1 2 3 4 5<!DOCTYPE html> 6<html lang="en"> 7<head> 8 <title>ImageMagick: Architecture</title> 9 <meta charset="utf-8" /> 10 <meta http-equiv="X-UA-Compatible" content="IE=edge" /> 11 <meta name="viewport" content="width=device-width, initial-scale=1" /> 12 <meta http-equiv="content-type" content="text/html; charset=utf-8"/> 13 <meta name="application-name" content="ImageMagick"/> 14 <meta name="description" content="ImageMagick® is a software suite to create, edit, compose, or convert bitmap images. It can read and write images in a variety of formats (over 200) including PNG, JPEG, JPEG-2000, GIF, WebP, Postscript, PDF, and SVG. Use ImageMagick to resize, flip, mirror, rotate, distort, shear and transform images, adjust image colors, apply various special effects, or draw text, lines, polygons, ellipses and Bézier curves."/> 15 <meta name="application-url" content="http://www.imagemagick.org"/> 16 <meta name="generator" content="PHP"/> 17 <meta name="keywords" content="architecture, ImageMagick, PerlMagick, image processing, image, photo, software, Magick++, OpenMP, convert"/> 18 <meta name="rating" content="GENERAL"/> 19 <meta name="robots" content="INDEX, FOLLOW"/> 20 <meta name="generator" content="ImageMagick Studio LLC"/> 21 <meta name="author" content="ImageMagick Studio LLC"/> 22 <meta name="revisit-after" content="2 DAYS"/> 23 <meta name="resource-type" content="document"/> 24 <meta name="copyright" content="Copyright (c) 1999-2016 ImageMagick Studio LLC"/> 25 <meta name="distribution" content="Global"/> 26 <meta name="magick-serial" content="P131-S030410-R485315270133-P82224-A6668-G1245-1"/> 27 <meta name="google-site-verification" content="_bMOCDpkx9ZAzBwb2kF3PRHbfUUdFj2uO8Jd1AXArz4" /> 28 <link rel="icon" href="../images/wand.png"/> 29 <link rel="shortcut icon" href="../images/wand.ico"/> 30 <link rel="stylesheet" type="text/css" href="http://fonts.googleapis.com/css?family=Roboto:900,400,400italic,700,700italic,300,300italic|Open+Sans:300italic,400italic,700italic,300,400,600,700"> 31 <link rel="stylesheet" href="css/magick.css"/> 32</head> 33<body> 34<div class="main"> 35<div class="magick-masthead"> 36 <div class="container"> 37 <script async="async" src="http://localhost/pagead/js/adsbygoogle.js"></script> <ins class="adsbygoogle" 38 style="display:block" 39 data-ad-client="ca-pub-3129977114552745" 40 data-ad-slot="6345125851" 41 data-ad-format="auto"></ins> 42 <script> 43 (adsbygoogle = window.adsbygoogle || []).push({}); 44 </script> 45 <nav class="magick-nav"> 46 <a class="magick-nav-item " href="../index.html">Home</a> 47 <a class="magick-nav-item " href="binary-releases.html">Download</a> 48 <a class="magick-nav-item " href="command-line-tools.html">Tools</a> 49 <a class="magick-nav-item " href="command-line-processing.html">Command-line</a> 50 <a class="magick-nav-item " href="resources.html">Resources</a> 51 <a class="magick-nav-item " href="api.html">Develop</a> 52 <a class="magick-nav-item " href="http://www.imagemagick.org/script/search.php">Search</a> 53 <a class="magick-nav-item pull-right" href="https://www.imagemagick.org/discourse-server/">Community</a> 54 </nav> 55 </div> 56</div> 57<div class="container"> 58<div class="magick-header"> 59<p class="text-center"><a href="architecture.html#cache">The Pixel Cache</a> • <a href="architecture.html#stream">Streaming Pixels</a> • <a href="architecture.html#properties">Image Properties and Profiles</a> • <a href="architecture.html#tera-pixel">Large Image Support</a> • <a href="architecture.html#threads">Threads of Execution</a> • <a href="architecture.html#distributed">Heterogeneous Distributed Processing</a> • <a href="architecture.html#coders">Custom Image Coders</a> • <a href="architecture.html#filters">Custom Image Filters</a></p> 60 61<p class="lead magick-description">The citizens of Oz were quite content with their benefactor, the all-powerful Wizard. They accepted his wisdom and benevolence without ever questioning the who, why, and where of his power. Like the citizens of Oz, if you feel comfortable that ImageMagick can help you convert, edit, or compose your images without knowing what goes on behind the curtain, feel free to skip this section. However, if you want to know more about the software and algorithms behind ImageMagick, read on. To fully benefit from this discussion, you should be comfortable with image nomenclature and be familiar with computer programming.</p> 62 63<h2 class="magick-header"><a id="overview"></a>Architecture Overview</h2> 64 65<p>An image typically consists of a rectangular region of pixels and metadata. To convert, edit, or compose an image in an efficient manner, we need convenient access to any pixel anywhere within the region (and sometimes outside the region). And in the case of an image sequence, we need access to any pixel of any region of any image in the sequence. However, there are hundreds of image formats such JPEG, TIFF, PNG, GIF, etc., that makes it difficult to access pixels on demand. Within these formats we find differences in:</p> 66 67<ul> 68 <li>colorspace (e.g sRGB, linear RGB, linear GRAY, CMYK, YUV, Lab, etc.)</li> 69 <li>bit depth (.e.g 1, 4, 8, 12, 16, etc.)</li> 70 <li>storage format (e.g. unsigned, signed, float, double, etc.)</li> 71 <li>compression (e.g. uncompressed, RLE, Zip, BZip, etc.)</li> 72 <li>orientation (i.e. top-to-bottom, right-to-left, etc.),</li> 73 <li>layout (.e.g. raw, interspersed with opcodes, etc.)</li> 74</ul> 75 76<p>In addition, some image pixels may require attenuation, some formats permit more than one frame, and some formats contain vector graphics that must first be rasterized (converted from vector to pixels).</p> 77 78<p>An efficient implementation of an image processing algorithm may require we get or set:</p> 79 80<ul> 81 <li>one pixel a time (e.g. pixel at location 10,3)</li> 82 <li>a single scanline (e.g. all pixels from row 4)</li> 83 <li>a few scanlines at once (e.g. pixel rows 4-7)</li> 84 <li>a single column or columns of pixels (e.g. all pixels from column 11)</li> 85 <li>an arbitrary region of pixels from the image (e.g. pixels defined at 10,7 to 10,19)</li> 86 <li>a pixel in random order (e.g. pixel at 14,15 and 640,480)</li> 87 <li>pixels from two different images (e.g. pixel at 5,1 from image 1 and pixel at 5,1 from image 2)</li> 88 <li>pixels outside the boundaries of the image (e.g. pixel at -1,-3)</li> 89 <li>a pixel component that is unsigned (65311) or in a floating-point representation (e.g. 0.17836)</li> 90 <li>a high-dynamic range pixel that can include negative values (e.g. -0.00716) as well as values that exceed the quantum depth (e.g. 65931)</li> 91 <li>one or more pixels simultaneously in different threads of execution</li> 92 <li>all the pixels in memory to take advantage of speed-ups offered by executing in concert across heterogeneous platforms consisting of CPUs, GPUs, and other processors</li> 93</ul> 94 95<p>Some images include a clip mask that define which pixels are eligible to be updated. Pixels outside the area defined by the clip mask remain untouched.</p> 96 97<p>Given the varied image formats and image processing requirements, we implemented the ImageMagick <a href="architecture.html#cache">pixel cache</a> to provide convenient sequential or parallel access to any pixel on demand anywhere inside the image region (i.e. <a href="architecture.html#authentic-pixels">authentic pixels</a>) and from any image in a sequence. In addition, the pixel cache permits access to pixels outside the boundaries defined by the image (i.e. <a href="architecture.html#virtual-pixels">virtual pixels</a>).</p> 98 99<p>In addition to pixels, images have a plethora of <a href="architecture.html#properties">image properties and profiles</a>. Properties include the well known attributes such as width, height, depth, and colorspace. An image may have optional properties which might include the image author, a comment, a create date, and others. Some images also include profiles for color management, or EXIF, IPTC, 8BIM, or XMP informational profiles. ImageMagick provides command line options and programming methods to get, set, or view image properties or profiles or apply profiles.</p> 100 101<p>ImageMagick consists of nearly a half million lines of C code and optionally depends on several million lines of code in dependent libraries (e.g. JPEG, PNG, TIFF libraries). Given that, one might expect a huge architecture document. However, a great majority of image processing is simply accessing pixels and its metadata and our simple, elegant, and efficient implementation makes this easy for the ImageMagick developer. We discuss the implementation of the pixel cache and getting and setting image properties and profiles in the next few sections. Next, we discuss using ImageMagick within a <a href="architecture.html#threads">thread</a> of execution. In the final sections, we discuss <a href="architecture.html#coders">image coders</a> to read or write a particular image format followed by a few words on creating a <a href="architecture.html#filters">filter</a> to access or update pixels based on your custom requirements.