with these memory | 2883 | DOS will require a reboot to recover the memory.) Thus, with these memory |
2884 | managers, it's a good idea to provide a signal handler that will trap any | 2884 | managers, it's a good idea to provide a signal handler that will trap any |
2885 | early exit from your program. The handler should call either jpeg_abort() | 2885 | early exit from your program. The handler should call either jpeg_abort() |
2886 | or jpeg_destroy() for any active JPEG objects. A handler is not needed with | 2886 | or jpeg_destroy() for any active JPEG objects. A handler is not needed with |
2887 | jmemnobs.c, and shouldn't be necessary with jmemansi.c or jmemmac.c either, | 2887 | jmemnobs.c, and shouldn't be necessary with jmemansi.c or jmemmac.c either, |
2888 | since the C library is supposed to take care of deleting files made with | 2888 | since the C library is supposed to take care of deleting files made with |
2889 | tmpfile(). | 2889 | tmpfile(). |
2890 | 2890 | ||
2891 | 2891 | ||
2892 | Memory usage | 2892 | Memory usage |
2893 | ------------ | 2893 | ------------ |
2894 | 2894 | ||
2895 | Working memory requirements while performing compression or decompression | 2895 | Working memory requirements while performing compression or decompression |
2896 | depend on image dimensions, image characteristics (such as colorspace and | 2896 | depend on image dimensions, image characteristics (such as colorspace and |
2897 | JPEG process), and operating mode (application-selected options). | 2897 | JPEG process), and operating mode (application-selected options). |
2898 | 2898 | ||
2899 | As of v6b, the decompressor requires: | 2899 | As of v6b, the decompressor requires: |
2900 | 1. About 24K in more-or-less-fixed-size data. This varies a bit depending | 2900 | 1. About 24K in more-or-less-fixed-size data. This varies a bit depending |
2901 | on operating mode and image characteristics (particularly color vs. | 2901 | on operating mode and image characteristics (particularly color vs. |
2902 | grayscale), but it doesn't depend on image dimensions. | 2902 | grayscale), but it doesn't depend on image dimensions. |
2903 | 2. Strip buffers (of size proportional to the image width) for IDCT and | 2903 | 2. Strip buffers (of size proportional to the image width) for IDCT and |
2904 | upsampling results. The worst case for commonly used sampling factors | 2904 | upsampling results. The worst case for commonly used sampling factors |
2905 | is about 34 bytes * width in pixels for a color image. A grayscale image | 2905 | is about 34 bytes * width in pixels for a color image. A grayscale image |
2906 | only needs about 8 bytes per pixel column. | 2906 | only needs about 8 bytes per pixel column. |
2907 | 3. A full-image DCT coefficient buffer is needed to decode a multi-scan JPEG | 2907 | 3. A full-image DCT coefficient buffer is needed to decode a multi-scan JPEG |
2908 | file (including progressive JPEGs), or whenever you select buffered-image | 2908 | file (including progressive JPEGs), or whenever you select buffered-image |
2909 | mode. This takes 2 bytes/coefficient. At typical 2x2 sampling, that's | 2909 | mode. This takes 2 bytes/coefficient. At typical 2x2 sampling, that's |
2910 | 3 bytes per pixel for a color image. Worst case (1x1 sampling) requires | 2910 | 3 bytes per pixel for a color image. Worst case (1x1 sampling) requires |
2911 | 6 bytes/pixel. For grayscale, figure 2 bytes/pixel. | 2911 | 6 bytes/pixel. For grayscale, figure 2 bytes/pixel. |
2912 | 4. To perform 2-pass color quantization, the decompressor also needs a | 2912 | 4. To perform 2-pass color quantization, the decompressor also needs a |
2913 | 128K color lookup table and a full-image pixel buffer (3 bytes/pixel). | 2913 | 128K color lookup table and a full-image pixel buffer (3 bytes/pixel). |
2914 | This does not count any memory allocated by the application, such as a | 2914 | This does not count any memory allocated by the application, such as a |
2915 | buffer to hold the final output image. | 2915 | buffer to hold the final output image. |
2916 | 2916 | ||
2917 | The above figures are valid for 8-bit JPEG data precision and a machine with | 2917 | The above figures are valid for 8-bit JPEG data precision and a machine with |