</p> 102 103<h2 class="magick-header"><a id="cache"></a>The Pixel Cache</h2> 104 105<p>The ImageMagick pixel cache is a repository for image pixels with up to 32 channels. The channels are stored contiguously at the depth specified when ImageMagick was built. The channel depths are 8 bits-per-pixel component for the Q8 version of ImageMagick, 16 bits-per-pixel component for the Q16 version, and 32 bits-per-pixel component for the Q32 version. By default pixel components are 32-bit floating-bit <a href="high-dynamic-range.html">high dynamic-range</a> quantities. The channels can hold any value but typically contain red, green, blue, and alpha intensities or cyan, magenta, yellow, alpha intensities. A channel might contain the colormap indexes for colormapped images or the black channel for CMYK images. The pixel cache storage may be heap memory, disk-backed memory mapped, or on disk. The pixel cache is reference-counted. Only the cache properties are copied when the cache is cloned. The cache pixels are subsequently copied only when you signal your intention to update any of the pixels.</p> 106 107<h3>Create the Pixel Cache</h3> 108 109<p>The pixel cache is associated with an image when it is created and it is initialized when you try to get or put pixels. Here are three common methods to associate a pixel cache with an image:</p> 110 111<dl> 112<dt>Create an image canvas initialized to the background color:</dt><br/> 113<dd><pre>image=AllocateImage(image_info); 114if (SetImageExtent(image,640,480) == MagickFalse) 115 { /* an exception was thrown */ } 116(void) QueryMagickColor("red",&image->background_color,&image->exception); 117SetImageBackgroundColor(image); 118</pre></dd> 119 120<dt>Create an image from a JPEG image on disk:</dt><br/> 121<dd><pre>(void) strcpy(image_info->filename,"image.jpg"): 122image=ReadImage(image_info,exception); 123if (image == (Image *) NULL) 124 { /* an exception was thrown */ } 125</pre></dd> 126<dt>Create an image from a memory based image:</dt><br/> 127<dd><pre>image=BlobToImage(blob_info,blob,extent,exception); 128if (image == (Image *) NULL) 129 { /* an exception was thrown */ } 130</pre></dd> 131</dl> 132 133<p>In our discussion of the pixel cache, we use the <a href="magick-core.html">MagickCore API</a> to illustrate our points, however, the principles are the same for other program interfaces to ImageMagick.</p> 134 135<p>When the pixel cache is initialized, pixels are scaled from whatever bit depth they originated from to that required by the pixel cache. For example, a 1-channel 1-bit monochrome PBM image is scaled to 8-bit gray image, if you are using the Q8 version of ImageMagick, and 16-bit RGBA for the Q16 version. You can determine which version you have with the <a href="command-line-options.html#version">‑version</a> option: </p> 136 137<pre><span class="crtprompt"> </span><span class='crtin'>identify -version</span><span class='crtout'>Version: ImageMagick 7.0.2-0 2016-06-08 Q16 http://www.imagemagick.org</span></pre> 138<p>As you can see, the convenience of the pixel cache sometimes comes with a trade-off in storage (e.g. storing a 1-bit monochrome image as 16-bit is wasteful) and speed (i.e. storing the entire image in memory is generally slower than accessing one scanline of pixels at a time). In most cases, the benefits of the pixel cache typically outweigh any disadvantages.</p> 139 140<h3><a id="authentic-pixels"></a>Access the Pixel Cache</h3> 141 142<p>Once the pixel cache is associated with an image, you typically want to get, update, or put pixels into it. We refer to pixels inside the image region as <a href="architecture.html#authentic-pixels">authentic pixels</a> and outside the region as <a href="architecture.html#virtual-pixels">virtual pixels</a>. Use these methods to access the pixels in the cache:</p> 143<ul> 144 <li><a href="api/cache.html#GetVirtualPixels">GetVirtualPixels()</a>: gets pixels that you do not intend to modify or pixels that lie outside the image region (e.g. pixel @ -1,-3)</li> 145 <li><a href="api/cache.html#GetAuthenticPixels">GetAuthenticPixels()</a>: gets pixels that you intend to modify</li> 146 <li><a href="api/cache.html#QueueAuthenticPixels">QueueAuthenticPixels()</a>: queue pixels that you intend to set</li> 147 <li><a href="api/cache.html#SyncAuthenticPixels">SyncAuthenticPixels()</a>: update the pixel cache with any modified pixels</li> 148</ul> 149 150<p>Here is a typical <a href="magick-core.html">MagickCore</a> code snippet for manipulating pixels in the pixel cache. In our example, we copy pixels from the input image to the output image and decrease the intensity by 10%:</p> 151 152<pre class="pre-scrollable">const Quantum 153 *p; 154 155Quantum 156 *q; 157 158ssize_t 159 x, 160 y; 161 162destination=CloneImage(source,source->columns,source->rows,MagickTrue, 163 exception); 164if (destination == (Image *) NULL) 165 { /* an exception was thrown */ } 166for (y=0; y < (ssize_t) source->rows; y++) 167{ 168 p=GetVirtualPixels(source,0,y,source->columns,1,exception); 169 q=GetAuthenticPixels(destination,0,y,destination->columns,1,exception); 170 if ((p == (const Quantum *) NULL) || (q == (Quantum *) NULL) 171 break; 172 for (x=0; x < (ssize_t) source->columns; x++) 173 { 174 SetPixelRed(image,90*p->red/100,q); 175 SetPixelGreen(image,90*p->green/100,q); 176 SetPixelBlue(image,90*p->blue/100,q); 177 SetPixelAlpha(image,90*p->opacity/100,q); 178 p+=GetPixelChannels(source); 179 q+=GetPixelChannels(destination); 180 } 181 if (SyncAuthenticPixels(destination,exception) == MagickFalse) 182 break; 183} 184if (y < (ssize_t) source->rows) 185 { /* an exception was thrown */ } 186</pre> 187 188<p>When we first create the destination image by cloning the source image, the pixel cache pixels are not copied. They are only copied when you signal your intentions to modify or set the pixel cache by calling <a href="api/cache.html#GetAuthenticPixels">GetAuthenticPixels()</a> or <a href="api/cache.html#QueueAuthenticPixels">QueueAuthenticPixels()</a>. Use <a href="api/cache.html#QueueAuthenticPixels">QueueAuthenticPixels()</a> if you want to set new pixel values rather than update existing ones. You could use GetAuthenticPixels() to set pixel values but it is slightly more efficient to use QueueAuthenticPixels() instead. Finally, use <a href="api/cache.html#SyncAuthenticPixels">SyncAuthenticPixels()</a> to ensure any updated pixels are pushed to the pixel cache.</p> 189 190<p>Recall how we mentioned that the indexes of a colormapped image or the black channel of a CMYK image are stored separately. Use <a href="api/cache.html#GetVirtualIndexQueue">GetVirtualIndexQueue()</a> (to read the indexes) or <a href="api/cache.html#GetAuthenticIndexQueue">GetAuthenticIndexQueue()</a> (to update the indexes) to gain access to this channel. For example, to print the colormap indexes, use:</p> 191 192<pre>const IndexPacket 193 *indexes; 194 195for (y=0; y < (ssize_t) source->rows; y++) 196{ 197 p=GetVirtualPixels(source,0,y,source->columns,1); 198 if (p == (const Quantum *) NULL) 199 break; 200 indexes=GetVirtualIndexQueue(source); 201 for (x=0; x < (ssize_t) source->columns; x++) 202 (void) printf("%d\n",GetPixelIndex(indexes+x)); 203} 204if (y < (ssize_t) source->rows) 205 /* an exception was thrown */ 206</pre> 207 208<p>The pixel cache manager decides whether to give you direct or indirect access to the image pixels. In some cases the pixels are staged to an intermediate buffer-- and that is why you must call SyncAuthenticPixels() to ensure this buffer is <var>pushed</var> out to the pixel cache to guarantee the corresponding pixels in the cache are updated. For this reason we recommend that you only read or update a scanline or a few scanlines of pixels at a time. However, you can get any rectangular region of pixels you want. GetAuthenticPixels() requires that the region you request is within the bounds of the image area. For a 640 by 480 image, you can get a scanline of 640 pixels at row 479 but if you ask for a scanline at row 480, an exception is returned (rows are numbered starting at 0). GetVirtualPixels() does not have this constraint. For example,</p> 209 210<pre>p=GetVirtualPixels(source,-3,-3,source->columns+3,6,exception); 211</pre> 212 213<p>gives you the pixels you asked for without complaint, even though some are not within the confines of the image region.</p> 214 215<h3><a id="virtual-pixels"></a>Virtual Pixels</h3> 216 217<p>There are a plethora of image processing algorithms that require a neighborhood of pixels about a pixel of interest. The algorithm typically includes a caveat concerning how to handle pixels around the image boundaries, known as edge pixels. With virtual pixels, you do not need to concern yourself about special edge processing other than choosing which virtual pixel method is most appropriate for your algorithm.