2918 | 32-bit ints. For 12-bit JPEG data, double the size of the strip buffers and | 2918 | 32-bit ints. For 12-bit JPEG data, double the size of the strip buffers and |
2919 | quantization pixel buffer. The "fixed-size" data will be somewhat smaller | 2919 | quantization pixel buffer. The "fixed-size" data will be somewhat smaller |
2920 | with 16-bit ints, larger with 64-bit ints. Also, CMYK or other unusual | 2920 | with 16-bit ints, larger with 64-bit ints. Also, CMYK or other unusual |
2921 | color spaces will require different amounts of space. | 2921 | color spaces will require different amounts of space. |
2922 | 2922 | ||
2923 | The full-image coefficient and pixel buffers, if needed at all, do not | 2923 | The full-image coefficient and pixel buffers, if needed at all, do not |
2924 | have to be fully RAM resident; you can have the library use temporary | 2924 | have to be fully RAM resident; you can have the library use temporary |
2925 | files instead when the total memory usage would exceed a limit you set. | 2925 | files instead when the total memory usage would exceed a limit you set. |
2926 | (But if your OS supports virtual memory, it's probably better to just use | 2926 | (But if your OS supports virtual memory, it's probably better to just use |
2927 | jmemnobs and let the OS do the swapping.) | 2927 | jmemnobs and let the OS do the swapping.) |
2928 | 2928 | ||
2929 | The compressor's memory requirements are similar, except that it has no need | 2929 | The compressor's memory requirements are similar, except that it has no need |
2930 | for color quantization. Also, it needs a full-image DCT coefficient buffer | 2930 | for color quantization. Also, it needs a full-image DCT coefficient buffer |
2931 | if Huffman-table optimization is asked for, even if progressive mode is not | 2931 | if Huffman-table optimization is asked for, even if progressive mode is not |
2932 | requested. | 2932 | requested. |
2933 | 2933 | ||
2934 | If you need more detailed information about memory usage in a particular | 2934 | If you need more detailed information about memory usage in a particular |
2935 | situation, you can enable the MEM_STATS code in jmemmgr.c. | 2935 | situation, you can enable the MEM_STATS code in jmemmgr.c. |
2936 | 2936 | ||
2937 | 2937 | ||
2938 | Library compile-time options | 2938 | Library compile-time options |
2939 | ---------------------------- | 2939 | ---------------------------- |
2940 | 2940 | ||
2941 | A number of compile-time options are available by modifying jmorecfg.h. | 2941 | A number of compile-time options are available by modifying jmorecfg.h. |
2942 | 2942 | ||
2943 | The JPEG standard provides for both the baseline 8-bit DCT process and | 2943 | The JPEG standard provides for both the baseline 8-bit DCT process and |
2944 | a 12-bit DCT process. The IJG code supports 12-bit JPEG if you define | 2944 | a 12-bit DCT process. The IJG code supports 12-bit JPEG if you define |
2945 | BITS_IN_JSAMPLE as 12 rather than 8. Note that this causes JSAMPLE to be | 2945 | BITS_IN_JSAMPLE as 12 rather than 8. Note that this causes JSAMPLE to be |
2946 | larger than a char, so it affects the surrounding application's image data. | 2946 | larger than a char, so it affects the surrounding application's image data. |
2947 | The sample applications cjpeg and djpeg can support 12-bit mode only for PPM | 2947 | The sample applications cjpeg and djpeg can support 12-bit mode only for PPM |
2948 | and GIF file formats; you must disable the other file formats to compile a | 2948 | and GIF file formats; you must disable the other file formats to compile a |
2949 | 12-bit cjpeg or djpeg. (install.txt has more information about that.) | 2949 | 12-bit cjpeg or djpeg. (install.txt has more information about that.) |
2950 | At present, a 12-bit library can handle *only* 12-bit images, not both | 2950 | At present, a 12-bit library can handle *only* 12-bit images, not both |
2951 | precisions. (If you need to include both 8- and 12-bit libraries in a single | 2951 | precisions. (If you need to include both 8- and 12-bit libraries in a single |
2952 | application, you could probably do it by defining NEED_SHORT_EXTERNAL_NAMES | 2952 | application, you could probably do it by defining NEED_SHORT_EXTERNAL_NAMES |
2953 | for just one of the copies. You'd have to access the 8-bit and 12-bit copies | 2953 | for just one of the copies. You'd have to access the 8-bit and 12-bit copies |
2954 | from separate application source files. This is untested ... if you try it, | 2954 | from separate application source files. This is untested ... if you try it, |
2955 | we'd like to hear whether it works!) | 2955 | we'd like to hear whether it works!) |
2956 | 2956 | ||
2957 | Note that a 12-bit library always compresses in Huffman optimization mode, | 2957 | Note that a 12-bit library always compresses in Huffman optimization mode, |
2958 | in order to generate valid Huffman tables. This is necessary because our | 2958 | in order to generate valid Huffman tables. This is necessary because our |
2959 | default Huffman tables only cover 8-bit data. If you need to output 12-bit | 2959 | default Huffman tables only cover 8-bit data. If you need to output 12-bit |
2960 | files in one pass, you'll have to supply suitable default Huffman tables. | 2960 | files in one pass, you'll have to supply suitable default Huffman tables. |
2961 | You may also want to supply your own DCT quantization tables; the existing | 2961 | You may also want to supply your own DCT quantization tables; the existing |
2962 | quality-scaling code has been developed for 8-bit use, and probably doesn't | 2962 | quality-scaling code has been developed for 8-bit use, and probably doesn't |
2963 | generate especially good tables for 12-bit. | 2963 | generate especially good tables for 12-bit. |
2964 | 2964 | ||
2965 | The maximum number of components (color channels) in the image is determined | 2965 | The maximum number of components (color channels) in the image is determined |
2966 | by MAX_COMPONENTS. The JPEG standard allows up to 255 components, but we | 2966 | by MAX_COMPONENTS. The JPEG standard allows up to 255 components, but we |
2967 | expect that few applications will need more than four or so. | 2967 | expect that few applications will need more than four or so. |
2968 | 2968 | ||
2969 | On machines with unusual data type sizes, you may be able to improve | 2969 | On machines with unusual data type sizes, you may be able to improve |
2970 | performance or reduce memory space by tweaking the various typedefs in | 2970 | performance or reduce memory space by tweaking the various typedefs in |
2971 | jmorecfg.h. In particular, on some RISC CPUs, access to arrays of "short"s | 2971 | jmorecfg.h. In particular, on some RISC CPUs, access to arrays of "short"s |
2972 | is quite slow; consider trading memory for speed by making JCOEF, INT16, and | 2972 | is quite slow; consider trading memory for speed by making JCOEF, INT16, and |
2973 | UINT16 be "int" or "unsigned int". UINT8 is also a candidate to become int. | 2973 | UINT16 be "int" or "unsigned int". UINT8 is also a candidate to become int. |
2974 | You probably don't want to make JSAMPLE be int unless you have lots of memory | 2974 | You probably don't want to make JSAMPLE be int unless you have lots of memory |
2975 | to burn. | 2975 | to burn. |
2976 | 2976 | ||
2977 | You can reduce the size of the library by compiling out various optional | 2977 | You can reduce the size of the library by compiling out various optional |
2978 | functions. To do this, undefine xxx_SUPPORTED symbols as necessary. | 2978 | functions. To do this, undefine xxx_SUPPORTED symbols as necessary. |
2979 | 2979 | ||
2980 | You can also save a few K by not having text error messages in the library; | 2980 | You can also save a few K by not having text error messages in the library; |
2981 | the standard error message table occupies about 5Kb. This is particularly | 2981 | the standard error message table occupies about 5Kb. This is particularly |
2982 | reasonable for embedded applications where there's no good way to display | 2982 | reasonable for embedded applications where there's no good way to display |