</p> 218 <p>Access to the virtual pixels are controlled by the <a href="api/cache.html#SetImageVirtualPixelMethod">SetImageVirtualPixelMethod()</a> method from the MagickCore API or the <a href="command-line-options.html#virtual-pixel">‑virtual‑pixel</a> option from the command line. The methods include:</p> 219 220<dl class="dl-horizontal"> 221<dt>background</dt> 222<dd>the area surrounding the image is the background color</dd> 223<dt>black</dt> 224<dd>the area surrounding the image is black</dd> 225<dt>checker-tile</dt> 226<dd>alternate squares with image and background color</dd> 227<dt>dither</dt> 228<dd>non-random 32x32 dithered pattern</dd> 229<dt>edge</dt> 230<dd>extend the edge pixel toward infinity (default)</dd> 231<dt>gray</dt> 232<dd>the area surrounding the image is gray</dd> 233<dt>horizontal-tile</dt> 234<dd>horizontally tile the image, background color above/below</dd> 235<dt>horizontal-tile-edge</dt> 236<dd>horizontally tile the image and replicate the side edge pixels</dd> 237<dt>mirror</dt> 238<dd>mirror tile the image</dd> 239<dt>random</dt> 240<dd>choose a random pixel from the image</dd> 241<dt>tile</dt> 242<dd>tile the image</dd> 243<dt>transparent</dt> 244<dd>the area surrounding the image is transparent blackness</dd> 245<dt>vertical-tile</dt> 246<dd>vertically tile the image, sides are background color</dd> 247<dt>vertical-tile-edge</dt> 248<dd>vertically tile the image and replicate the side edge pixels</dd> 249<dt>white</dt> 250<dd>the area surrounding the image is white</dd> 251</dl> 252 253 254<h3>Cache Storage and Resource Requirements</h3> 255 256<p>Recall that this simple and elegant design of the ImageMagick pixel cache comes at a cost in terms of storage and processing speed. The pixel cache storage requirements scales with the area of the image and the bit depth of the pixel components. For example, if we have a 640 by 480 image and we are using the Q16 version of ImageMagick, the pixel cache consumes image <var>width * height * bit-depth / 8 * channels</var> bytes or approximately 2.3 mebibytes (i.e. 640 * 480 * 2 * 4). Not too bad, but what if your image is 25000 by 25000 pixels? The pixel cache requires approximately 4.7 gibibytes of storage. Ouch. ImageMagick accounts for possible huge storage requirements by caching large images to disk rather than memory. Typically the pixel cache is stored in memory using heap memory. If heap memory is exhausted, we create the pixel cache on disk and attempt to memory-map it. If memory-map memory is exhausted, we simply use standard disk I/O. Disk storage is cheap but it is also very slow, upwards of 1000 times slower than memory. We can get some speed improvements, up to 5 times, if we use memory mapping to the disk-based cache. These decisions about storage are made <var>automagically</var> by the pixel cache manager negotiating with the operating system. However, you can influence how the pixel cache manager allocates the pixel cache with <var>cache resource limits</var>. The limits include:</p> 257 258<dl class="dl-horizontal"> 259 <dt>width</dt> 260 <dd>maximum width of an image. Exceed this limit and an exception is thrown and processing stops.</dd> 261 <dt>height</dt> 262 <dd>maximum height of an image. Exceed this limit and an exception is thrown and processing stops.</dd> 263 <dt>area</dt> 264 <dd>maximum area in bytes of any one image that can reside in the pixel cache memory. If this limit is exceeded, the image is automagically cached to disk and optionally memory-mapped.</dd> 265 <dt>memory</dt> 266 <dd>maximum amount of memory in bytes to allocate for the pixel cache from the heap.</dd> 267 <dt>map</dt> 268 <dd>maximum amount of memory map in bytes to allocate for the pixel cache.</dd> 269 <dt>disk</dt> 270 <dd>maximum amount of disk space in bytes permitted for use by the pixel cache. If this limit is exceeded, the pixel cache is not created and a fatal exception is thrown.</dd> 271 <dt>files</dt> 272 <dd>maximum number of open pixel cache files. When this limit is exceeded, any subsequent pixels cached to disk are closed and reopened on demand. This behavior permits a large number of images to be accessed simultaneously on disk, but without a speed penalty due to repeated open/close calls.</dd> 273 <dt>thread</dt> 274 <dd>maximum number of threads that are permitted to run in parallel.</dd> 275 <dt>time</dt> 276 <dd>maximum number of seconds that the process is permitted to execute. Exceed this limit and an exception is thrown and processing stops.</dd> 277</dl> 278 279<p>To determine the current setting of these limits, use this command:</p> 280 281<pre> 282-> identify -list resource 283Resource limits: 284 Width: 100MP 285 Height: 100MP 286 Area: 25.181GB 287 Memory: 11.726GiB 288 Map: 23.452GiB 289 Disk: unlimited 290 File: 768 291 Thread: 12 292 Throttle: 0 293 Time: unlimited 294</pre> 295 296<p>You can set these limits either as a <a href="resources.html#configure">policy</a> (see <a href="../source/policy.xml">policy.xml</a>), with an <a href="resources.html#environment">environment variable</a>, with the <a href="command-line-options.html#limit">-limit</a> command line option, or with the <a href="api/resource.html#SetMagickResourceLimit">SetMagickResourceLimit()</a> MagickCore API method. As an example, our online web interface to ImageMagick, <a href="http://www.imagemagick.org/MagickStudio/scripts/MagickStudio.cgi">ImageMagick Studio</a>, includes these policy limits to help prevent a denial-of-service:</p> 297<pre> 298<policymap> 299 <policy domain="resource" name="temporary-path" value="/tmp"/> 300 <policy domain="resource" name="memory" value="256MiB"/> 301 <policy domain="resource" name="map" value="512MiB"/> 302 <policy domain="resource" name="width" value="8KP"/> 303 <policy domain="resource" name="height" value="8KP"/> 304 <policy domain="resource" name="area" value="128MB"/> 305 <policy domain="resource" name="disk" value="1GiB"/> 306 <policy domain="resource" name="file" value="768"/> 307 <policy domain="resource" name="thread" value="2"/> 308 <policy domain="resource" name="throttle" value="0"/> 309 <policy domain="resource" name="time" value="120"/> 310 <policy domain="system" name="precision" value="6"/> 311 <policy domain="cache" name="shared-secret" value="replace with your secret phrase" stealth="true"/> 312 <policy domain="delegate" rights="none" pattern="HTTPS" /> 313 <policy domain="path" rights="none" pattern="@*"/> <!-- indirect reads not permitted --> 314</policymap> 315</pre> 316<p>Since we process multiple simultaneous sessions, we don't want any one session consuming all the available memory.With this policy, large images are cached to disk. If the image is too large and exceeds the pixel cache disk limit, the program exits. In addition, we place a time limit to prevent any run-away processing tasks. If any one image has a width or height that exceeds 8192 pixels, an exception is thrown and processing stops. As of ImageMagick 7.0.1-8 you can prevent the use of any delegate or all delegates (set the pattern to "*"). Note, prior to this release, use a domain of "coder" to prevent delegate usage (e.g. domain="coder" rights="none" pattern="HTTPS"). The policy also prevents indirect reads. If you want to, for example, read text from a file (e.g. caption:@myCaption.txt), you'll need to remove this policy.</p> 317 318<p>Note, the cache limits are global to each invocation of ImageMagick, meaning if you create several images, the combined resource requirements are compared to the limit to determine the pixel cache storage disposition.</p> 319 320<p>To determine which type and how much resources are consumed by the pixel cache, add the <a href="command-line-options.html#debug">-debug cache</a> option to the command-line:</p> 321<pre>-> convert -debug cache logo: -sharpen 3x2 null: 3222016-12-17T13:33:42-05:00 0:00.000 0.000u 7.0.0 Cache convert: cache.c/DestroyPixelCache/1275/Cache 323 destroy 3242016-12-17T13:33:42-05:00 0:00.000 0.000u 7.0.0 Cache convert: cache.c/OpenPixelCache/3834/Cache 325 open LOGO[0] (Heap Memory, 640x480x4 4.688MiB) 3262016-12-17T13:33:42-05:00 0:00.010 0.000u 7.0.0 Cache convert: cache.c/OpenPixelCache/3834/Cache 327 open LOGO[0] (Heap Memory, 640x480x3 3.516MiB) 3282016-12-17T13:33:42-05:00 0:00.010 0.000u 7.0.0 Cache convert: cache.c/ClonePixelCachePixels/1044/Cache 329 Memory => Memory 3302016-12-17T13:33:42-05:00 0:00.020 0.010u 7.0.0 Cache convert: cache.c/ClonePixelCachePixels/1044/Cache 331 Memory => Memory 3322016-12-17T13:33:42-05:00 0:00.020 0.010u 7.0.0 Cache convert: cache.c/OpenPixelCache/3834/Cache 333 open LOGO[0] (Heap Memory, 640x480x3 3.516MiB) 3342016-12-17T13:33:42-05:00 0:00.050 0.100u 7.0.0 Cache convert: cache.c/DestroyPixelCache/1275/Cache 335 destroy LOGO[0] 3362016-12-17T13:33:42-05:00 0:00.050 0.100u 7.0.0 Cache convert: cache.c/DestroyPixelCache/1275/Cache 337 destroy LOGO[0] 338</pre> 339<p>This command utilizes a pixel cache in memory. The logo consumed 4.688MiB and after it was sharpened, 3.516MiB.</p> 340 341 342<h3>Distributed Pixel Cache</h3> 343<p>A distributed pixel cache is an extension of the traditional pixel cache available on a single host. The distributed pixel cache may span multiple servers so that it can grow in size and transactional capacity to support very large images. Start up the pixel cache server on one or more machines. When you read or operate on an image and the local pixel cache resources are exhausted, ImageMagick contacts one or more of these remote pixel servers to store or retrieve pixels. The distributed pixel cache relies on network bandwidth to marshal pixels to and from the remote server. As such, it will likely be significantly slower than a pixel cache utilizing local storage (e.g. memory, disk, etc.).