2983 | a message anyway. To do this, remove the creation of the message table | 2983 | a message anyway. To do this, remove the creation of the message table |
2984 | (jpeg_std_message_table[]) from jerror.c, and alter format_message to do | 2984 | (jpeg_std_message_table[]) from jerror.c, and alter format_message to do |
2985 | something reasonable without it. You could output the numeric value of the | 2985 | something reasonable without it. You could output the numeric value of the |
2986 | message code number, for example. If you do this, you can also save a couple | 2986 | message code number, for example. If you do this, you can also save a couple |
2987 | more K by modifying the TRACEMSn() macros in jerror.h to expand to nothing; | 2987 | more K by modifying the TRACEMSn() macros in jerror.h to expand to nothing; |
2988 | you don't need trace capability anyway, right? | 2988 | you don't need trace capability anyway, right? |
2989 | 2989 | ||
2990 | 2990 | ||
2991 | Portability considerations | 2991 | Portability considerations |
2992 | -------------------------- | 2992 | -------------------------- |
2993 | 2993 | ||
2994 | The JPEG library has been written to be extremely portable; the sample | 2994 | The JPEG library has been written to be extremely portable; the sample |
2995 | applications cjpeg and djpeg are slightly less so. This section summarizes | 2995 | applications cjpeg and djpeg are slightly less so. This section summarizes |
2996 | the design goals in this area. (If you encounter any bugs that cause the | 2996 | the design goals in this area. (If you encounter any bugs that cause the |
2997 | library to be less portable than is claimed here, we'd appreciate hearing | 2997 | library to be less portable than is claimed here, we'd appreciate hearing |
2998 | about them.) | 2998 | about them.) |
2999 | 2999 | ||
3000 | The code works fine on ANSI C, C++, and pre-ANSI C compilers, using any of | 3000 | The code works fine on ANSI C, C++, and pre-ANSI C compilers, using any of |
3001 | the popular system include file setups, and some not-so-popular ones too. | 3001 | the popular system include file setups, and some not-so-popular ones too. |
3002 | See install.txt for configuration procedures. | 3002 | See install.txt for configuration procedures. |
3003 | 3003 | ||
3004 | The code is not dependent on the exact sizes of the C data types. As | 3004 | The code is not dependent on the exact sizes of the C data types. As |
3005 | distributed, we make the assumptions that | 3005 | distributed, we make the assumptions that |
3006 | char is at least 8 bits wide | 3006 | char is at least 8 bits wide |
3007 | short is at least 16 bits wide | 3007 | short is at least 16 bits wide |
3008 | int is at least 16 bits wide | 3008 | int is at least 16 bits wide |
3009 | long is at least 32 bits wide | 3009 | long is at least 32 bits wide |
3010 | (These are the minimum requirements of the ANSI C standard.) Wider types will | 3010 | (These are the minimum requirements of the ANSI C standard.) Wider types will |
3011 | work fine, although memory may be used inefficiently if char is much larger | 3011 | work fine, although memory may be used inefficiently if char is much larger |
3012 | than 8 bits or short is much bigger than 16 bits. The code should work | 3012 | than 8 bits or short is much bigger than 16 bits. The code should work |
3013 | equally well with 16- or 32-bit ints. | 3013 | equally well with 16- or 32-bit ints. |
3014 | 3014 | ||
3015 | In a system where these assumptions are not met, you may be able to make the | 3015 | In a system where these assumptions are not met, you may be able to make the |
3016 | code work by modifying the typedefs in jmorecfg.h. However, you will probably | 3016 | code work by modifying the typedefs in jmorecfg.h. However, you will probably |
3017 | have difficulty if int is less than 16 bits wide, since references to plain | 3017 | have difficulty if int is less than 16 bits wide, since references to plain |
3018 | int abound in the code. | 3018 | int abound in the code. |
3019 | 3019 | ||
3020 | char can be either signed or unsigned, although the code runs faster if an | 3020 | char can be either signed or unsigned, although the code runs faster if an |
3021 | unsigned char type is available. If char is wider than 8 bits, you will need | 3021 | unsigned char type is available. If char is wider than 8 bits, you will need |
3022 | to redefine JOCTET and/or provide custom data source/destination managers so | 3022 | to redefine JOCTET and/or provide custom data source/destination managers so |
3023 | that JOCTET represents exactly 8 bits of data on external storage. | 3023 | that JOCTET represents exactly 8 bits of data on external storage. |
3024 | 3024 | ||
3025 | The JPEG library proper does not assume ASCII representation of characters. | 3025 | The JPEG library proper does not assume ASCII representation of characters. |
3026 | But some of the image file I/O modules in cjpeg/djpeg do have ASCII | 3026 | But some of the image file I/O modules in cjpeg/djpeg do have ASCII |
3027 | dependencies in file-header manipulation; so does cjpeg's select_file_type() | 3027 | dependencies in file-header manipulation; so does cjpeg's select_file_type() |
3028 | routine. | 3028 | routine. |
3029 | 3029 | ||
3030 | The JPEG library does not rely heavily on the C library. In particular, C | 3030 | The JPEG library does not rely heavily on the C library. In particular, C |
3031 | stdio is used only by the data source/destination modules and the error | 3031 | stdio is used only by the data source/destination modules and the error |
3032 | handler, all of which are application-replaceable. (cjpeg/djpeg are more | 3032 | handler, all of which are application-replaceable. (cjpeg/djpeg are more |
3033 | heavily dependent on stdio.) malloc and free are called only from the memory | 3033 | heavily dependent on stdio.) malloc and free are called only from the memory |
3034 | manager "back end" module, so you can use a different memory allocator by | 3034 | manager "back end" module, so you can use a different memory allocator by |
3035 | replacing that one file. | 3035 | replacing that one file. |
3036 | 3036 | ||
3037 | The code generally assumes that C names must be unique in the first 15 | 3037 | The code generally assumes that C names must be unique in the first 15 |
3038 | characters. However, global function names can be made unique in the | 3038 | characters. However, global function names can be made unique in the |
3039 | first 6 characters by defining NEED_SHORT_EXTERNAL_NAMES. | 3039 | first 6 characters by defining NEED_SHORT_EXTERNAL_NAMES. |
3040 | 3040 | ||
3041 | More info about porting the code may be gleaned by reading jconfig.txt, | 3041 | More info about porting the code may be gleaned by reading jconfig.txt, |
3042 | jmorecfg.h, and jinclude.h. | 3042 | jmorecfg.h, and jinclude.h. |
3043 | 3043 | ||
3044 | 3044 | ||
3045 | Notes for MS-DOS implementors | 3045 | Notes for MS-DOS implementors |
3046 | ----------------------------- | 3046 | ----------------------------- |
3047 | 3047 | ||
3048 | The IJG code is designed to work efficiently in 80x86 "small" or "medium" | 3048 | The IJG code is designed to work efficiently in 80x86 "small" or "medium" |
3049 | memory models (i.e., data pointers are 16 bits unless explicitly declared | 3049 | memory models (i.e., data pointers are 16 bits unless explicitly declared |
3050 | "far"; code pointers can be either size). You may be able to use small | 3050 | "far"; code pointers can be either size). You may be able to use small |
3051 | model to compile cjpeg or djpeg by itself, but you will probably have to use | 3051 | model to compile cjpeg or djpeg by itself, but you will probably have to use |
3052 | medium model for any larger application. This won't make much difference in | 3052 | medium model for any larger application. This won't make much difference in |
3053 | performance. You *will* take a noticeable performance hit if you use a | 3053 | performance. You *will* take a noticeable performance hit if you use a |
3054 | large-data memory model (perhaps 10%-25%), and you should avoid "huge" model | 3054 | large-data memory model (perhaps 10%-25%), and you should avoid "huge" model |
3055 | if at all possible. | 3055 | if at all possible. |
3056 | 3056 | ||