</p> 344<pre> 345convert -distribute-cache 6668 & // start on 192.168.100.50 346convert -define registry:cache:hosts=192.168.100.50:6668 myimage.jpg -sharpen 5x2 mimage.png 347</pre> 348 349<h3>Cache Views</h3> 350 351<p>GetVirtualPixels(), GetAuthenticPixels(), QueueAuthenticPixels(), and SyncAuthenticPixels(), from the MagickCore API, can only deal with one pixel cache area per image at a time. Suppose you want to access the first and last scanline from the same image at the same time? The solution is to use a <var>cache view</var>. A cache view permits you to access as many areas simultaneously in the pixel cache as you require. The cache view <a href="api/cache-view.html">methods</a> are analogous to the previous methods except you must first open a view and close it when you are finished with it. Here is a snippet of MagickCore code that permits us to access the first and last pixel row of the image simultaneously:</p> 352 353<pre class="pre-scrollable">CacheView 354 *view_1, 355 *view_2; 356 357view_1=AcquireVirtualCacheView(source,exception); 358view_2=AcquireVirtualCacheView(source,exception); 359for (y=0; y < (ssize_t) source->rows; y++) 360{ 361 u=GetCacheViewVirtualPixels(view_1,0,y,source->columns,1,exception); 362 v=GetCacheViewVirtualPixels(view_2,0,source->rows-y-1,source->columns,1,exception); 363 if ((u == (const Quantum *) NULL) || (v == (const Quantum *) NULL)) 364 break; 365 for (x=0; x < (ssize_t) source->columns; x++) 366 { 367 /* do something with u & v here */ 368 } 369} 370view_2=DestroyCacheView(view_2); 371view_1=DestroyCacheView(view_1); 372if (y < (ssize_t) source->rows) 373 { /* an exception was thrown */ } 374</pre> 375 376<h3>Magick Persistent Cache Format</h3> 377 378<p>Recall that each image format is decoded by ImageMagick and the pixels are deposited in the pixel cache. If you write an image, the pixels are read from the pixel cache and encoded as required by the format you are writing (e.g. GIF, PNG, etc.). The Magick Persistent Cache (MPC) format is designed to eliminate the overhead of decoding and encoding pixels to and from an image format. MPC writes two files. One, with the extension <code>.mpc</code>, retains all the properties associated with the image or image sequence (e.g. width, height, colorspace, etc.) and the second, with the extension <code>.cache</code>, is the pixel cache in the native raw format. When reading an MPC image file, ImageMagick reads the image properties and memory maps the pixel cache on disk eliminating the need for decoding the image pixels. The tradeoff is in disk space. MPC is generally larger in file size than most other image formats.</p> 379<p>The most efficient use of MPC image files is a write-once, read-many-times pattern. For example, your workflow requires extracting random blocks of pixels from the source image. Rather than re-reading and possibly decompressing the source image each time, we use MPC and map the image directly to memory.</p> 380 381<h3>Best Practices</h3> 382 383<p>Although you can request any pixel from the pixel cache, any block of pixels, any scanline, multiple scanlines, any row, or multiple rows with the GetVirtualPixels(), GetAuthenticPixels(), QueueAuthenticPixels, GetCacheViewVirtualPixels(), GetCacheViewAuthenticPixels(), and QueueCacheViewAuthenticPixels() methods, ImageMagick is optimized to return a few pixels or a few pixels rows at time. There are additional optimizations if you request a single scanline or a few scanlines at a time. These methods also permit random access to the pixel cache, however, ImageMagick is optimized for sequential access. Although you can access scanlines of pixels sequentially from the last row of the image to the first, you may get a performance boost if you access scanlines from the first row of the image to the last, in sequential order.</p> 384 385<p>You can get, modify, or set pixels in row or column order. However, it is more efficient to access the pixels by row rather than by column.</p> 386 387<p>If you update pixels returned from GetAuthenticPixels() or GetCacheViewAuthenticPixels(), don't forget to call SyncAuthenticPixels() or SyncCacheViewAuthenticPixels() respectively to ensure your changes are synchronized with the pixel cache.</p> 388 389<p>Use QueueAuthenticPixels() or QueueCacheViewAuthenticPixels() if you are setting an initial pixel value. The GetAuthenticPixels() or GetCacheViewAuthenticPixels() method reads pixels from the cache and if you are setting an initial pixel value, this read is unnecessary. Don't forget to call SyncAuthenticPixels() or SyncCacheViewAuthenticPixels() respectively to push any pixel changes to the pixel cache.</p> 390 391<p>GetVirtualPixels(), GetAuthenticPixels(), QueueAuthenticPixels(), and SyncAuthenticPixels() are slightly more efficient than their cache view counter-parts. However, cache views are required if you need access to more than one region of the image simultaneously or if more than one <a href="architecture.html#threads">thread of execution</a> is accessing the image.</p> 392 393<p>You can request pixels outside the bounds of the image with GetVirtualPixels() or GetCacheViewVirtualPixels(), however, it is more efficient to request pixels within the confines of the image region.</p> 394 395<p>Although you can force the pixel cache to disk using appropriate resource limits, disk access can be upwards of 1000 times slower than memory access. For fast, efficient, access to the pixel cache, try to keep the pixel cache in heap memory.</p> 396 397<p>The ImageMagick Q16 version of ImageMagick permits you to read and write 16 bit images without scaling but the pixel cache consumes twice as many resources as the Q8 version. If your system has constrained memory or disk resources, consider the Q8 version of ImageMagick. In addition, the Q8 version typically executes faster than the Q16 version.</p> 398 399<p>A great majority of image formats and algorithms restrict themselves to a fixed range of pixel values from 0 to some maximum value, for example, the Q16 version of ImageMagick permit intensities from 0 to 65535. High dynamic-range imaging (HDRI), however, permits a far greater dynamic range of exposures (i.e. a large difference between light and dark areas) than standard digital imaging techniques. HDRI accurately represents the wide range of intensity levels found in real scenes ranging from the brightest direct sunlight to the deepest darkest shadows. Enable <a href="high-dynamic-range.html">HDRI</a> at ImageMagick build time to deal with high dynamic-range images, but be mindful that each pixel component is a 32-bit floating point value. In addition, pixel values are not clamped by default so some algorithms may have unexpected results due to out-of-band pixel values than the non-HDRI version.</p> 400 401<p>If you are dealing with large images, make sure the pixel cache is written to a disk area with plenty of free space. Under Unix, this is typically <code>/tmp</code> and for Windows, <code>c:/temp</code>. You can tell ImageMagick to write the pixel cache to an alternate location and conserve memory with these options:</p> 402 403<pre> 404convert -limit memory 2GB -limit map 4GB -define registry:temporary-path=/data/tmp ... 405</pre> 406 407<p>Set global resource limits for your environment in the <code>policy.xml</code> configuration file.</p> 408 409<p>If you plan on processing the same image many times, consider the MPC format. Reading a MPC image has near-zero overhead because its in the native pixel cache format eliminating the need for decoding the image pixels. Here is an example:</p> 410 411<pre> 412convert image.tif image.mpc 413convert image.mpc -crop 100x100+0+0 +repage 1.png 414convert image.mpc -crop 100x100+100+0 +repage 2.png 415convert image.mpc -crop 100x100+200+0 +repage 3.png 416</pre> 417 418<p>MPC is ideal for web sites. It reduces the overhead of reading and writing an image. We use it exclusively at our <a href="http://www.imagemagick.org/MagickStudio/scripts/MagickStudio.cgi">online image studio</a>.</p> 419 420<h2 class="magick-header"><a id="stream"></a>Streaming Pixels</h2> 421 422<p>ImageMagick provides for streaming pixels as they are read from or written to an image. This has several advantages over the pixel cache. The time and resources consumed by the pixel cache scale with the area of an image, whereas the pixel stream resources scale with the width of an image. The disadvantage is the pixels must be consumed as they are streamed so there is no persistence.</p> 423 424<p>Use <a href="api/stream.html#ReadStream">ReadStream()</a> or <a href="api/stream.html#WriteStream">WriteStream()</a> with an appropriate callback method in your MagickCore program to consume the pixels as they are streaming. Here's an abbreviated example of using ReadStream:</p> 425 426<pre class="pre-scrollable">static size_t StreamPixels(const Image *image,const void *pixels,const size_t columns) 427{ 428 register const Quantum 429 *p; 430 431 MyData 432 *my_data; 433 434 my_data=(MyData *) image->client_data; 435 p=(Quantum *) pixels; 436 if (p != (const Quantum *) NULL) 437 { 438 /* process pixels here */ 439 } 440 return(columns); 441} 442 443... 444 445/* invoke the pixel stream here */ 446image_info->client_data=(void *) MyData; 447image=ReadStream(image_info,&StreamPixels,exception); 448</pre> 449 450<p>We also provide a lightweight tool, <a href="stream.html">stream</a>, to stream one or more pixel components of the image or portion of the image to your choice of storage formats. It writes the pixel components as they are read from the input image a row at a time making <a href="stream.html">stream</a> desirable when working with large images or when you require raw pixel components. A majority of the image formats stream pixels (red, green, and blue) from left to right and top to bottom. However, a few formats do not support this common ordering (e.g. the PSD format).</p> 451 452<h2 class="magick-header"><a id="properties"></a>Image Properties and Profiles</h2> 453 454<p>Images have metadata associated with them in the form of properties (e.g. width, height, description, etc.) and profiles (e.g. EXIF, IPTC, color management). ImageMagick provides convenient methods to get, set, or update image properties and get, set, update, or apply profiles. Some of the more popular image properties are associated with the Image structure in the MagickCore API. For example:</p> 455 456<pre>(void) printf("image width: %lu, height: %lu\n",image->columns,image->rows); 457</pre> 458 459<p>For a great majority of image properties, such as an image comment or description, we use the <a href="api/property.html#GetImageProperty">GetImageProperty()</a> and <a href="api/property.html#SetImageProperty">SetImageProperty()</a> methods. Here we set a property and fetch it right back:</p> 460 461<pre>const char 462 *comment; 463 464(void) SetImageProperty(image,"comment","This space for rent"); 465comment=GetImageProperty(image,"comment"); 466if (comment == (const char *) NULL) 467 (void) printf("Image comment: %s\n",comment); 468</pre> 469 470<p>ImageMagick supports artifacts with the GetImageArtifact() and SetImageArtifact() methods. Artifacts are stealth properties that are not exported to image formats (e.g. PNG).</p> 471 472<p>Image profiles are handled with <a href="api/profile.html#GetImageProfile">GetImageProfile()</a>, <a href="api/profile.html#SetImageProfile">SetImageProfile()</a>, and <a href="api/profile.html#ProfileImage">ProfileImage()</a> methods. Here we set a profile and fetch it right back:</p> 473 474<pre>StringInfo 475 *profile; 476 477profile=AcquireStringInfo(length); 478SetStringInfoDatum(profile,my_exif_profile); 479(void) SetImageProfile(image,"EXIF",profile); 480DestroyStringInfo(profile); 481profile=GetImageProfile(image,"EXIF"); 482if (profile != (StringInfo *) NULL) 483 (void) PrintStringInfo(stdout,"EXIF",profile); 484</pre> 485 486<h2 class="magick-header"><a id="tera-pixel"></a>Large Image Support</h2> 487<p>ImageMagick can read, process, or write mega-, giga-, or tera-pixel image sizes. An image width or height can range from 1 to 2 giga-pixels on a 32 bit OS and up to 9 exa-pixels on a 64-bit OS. Note, that some image formats have restrictions on image size. For example, Photoshop images are limited to 300,000 pixels for width or height. Here we resize an image to a quarter million pixels square:</p> 488 489<pre> 490convert logo: -resize 250000x250000 logo.miff 491</pre> 492 493<p>For large images, ImageMagick will likely create a pixel cache on disk. Make sure you have plenty of temporary disk space. If your default temporary disk partition is too small, tell ImageMagick to use another partition with plenty of free space. For example:</p> 494 495<pre> 496convert -define registry:temporary-path=/data/tmp logo: \ <br/> -resize 250000x250000 logo.miff 497</pre> 498 499<p>To ensure large images do not consume all the memory on your system, force the image pixels to memory-mapped disk with resource limits:</p> 500 501<pre> 502convert -define registry:temporary-path=/data/tmp -limit memory 16mb \ 503 logo: -resize 250000x250000 logo.miff 504</pre> 505 506<p>Here we force all image pixels to disk:</p> 507 508<pre> 509convert -define registry:temporary-path=/data/tmp -limit area 0 \ 510 logo: -resize 250000x250000 logo.miff 511</pre> 512 513<p>Caching pixels to disk is about 1000 times slower than memory. Expect long run times when processing large images on disk with ImageMagick. You can monitor progress with this command:</p> 514 515<pre>convert -monitor -limit memory 2GiB -limit map 4GiB -define registry:temporary-path=/data/tmp \ 516 logo: -resize 250000x250000 logo.miff 517</pre> 518 519<p>For really large images, or if there is limited resources on your host, you can utilize a distributed pixel cache on one or more remote hosts:</p> 520<pre> 521convert -distribute-cache 6668 & // start on 192.168.100.50 522convert -distribute-cache 6668 & // start on 192.168.100.51 523convert -limit memory 2mb -limit map 2mb -limit disk 2gb \ 524 -define registry:cache:hosts=192.168.100.50:6668,192.168.100.51:6668 \ 525 myhugeimage.jpg -sharpen 5x2 myhugeimage.png 526</pre> 527 528<h2 class="magick-header"><a id="threads"></a>Threads of Execution</h2> 529 530<p>Many of ImageMagick's internal algorithms are threaded to take advantage of speed-ups offered by the multicore processor chips. However, you are welcome to use ImageMagick algorithms in your threads of execution with the exception of the MagickCore's GetVirtualPixels(), GetAuthenticPixels(), QueueAuthenticPixels(), or SyncAuthenticPixels() pixel cache methods. These methods are intended for one thread of execution only with the exception of an OpenMP parallel section. To access the pixel cache with more than one thread of execution, use a cache view. We do this for the <a href="api/composite.html#CompositeImage">CompositeImage()</a> method, for example. Suppose we want to composite a single image over a different image in each thread of execution. If we use GetVirtualPixels(), the results are unpredictable because multiple threads would likely be asking for different areas of the pixel cache simultaneously. Instead we use GetCacheViewVirtualPixels() which creates a unique view for each thread of execution ensuring our program behaves properly regardless of how many threads are invoked. The other program interfaces, such as the <a href="magick-wand.html">MagickWand API</a>, are completely thread safe so there are no special precautions for threads of execution.</p> 531 532<p>Here is an MagickCore code snippet that takes advantage of threads of execution with the <a href="openmp.html">OpenMP</a> programming paradigm:</p> 533 534<pre class="pre-scrollable">CacheView 535 *image_view; 536 537MagickBooleanType 538 status; 539 540ssize_t 541 y; 542 543status=MagickTrue; 544image_view=AcquireVirtualCacheView(image,exception); 545#pragma omp parallel for schedule(dynamic,4) shared(status) 546for (y=0; y < (ssize_t) image->rows; y++) 547{ 548 register IndexPacket 549 *indexes; 550 551 register Quantum 552 *q; 553 554 register ssize_t 555 x; 556 557 if (status == MagickFalse) 558 continue; 559 q=GetCacheViewAuthenticPixels(image_view,0,y,image->columns,1,exception); 560 if (q == (Quantum *) NULL) 561 { 562 status=MagickFalse; 563 continue; 564 } 565 indexes=GetCacheViewAuthenticIndexQueue(image_view); 566 for (x=0; x < (ssize_t) image->columns; x++) 567 { 568 SetPixelRed(image,...,q); 569 SetPixelGreen(image,...,q); 570 SetPixelBlue(image,...,q); 571 SetPixelAlpha(image,...,q); 572 if (indexes != (IndexPacket *) NULL) 573 SetPixelIndex(indexes+x,...); 574 q+=GetPixelChannels(image); 575 } 576 if (SyncCacheViewAuthenticPixels(image_view,exception) == MagickFalse) 577 status=MagickFalse; 578} 579image_view=DestroyCacheView(image_view); 580if (status == MagickFalse) 581 perror("something went wrong"); 582</pre> 583 584<p>This code snippet converts an uncompressed Windows bitmap to a Magick++ image:</p> 585 586<pre class="pre-scrollable">#include "Magick++.h" 587#include <assert.h> 588#include "omp.h" 589 590void ConvertBMPToImage(const BITMAPINFOHEADER *bmp_info, 591 const unsigned char *restrict pixels,Magick::Image *image) 592{ 593 /* 594 Prepare the image so that we can modify the pixels directly. 595 */ 596 assert(bmp_info->biCompression == BI_RGB); 597 assert(bmp_info->biWidth == image->columns()); 598 assert(abs(bmp_info->biHeight) == image->rows()); 599 image->modifyImage(); 600 if (bmp_info->biBitCount == 24) 601 image->type(MagickCore::TrueColorType); 602 else 603 image->type(MagickCore::TrueColorMatteType); 604 register unsigned int bytes_per_row=bmp_info->biWidth*bmp_info->biBitCount/8; 605 if (bytes_per_row % 4 != 0) { 606 bytes_per_row=bytes_per_row+(4-bytes_per_row % 4); // divisible by 4. 607 } 608 /* 609 Copy all pixel data, row by row. 610 */ 611 #pragma omp parallel for 612 for (int y=0; y < int(image->rows()); y++) 613 { 614 int 615 row; 616 617 register const unsigned char 618 *restrict p; 619 620 register MagickCore::Quantum 621 *restrict q; 622 623 row=(bmp_info->biHeight > 0) ? (image->rows()-y-1) : y; 624 p=pixels+row*bytes_per_row; 625 q=image->setPixels(0,y,image->columns(),1); 626 for (int x=0; x < int(image->columns()); x++) 627 { 628 SetPixelBlue(image,p[0],q); 629 SetPixelGreen(image,p[1],q); 630 SetPixelRed(image,p[2],q); 631 if (bmp_info->biBitCount == 32) { 632 SetPixelAlpha(image,p[3],q); 633 } 634 q+=GetPixelChannels(image); 635 p+=bmp_info->biBitCount/8; 636 } 637 image->syncPixels(); // sync pixels to pixel cache. 638 } 639 return; 640}</pre> 641 642<p>If you call the ImageMagick API from your OpenMP-enabled application and you intend to dynamically increase the number of threads available in subsequent parallel regions, be sure to perform the increase <var>before</var> you call the API otherwise ImageMagick may fault.</p> 643 644<p><a href="api/wand-view.html">MagickWand</a> supports wand views. A view iterates over the entire, or portion, of the image in parallel and for each row of pixels, it invokes a callback method you provide. This limits most of your parallel programming activity to just that one module. There are similar methods in <a href="api/image-view.html">MagickCore</a>. For an example, see the same sigmoidal contrast algorithm implemented in both <a href="magick-wand.html#wand-view">MagickWand</a> and <a href="magick-core.html#image-view">MagickCore</a>.