3057 | The JPEG library typically needs 2Kb-3Kb of stack space. It will also | 3057 | The JPEG library typically needs 2Kb-3Kb of stack space. It will also |
3058 | malloc about 20K-30K of near heap space while executing (and lots of far | 3058 | malloc about 20K-30K of near heap space while executing (and lots of far |
3059 | heap, but that doesn't count in this calculation). This figure will vary | 3059 | heap, but that doesn't count in this calculation). This figure will vary |
3060 | depending on selected operating mode, and to a lesser extent on image size. | 3060 | depending on selected operating mode, and to a lesser extent on image size. |
3061 | There is also about 5Kb-6Kb of constant data which will be allocated in the | 3061 | There is also about 5Kb-6Kb of constant data which will be allocated in the |
3062 | near data segment (about 4Kb of this is the error message table). | 3062 | near data segment (about 4Kb of this is the error message table). |
3063 | Thus you have perhaps 20K available for other modules' static data and near | 3063 | Thus you have perhaps 20K available for other modules' static data and near |
3064 | heap space before you need to go to a larger memory model. The C library's | 3064 | heap space before you need to go to a larger memory model. The C library's |
3065 | static data will account for several K of this, but that still leaves a good | 3065 | static data will account for several K of this, but that still leaves a good |
3066 | deal for your needs. (If you are tight on space, you could reduce the sizes | 3066 | deal for your needs. (If you are tight on space, you could reduce the sizes |
3067 | of the I/O buffers allocated by jdatasrc.c and jdatadst.c, say from 4K to | 3067 | of the I/O buffers allocated by jdatasrc.c and jdatadst.c, say from 4K to |
3068 | 1K. Another possibility is to move the error message table to far memory; | 3068 | 1K. Another possibility is to move the error message table to far memory; |
3069 | this should be doable with only localized hacking on jerror.c.) | 3069 | this should be doable with only localized hacking on jerror.c.) |
3070 | 3070 | ||
3071 | About 2K of the near heap space is "permanent" memory that will not be | 3071 | About 2K of the near heap space is "permanent" memory that will not be |
3072 | released until you destroy the JPEG object. This is only an issue if you | 3072 | released until you destroy the JPEG object. This is only an issue if you |
3073 | save a JPEG object between compression or decompression operations. | 3073 | save a JPEG object between compression or decompression operations. |
3074 | 3074 | ||
3075 | Far data space may also be a tight resource when you are dealing with large | 3075 | Far data space may also be a tight resource when you are dealing with large |
3076 | images. The most memory-intensive case is decompression with two-pass color | 3076 | images. The most memory-intensive case is decompression with two-pass color |
3077 | quantization, or single-pass quantization to an externally supplied color | 3077 | quantization, or single-pass quantization to an externally supplied color |
3078 | map. This requires a 128Kb color lookup table plus strip buffers amounting | 3078 | map. This requires a 128Kb color lookup table plus strip buffers amounting |
3079 | to about 40 bytes per column for typical sampling ratios (eg, about 25600 | 3079 | to about 40 bytes per column for typical sampling ratios (eg, about 25600 |
3080 | bytes for a 640-pixel-wide image). You may not be able to process wide | 3080 | bytes for a 640-pixel-wide image). You may not be able to process wide |
3081 | images if you have large data structures of your own. | 3081 | images if you have large data structures of your own. |
3082 | 3082 | ||
3083 | Of course, all of these concerns vanish if you use a 32-bit flat-memory-model | 3083 | Of course, all of these concerns vanish if you use a 32-bit flat-memory-model |
3084 | compiler, such as DJGPP or Watcom C. We highly recommend flat model if you | 3084 | compiler, such as DJGPP or Watcom C. We highly recommend flat model if you |
3085 | can use it; the JPEG library is significantly faster in flat model. | 3085 | can use it; the JPEG library is significantly faster in flat model. |