</p> 645 646<p>In most circumstances, the default number of threads is set to the number of processor cores on your system for optimal performance. However, if your system is hyperthreaded or if you are running on a virtual host and only a subset of the processors are available to your server instance, you might get an increase in performance by setting the thread <a href="resources.html#configure">policy</a> or the <a href="resources.html#environment">MAGICK_THREAD_LIMIT</a> environment variable. For example, your virtual host has 8 processors but only 2 are assigned to your server instance. The default of 8 threads can cause severe performance problems. One solution is to limit the number of threads to the available processors in your <a href="../source/policy.xml">policy.xml</a> configuration file:</p> 647 648<pre> 649<policy domain="resource" name="thread" value="2"/> 650</pre> 651 652<p>Or suppose your 12 core hyperthreaded computer defaults to 24 threads. Set the MAGICK_THREAD_LIMIT environment variable and you will likely get improved performance:</p> 653 654<pre> 655export MAGICK_THREAD_LIMIT=12 656</pre> 657 658<p>The OpenMP committee has not defined the behavior of mixing OpenMP with other threading models such as Posix threads. However, using modern releases of Linux, OpenMP and Posix threads appear to interoperate without complaint. If you want to use Posix threads from a program module that calls one of the ImageMagick application programming interfaces (e.g. MagickCore, MagickWand, Magick++, etc.) from Mac OS X or an older Linux release, you may need to disable OpenMP support within ImageMagick. Add the <code>--disable-openmp</code> option to the configure script command line and rebuild and reinstall ImageMagick.</p> 659 660<h4>Threading Performance</h4> 661<p>It can be difficult to predict behavior in a parallel environment. Performance might depend on a number of factors including the compiler, the version of the OpenMP library, the processor type, the number of cores, the amount of memory, whether hyperthreading is enabled, the mix of applications that are executing concurrently with ImageMagick, or the particular image-processing algorithm you utilize. The only way to be certain of optimal performance, in terms of the number of threads, is to benchmark. ImageMagick includes progressive threading when benchmarking a command and returns the elapsed time and efficiency for one or more threads. This can help you identify how many threads is the most efficient in your environment. For this benchmark we sharpen a 1920x1080 image of a model 10 times with 1 to 12 threads:</p> 662<pre> 663-> convert -bench 10 model.png -sharpen 5x2 null: 664Performance[1]: 10i 1.135ips 1.000e 8.760u 0:08.810 665Performance[2]: 10i 2.020ips 0.640e 9.190u 0:04.950 666Performance[3]: 10i 2.786ips 0.710e 9.400u 0:03.590 667Performance[4]: 10i 3.378ips 0.749e 9.580u 0:02.960 668Performance[5]: 10i 4.032ips 0.780e 9.580u 0:02.480 669Performance[6]: 10i 4.566ips 0.801e 9.640u 0:02.190 670Performance[7]: 10i 3.788ips 0.769e 10.980u 0:02.640 671Performance[8]: 10i 4.115ips 0.784e 12.030u 0:02.430 672Performance[9]: 10i 4.484ips 0.798e 12.860u 0:02.230 673Performance[10]: 10i 4.274ips 0.790e 14.830u 0:02.340 674Performance[11]: 10i 4.348ips 0.793e 16.500u 0:02.300 675Performance[12]: 10i 4.525ips 0.799e 18.320u 0:02.210 676</pre> 677<p>The sweet spot for this example is 6 threads. This makes sense since there are 6 physical cores. The other 6 are hyperthreads. It appears that sharpening does not benefit from hyperthreading.</p> 678<p>In certain cases, it might be optimal to set the number of threads to 1 or to disable OpenMP completely with the <a href="resources.html#environment">MAGICK_THREAD_LIMIT</a> environment variable, <a href="command-line-options.html#limit">-limit</a> command line option, or the <a href="resources.html#configure">policy.xml</a> configuration file.</p> 679 680<h2 class="magick-header"><a id="distributed"></a>Heterogeneous Distributed Processing</h2> 681<p>ImageMagick includes support for heterogeneous distributed processing with the <a href="http://en.wikipedia.org/wiki/OpenCL">OpenCL</a> framework. OpenCL kernels within ImageMagick permit image processing algorithms to execute across heterogeneous platforms consisting of CPUs, GPUs, and other processors. Depending on your platform, speed-ups can be an order of magnitude faster than the traditional single CPU.</p> 682 683<p>First verify that your version of ImageMagick includes support for the OpenCL feature:</p> 684 685<pre> 686identify -version 687Features: DPC Cipher Modules OpenCL OpenMP 688</pre> 689 690<p>If so, run this command to realize a significant speed-up for image convolution:</p> 691 692<pre> 693convert image.png -convolve '-1, -1, -1, -1, 9, -1, -1, -1, -1' convolve.png 694</pre> 695 696<p>If an accelerator is not available or if the accelerator fails to respond, ImageMagick reverts to the non-accelerated convolution algorithm.</p> 697 698<p>Here is an example OpenCL kernel that convolves an image:</p> 699 700<pre class="pre-scrollable">static inline long ClampToCanvas(const long offset,const ulong range) 701{ 702 if (offset < 0L) 703 return(0L); 704 if (offset >= range) 705 return((long) (range-1L)); 706 return(offset); 707} 708 709static inline CLQuantum ClampToQuantum(const float value) 710{ 711 if (value < 0.0) 712 return((CLQuantum) 0); 713 if (value >= (float) QuantumRange) 714 return((CLQuantum) QuantumRange); 715 return((CLQuantum) (value+0.5)); 716} 717 718__kernel void Convolve(const __global CLPixelType *source,__constant float *filter, 719 const ulong width,const ulong height,__global CLPixelType *destination) 720{ 721 const ulong columns = get_global_size(0); 722 const ulong rows = get_global_size(1); 723 724 const long x = get_global_id(0); 725 const long y = get_global_id(1); 726 727 const float scale = (1.0/QuantumRange); 728 const long mid_width = (width-1)/2; 729 const long mid_height = (height-1)/2; 730 float4 sum = { 0.0, 0.0, 0.0, 0.0 }; 731 float gamma = 0.0; 732 register ulong i = 0; 733 734 for (long v=(-mid_height); v <= mid_height; v++) 735 { 736 for (long u=(-mid_width); u <= mid_width; u++) 737 { 738 register const ulong index=ClampToCanvas(y+v,rows)*columns+ClampToCanvas(x+u, 739 columns); 740 const float alpha=scale*(QuantumRange-source[index].w); 741 sum.x+=alpha*filter[i]*source[index].x; 742 sum.y+=alpha*filter[i]*source[index].y; 743 sum.z+=alpha*filter[i]*source[index].z; 744 sum.w+=filter[i]*source[index].w; 745 gamma+=alpha*filter[i]; 746 i++; 747 } 748 } 749 750 gamma=1.0/(fabs(gamma) <= MagickEpsilon ? 1.0 : gamma); 751 const ulong index=y*columns+x; 752 destination[index].x=ClampToQuantum(gamma*sum.x); 753 destination[index].y=ClampToQuantum(gamma*sum.y); 754 destination[index].z=ClampToQuantum(gamma*sum.z); 755 destination[index].w=ClampToQuantum(sum.w); 756};</pre> 757 758<p>See <a href="https://github.com/ImageMagick/ImageMagick/tree/ImageMagick-6/magick/accelerate.c">magick/accelerate.c</a> for a complete implementation of image convolution with an OpenCL kernel.</p> 759 760<p>Note, that under Windows, you might have an issue with TDR (Timeout Detection and Recovery of GPUs). Its purpose is to detect runaway tasks hanging the GPU by using an execution time threshold. For some older low-end GPUs running the OpenCL filters in ImageMagick, longer execution times might trigger the TDR mechanism and pre-empt the GPU image filter. When this happens, ImageMagick automatically falls back to the CPU code path and returns the expected results. To avoid pre-emption, increase the <a href="http://msdn.microsoft.com/en-us/library/windows/hardware/gg487368.aspx">TdrDelay</a> registry key.</p> 761 762<h2 class="magick-header"><a id="coders"></a>Custom Image Coders</h2> 763 764<p>An image coder (i.e. encoder / decoder) is responsible for registering, optionally classifying, optionally reading, optionally writing, and unregistering one image format (e.g. PNG, GIF, JPEG, etc.). Registering an image coder alerts ImageMagick a particular format is available to read or write. While unregistering tells ImageMagick the format is no longer available. The classifying method looks at the first few bytes of an image and determines if the image is in the expected format. The reader sets the image size, colorspace, and other properties and loads the pixel cache with the pixels. The reader returns a single image or an image sequence (if the format supports multiple images per file), or if an error occurs, an exception and a null image. The writer does the reverse. It takes the image properties and unloads the pixel cache and writes them as required by the image format.</p> 765 766<p>Here is a listing of a sample <a href="../source/mgk.c">custom coder</a>. It reads and writes images in the MGK image format which is simply an ID followed by the image width and height followed by the RGB pixel values.</p> 767 768<pre class="pre-scrollable">/* 769 Include declarations. 770*/ 771#include "magick/studio.h" 772#include "magick/blob.h" 773#include "magick/blob-private.h" 774#include "magick/colorspace.h" 775#include "magick/exception.h" 776#include "magick/exception-private.h" 777#include "magick/image.h" 778#include "magick/image-private.h" 779#include "magick/list.h" 780#include "magick/magick.h" 781#include "magick/memory_.h" 782#include "magick/monitor.h" 783#include "magick/monitor-private.h" 784#include "magick/quantum-private.h" 785#include "magick/static.h" 786#include "magick/string_.h" 787#include "magick/module.h" 788 789/* 790 Forward declarations. 791*/ 792static MagickBooleanType 793 WriteMGKImage(const ImageInfo *,Image *); 794 795/* 796%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 797% % 798% % 799% % 800% I s M G K % 801% % 802% % 803% % 804%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 805% 806% IsMGK() returns MagickTrue if the image format type, identified by the 807% magick string, is MGK. 808% 809% The format of the IsMGK method is: 810% 811% MagickBooleanType IsMGK(const unsigned char *magick,const size_t length) 812% 813% A description of each parameter follows: 814% 815% o magick: This string is generally the first few bytes of an image file 816% or blob. 817% 818% o length: Specifies the length of the magick string. 819% 820*/ 821static MagickBooleanType IsMGK(const unsigned char *magick,const size_t length) 822{ 823 if (length < 7) 824 return(MagickFalse); 825 if (LocaleNCompare((char *) magick,"id=mgk",7) == 0) 826 return(MagickTrue); 827 return(MagickFalse); 828} 829 830/* 831%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 832% % 833% % 834% % 835% R e a d M G K I m a g e % 836% % 837% % 838% % 839%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 840% 841% ReadMGKImage() reads a MGK image file and returns it. It allocates 842% the memory necessary for the new Image structure and returns a pointer to 843% the new image. 844% 845% The format of the ReadMGKImage method is: 846% 847% Image *ReadMGKImage(const ImageInfo *image_info,ExceptionInfo *exception) 848% 849% A description of each parameter follows: 850% 851% o image_info: the image info. 852% 853% o exception: return any errors or warnings in this structure. 854% 855*/ 856static Image *ReadMGKImage(const ImageInfo *image_info, 857 ExceptionInfo *exception) 858{ 859 char 860 buffer[MaxTextExtent]; 861 862 Image 863 *image; 864 865 MagickBooleanType 866 status; 867 868 register Quantum 869 *q; 870 871 register size_t 872 x; 873 874 register unsigned char 875 *p; 876 877 ssize_t 878 count, 879 y; 880 881 unsigned char 882 *pixels; 883 884 unsigned long 885 columns, 886 rows; 887 888 /* 889 Open image file. 890 */ 891 assert(image_info != (const ImageInfo *) NULL); 892 assert(image_info->signature == MagickSignature); 893 if (image_info->debug != MagickFalse) 894 (void) LogMagickEvent(TraceEvent,GetMagickModule(),"%s",image_info->filename); 895 assert(exception != (ExceptionInfo *) NULL); 896 assert(exception->signature == MagickSignature); 897 image=AcquireImage(image_info); 898 status=OpenBlob(image_info,image,ReadBinaryBlobMode,exception); 899 if (status == MagickFalse) 900 { 901 image=DestroyImageList(image); 902 return((Image *) NULL); 903 } 904 /* 905 Read MGK image. 906 */ 907 (void) ReadBlobString(image,buffer); /* read magic number */ 908 if (IsMGK(buffer,7) == MagickFalse) 909 ThrowReaderException(CorruptImageError,"ImproperImageHeader"); 910 (void) ReadBlobString(image,buffer); 911 count=(ssize_t) sscanf(buffer,"%lu %lu\n",&columns,&rows); 912 if (count <= 0) 913 ThrowReaderException(CorruptImageError,"ImproperImageHeader"); 914 do 915 { 916 /* 917 Initialize image structure. 918 */ 919 image->columns=columns; 920 image->rows=rows; 921 image->depth=8; 922 if ((image_info->ping != MagickFalse) && (image_info->number_scenes != 0)) 923 if (image->scene >= (image_info->scene+image_info->number_scenes-1)) 924 break; 925 /* 926 Convert MGK raster image to pixel packets. 927 */ 928 if (SetImageExtent(image,0,0) == MagickFalse) 929 { 930 InheritException(exception,&image->exception); 931 return(DestroyImageList(image)); 932 } 933 pixels=(unsigned char *) AcquireQuantumMemory((size_t) image->columns,3UL*sizeof(*pixels)); 934 if (pixels == (unsigned char *) NULL) 935 ThrowReaderException(ResourceLimitError,"MemoryAllocationFailed"); 936 for (y=0; y < (ssize_t) image->rows; y++) 937 { 938 count=(ssize_t) ReadBlob(image,(size_t) (3*image->columns),pixels); 939 if (count != (ssize_t) (3*image->columns)) 940 ThrowReaderException(CorruptImageError,"UnableToReadImageData"); 941 p=pixels; 942 q=QueueAuthenticPixels(image,0,y,image->columns,1,exception); 943 if (q == (Quantum *) NULL) 944 break; 945 for (x=0; x < (ssize_t) image->columns; x++) 946 { 947 SetPixelRed(image,ScaleCharToQuantum(*p++),q); 948 SetPixelGreen(image,ScaleCharToQuantum(*p++),q); 949 SetPixelBlue(image,ScaleCharToQuantum(*p++),q); 950 q+=GetPixelChannels(image); 951 } 952 if (SyncAuthenticPixels(image,exception) == MagickFalse) 953 break; 954 if ((image->previous == (Image *) NULL) && 955 (SetImageProgress(image,LoadImageTag,y,image>>rows) == MagickFalse)) 956 break; 957 } 958 pixels=(unsigned char *) RelinquishMagickMemory(pixels); 959 if (EOFBlob(image) != MagickFalse) 960 { 961 ThrowFileException(exception,CorruptImageError,"UnexpectedEndOfFile",image->filename); 962 break; 963 } 964 /* 965 Proceed to next image. 966 */ 967 if (image_info->number_scenes != 0) 968 if (image->scene >= (image_info->scene+image_info->number_scenes-1)) 969 break; 970 *buffer='\0'; 971 (void) ReadBlobString(image,buffer); 972 count=(ssize_t) sscanf(buffer,"%lu %lu\n",&columns,&rows); 973 if (count != 0) 974 { 975 /* 976 Allocate next image structure. 977 */ 978 AcquireNextImage(image_info,image); 979 if (GetNextImageInList(image) == (Image *) NULL) 980 { 981 image=DestroyImageList(image); 982 return((Image *) NULL); 983 } 984 image=SyncNextImageInList(image); 985 status=SetImageProgress(image,LoadImageTag,TellBlob(image),GetBlobSize(image)); 986 if (status == MagickFalse) 987 break; 988 } 989 } while (count != 0); 990 (void) CloseBlob(image); 991 return(GetFirstImageInList(image)); 992} 993 994/* 995%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 996% % 997% % 998% % 999% R e g i s t e r M G K I m a g e % 1000% % 1001% % 1002% % 1003%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1004% 1005% RegisterMGKImage() adds attributes for the MGK image format to 1006% the list of supported formats. The attributes include the image format 1007% tag, a method to read and/or write the format, whether the format 1008% supports the saving of more than one frame to the same file or blob, 1009% whether the format supports native in-memory I/O, and a brief 1010% description of the format. 1011% 1012% The format of the RegisterMGKImage method is: 1013% 1014% unsigned long RegisterMGKImage(void) 1015% 1016*/ 1017ModuleExport unsigned long RegisterMGKImage(void) 1018{ 1019 MagickInfo 1020 *entry; 1021 1022 entry=SetMagickInfo("MGK"); 1023 entry->decoder=(DecodeImageHandler *) ReadMGKImage; 1024 entry->encoder=(EncodeImageHandler *) WriteMGKImage; 1025 entry->magick=(IsImageFormatHandler *) IsMGK; 1026 entry->description=ConstantString("MGK"); 1027 entry->module=ConstantString("MGK"); 1028 (void) RegisterMagickInfo(entry); 1029 return(MagickImageCoderSignature); 1030} 1031 1032/* 1033%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1034% % 1035% % 1036% % 1037% U n r e g i s t e r M G K I m a g e % 1038% % 1039% % 1040% % 1041%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1042% 1043% UnregisterMGKImage() removes format registrations made by the 1044% MGK module from the list of supported formats. 1045% 1046% The format of the UnregisterMGKImage method is: 1047% 1048% UnregisterMGKImage(void) 1049% 1050*/ 1051ModuleExport void UnregisterMGKImage(void) 1052{ 1053 (void) UnregisterMagickInfo("MGK"); 1054} 1055 1056/* 1057%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1058% % 1059% % 1060% % 1061% W r i t e M G K I m a g e % 1062% % 1063% % 1064% % 1065%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1066% 1067% WriteMGKImage() writes an image to a file in red, green, and blue 1068% MGK rasterfile format. 1069% 1070% The format of the WriteMGKImage method is: 1071% 1072% MagickBooleanType WriteMGKImage(const ImageInfo *image_info,Image *image) 1073% 1074% A description of each parameter follows. 1075% 1076% o image_info: the image info. 1077% 1078% o image: The image. 1079% 1080*/ 1081static MagickBooleanType WriteMGKImage(const ImageInfo *image_info,Image *image) 1082{ 1083 char 1084 buffer[MaxTextExtent]; 1085 1086 MagickBooleanType 1087 status; 1088 1089 MagickOffsetType 1090 scene; 1091 1092 register const Quantum 1093 *p; 1094 1095 register ssize_t 1096 x; 1097 1098 register unsigned char 1099 *q; 1100 1101 ssize_t 1102 y; 1103 1104 unsigned char 1105 *pixels; 1106 1107 /* 1108 Open output image file. 1109 */ 1110 assert(image_info != (const ImageInfo *) NULL); 1111 assert(image_info->signature == MagickSignature); 1112 assert(image != (Image *) NULL); 1113 assert(image->signature == MagickSignature); 1114 if (image->debug != MagickFalse) 1115 (void) LogMagickEvent(TraceEvent,GetMagickModule(),"%s",image->filename); 1116 status=OpenBlob(image_info,image,WriteBinaryBlobMode,&image->exception); 1117 if (status == MagickFalse) 1118 return(status); 1119 scene=0; 1120 do 1121 { 1122 /* 1123 Allocate memory for pixels. 1124 */ 1125 if (image->colorspace != RGBColorspace) 1126 (void) SetImageColorspace(image,RGBColorspace); 1127 pixels=(unsigned char *) AcquireQuantumMemory((size_t) image->columns, 1128 3UL*sizeof(*pixels)); 1129 if (pixels == (unsigned char *) NULL) 1130 ThrowWriterException(ResourceLimitError,"MemoryAllocationFailed"); 1131 /* 1132 Initialize raster file header. 1133 */ 1134 (void) WriteBlobString(image,"id=mgk\n"); 1135 (void) FormatLocaleString(buffer,MaxTextExtent,"%lu %lu\n", 1136 image->columns,image->rows); 1137 (void) WriteBlobString(image,buffer); 1138 for (y=0; y < (ssize_t) image->rows; y++) 1139 { 1140 p=GetVirtualPixels(image,0,y,image->columns,1,&image->exception); 1141 if (p == (const Quantum *) NULL) 1142 break; 1143 q=pixels; 1144 for (x=0; x < (ssize_t) image->columns; x++) 1145 { 1146 *q++=ScaleQuantumToChar(GetPixelRed(p)); 1147 *q++=ScaleQuantumToChar(GetPixelGreen(p)); 1148 *q++=ScaleQuantumToChar(GetPixelBlue(p)); 1149 p+=GetPixelChannels(image); 1150 } 1151 (void) WriteBlob(image,(size_t) (q-pixels),pixels); 1152 if ((image->previous == (Image *) NULL) && 1153 (SetImageProgress(image,SaveImageTag,y,image->rows) == MagickFalse)) 1154 break; 1155 } 1156 pixels=(unsigned char *) RelinquishMagickMemory(pixels); 1157 if (GetNextImageInList(image) == (Image *) NULL) 1158 break; 1159 image=SyncNextImageInList(image); 1160 status=SetImageProgress(image,SaveImagesTag,scene, 1161 GetImageListLength(image)); 1162 if (status == MagickFalse) 1163 break; 1164 scene++; 1165 } while (image_info->adjoin != MagickFalse); 1166 (void) CloseBlob(image); 1167 return(MagickTrue); 1168}</pre> 1169 1170<p>To invoke the custom coder from the command line, use these commands:</p> 1171 1172<pre>convert logo: logo.mgk 1173display logo.mgk 1174</pre> 1175 1176<p>We provide the <a href="http://www.imagemagick.org/download/kits/">Magick Coder Kit</a> to help you get started writing your own custom coder.</p> 1177 1178<h2 class="magick-header"><a id="filters"></a>Custom Image Filters</h2> 1179 1180<p>ImageMagick provides a convenient mechanism for adding your own custom image processing algorithms. We call these image filters and they are invoked from the command line with the <a href="command-line-options.html#process">-process</a> option or from the MagickCore API method <a href="api/module.html#ExecuteModuleProcess">ExecuteModuleProcess()</a>.</p> 1181 1182<p>Here is a listing of a sample <a href="../source/analyze.c">custom image filter</a>. It computes a few statistics such as the pixel brightness and saturation mean and standard-deviation.</p> 1183 1184<pre class="pre-scrollable">#include <stdio.h> 1185#include <stdlib.h> 1186#include <string.h> 1187#include <time.h> 1188#include <assert.h> 1189#include <math.h> 1190#include "magick/MagickCore.h" 1191 1192/* 1193%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1194% % 1195% % 1196% % 1197% a n a l y z e I m a g e % 1198% % 1199% % 1200% % 1201%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1202% 1203% analyzeImage() computes the brightness and saturation mean, standard 1204% deviation, kurtosis and skewness and stores these values as attributes 1205% of the image. 1206% 1207% The format of the analyzeImage method is: 1208% 1209% size_t analyzeImage(Image *images,const int argc,char **argv, 1210% ExceptionInfo *exception) 1211% 1212% A description of each parameter follows: 1213% 1214% o image: the address of a structure of type Image. 1215% 1216% o argc: Specifies a pointer to an integer describing the number of 1217% elements in the argument vector. 1218% 1219% o argv: Specifies a pointer to a text array containing the command line 1220% arguments. 1221% 1222% o exception: return any errors or warnings in this structure. 1223% 1224*/ 1225ModuleExport size_t analyzeImage(Image **images,const int argc,const char **argv, 1226 ExceptionInfo *exception) 1227{ 1228 char 1229 text[MaxTextExtent]; 1230 1231 double 1232 area, 1233 brightness, 1234 brightness_mean, 1235 brightness_standard_deviation, 1236 brightness_kurtosis, 1237 brightness_skewness, 1238 brightness_sum_x, 1239 brightness_sum_x2, 1240 brightness_sum_x3, 1241 brightness_sum_x4, 1242 hue, 1243 saturation, 1244 saturation_mean, 1245 saturation_standard_deviation, 1246 saturation_kurtosis, 1247 saturation_skewness, 1248 saturation_sum_x, 1249 saturation_sum_x2, 1250 saturation_sum_x3, 1251 saturation_sum_x4; 1252 1253 Image 1254 *image; 1255 1256 assert(images != (Image **) NULL); 1257 assert(*images != (Image *) NULL); 1258 assert((*images)->signature == MagickSignature); 1259 (void) argc; 1260 (void) argv; 1261 image=(*images); 1262 for ( ; image != (Image *) NULL; image=GetNextImageInList(image)) 1263 { 1264 CacheView 1265 *image_view; 1266 1267 MagickBooleanType 1268 status; 1269 1270 ssize_t 1271 y; 1272 1273 brightness_sum_x=0.0; 1274 brightness_sum_x2=0.0; 1275 brightness_sum_x3=0.0; 1276 brightness_sum_x4=0.0; 1277 brightness_mean=0.0; 1278 brightness_standard_deviation=0.0; 1279 brightness_kurtosis=0.0; 1280 brightness_skewness=0.0; 1281 saturation_sum_x=0.0; 1282 saturation_sum_x2=0.0; 1283 saturation_sum_x3=0.0; 1284 saturation_sum_x4=0.0; 1285 saturation_mean=0.0; 1286 saturation_standard_deviation=0.0; 1287 saturation_kurtosis=0.0; 1288 saturation_skewness=0.0; 1289 area=0.0; 1290 status=MagickTrue; 1291 image_view=AcquireVirtualCacheView(image,exception); 1292#if defined(MAGICKCORE_OPENMP_SUPPORT) 1293 #pragma omp parallel for schedule(dynamic,4) shared(status) 1294#endif 1295 for (y=0; y < (ssize_t) image->rows; y++) 1296 { 1297 register const Quantum 1298 *p; 1299 1300 register ssize_t 1301 x; 1302 1303 if (status == MagickFalse) 1304 continue; 1305 p=GetCacheViewVirtualPixels(image_view,0,y,image->columns,1,exception); 1306 if (p == (const Quantum *) NULL) 1307 { 1308 status=MagickFalse; 1309 continue; 1310 } 1311 for (x=0; x < (ssize_t) image->columns; x++) 1312 { 1313 ConvertRGBToHSB(GetPixelRed(p),GetPixelGreen(p),GetPixelBue(p),&hue,&saturation,&brightness); 1314 brightness*=QuantumRange; 1315 brightness_sum_x+=brightness; 1316 brightness_sum_x2+=brightness*brightness; 1317 brightness_sum_x3+=brightness*brightness*brightness; 1318 brightness_sum_x4+=brightness*brightness*brightness*brightness; 1319 saturation*=QuantumRange; 1320 saturation_sum_x+=saturation; 1321 saturation_sum_x2+=saturation*saturation; 1322 saturation_sum_x3+=saturation*saturation*saturation; 1323 saturation_sum_x4+=saturation*saturation*saturation*saturation; 1324 area++; 1325 p+=GetPixelChannels(image); 1326 } 1327 } 1328 image_view=DestroyCacheView(image_view); 1329 if (area <= 0.0) 1330 break; 1331 brightness_mean=brightness_sum_x/area; 1332 (void) FormatLocaleString(text,MaxTextExtent,"%g",brightness_mean); 1333 (void) SetImageProperty(image,"filter:brightness:mean",text); 1334 brightness_standard_deviation=sqrt(brightness_sum_x2/area-(brightness_sum_x/ 1335 area*brightness_sum_x/area)); 1336 (void) FormatLocaleString(text,MaxTextExtent,"%g", 1337 brightness_standard_deviation); 1338 (void) SetImageProperty(image,"filter:brightness:standard-deviation",text); 1339 if (brightness_standard_deviation != 0) 1340 brightness_kurtosis=(brightness_sum_x4/area-4.0*brightness_mean* 1341 brightness_sum_x3/area+6.0*brightness_mean*brightness_mean* 1342 brightness_sum_x2/area-3.0*brightness_mean*brightness_mean* 1343 brightness_mean*brightness_mean)/(brightness_standard_deviation* 1344 brightness_standard_deviation*brightness_standard_deviation* 1345 brightness_standard_deviation)-3.0; 1346 (void) FormatLocaleString(text,MaxTextExtent,"%g",brightness_kurtosis); 1347 (void) SetImageProperty(image,"filter:brightness:kurtosis",text); 1348 if (brightness_standard_deviation != 0) 1349 brightness_skewness=(brightness_sum_x3/area-3.0*brightness_mean* 1350 brightness_sum_x2/area+2.0*brightness_mean*brightness_mean* 1351 brightness_mean)/(brightness_standard_deviation* 1352 brightness_standard_deviation*brightness_standard_deviation); 1353 (void) FormatLocaleString(text,MaxTextExtent,"%g",brightness_skewness); 1354 (void) SetImageProperty(image,"filter:brightness:skewness",text); 1355 saturation_mean=saturation_sum_x/area; 1356 (void) FormatLocaleString(text,MaxTextExtent,"%g",saturation_mean); 1357 (void) SetImageProperty(image,"filter:saturation:mean",text); 1358 saturation_standard_deviation=sqrt(saturation_sum_x2/area-(saturation_sum_x/ 1359 area*saturation_sum_x/area)); 1360 (void) FormatLocaleString(text,MaxTextExtent,"%g", 1361 saturation_standard_deviation); 1362 (void) SetImageProperty(image,"filter:saturation:standard-deviation",text); 1363 if (saturation_standard_deviation != 0) 1364 saturation_kurtosis=(saturation_sum_x4/area-4.0*saturation_mean* 1365 saturation_sum_x3/area+6.0*saturation_mean*saturation_mean* 1366 saturation_sum_x2/area-3.0*saturation_mean*saturation_mean* 1367 saturation_mean*saturation_mean)/(saturation_standard_deviation* 1368 saturation_standard_deviation*saturation_standard_deviation* 1369 saturation_standard_deviation)-3.0; 1370 (void) FormatLocaleString(text,MaxTextExtent,"%g",saturation_kurtosis); 1371 (void) SetImageProperty(image,"filter:saturation:kurtosis",text); 1372 if (saturation_standard_deviation != 0) 1373 saturation_skewness=(saturation_sum_x3/area-3.0*saturation_mean* 1374 saturation_sum_x2/area+2.0*saturation_mean*saturation_mean* 1375 saturation_mean)/(saturation_standard_deviation* 1376 saturation_standard_deviation*saturation_standard_deviation); 1377 (void) FormatLocaleString(text,MaxTextExtent,"%g",saturation_skewness); 1378 (void) SetImageProperty(image,"filter:saturation:skewness",text); 1379 } 1380 return(MagickImageFilterSignature); 1381}</pre> 1382 1383<p>To invoke the custom filter from the command line, use this command:</p> 1384 1385<pre>convert logo: -process \"analyze\" -verbose info: 1386 Image: logo: 1387 Format: LOGO (ImageMagick Logo) 1388 Class: PseudoClass 1389 Geometry: 640x480 1390 ... 1391 filter:brightness:kurtosis: 8.17947 1392 filter:brightness:mean: 60632.1 1393 filter:brightness:skewness: -2.97118 1394 filter:brightness:standard-deviation: 13742.1 1395 filter:saturation:kurtosis: 4.33554 1396 filter:saturation:mean: 5951.55 1397 filter:saturation:skewness: 2.42848 1398 filter:saturation:standard-deviation: 15575.9 1399</pre> 1400 1401 1402<p>We provide the <a href="http://www.imagemagick.org/download/kits/">Magick Filter Kit</a> to help you get started writing your own custom image filter.</p> 1403 1404</div> 1405 <footer class="magick-footer"> 1406 <p><a href="support.html">Donate</a> • 1407 <a href="sitemap.html">Sitemap</a> • 1408 <a href="links.html">Related</a> • 1409 <a href="architecture.html">Architecture</a> 1410</p> 1411 <p><a href="architecture.html#">Back to top</a> • 1412 <a href="http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x89AB63D48277377A">Public Key</a> • 1413 <a href="http://www.imagemagick.org/script/contact.php">Contact Us</a></p> 1414 <p><small>© 1999-2016 ImageMagick Studio LLC</small></p> 1415 </footer> 1416</div><!-- /.container --> 1417 1418 <script src="https://localhost/ajax/libs/jquery/1.11.3/jquery.min.js"></script> 1419 <script src="../js/magick.html"></script> 1420</div> 1421</body> 1422</html> 1423
