ImageMagick Command-line Processing

In Unix shells, certain characters such as the asterisk (*) and question mark (?) automagically cause lists of filenames to be generated based on pattern matches. This feature is known as globbing. ImageMagick supports filename globbing for systems, such as Windows, that does not natively support it. For example, suppose you want to convert 1.jpg, 2.jpg, 3.jpg, 4.jpg, and 5.jpg in your current directory to a GIF animation. You can conveniently refer to all of the JPEG files with this command:

$magick> convert *.jpg images.gif

Images are stored in a myriad of image formats including the better known JPEG, PNG, TIFF and others. ImageMagick must know the format of the image before it can be read and processed. Most formats have a signature within the image that uniquely identifies the format. Failing that, ImageMagick leverages the filename extension to determine the format. For example, image.jpg or image.JPG tells ImageMagick it is reading an image in the JPEG format.

In some cases the image may not contain a signature and/or the filename does not identify the image format. In these cases an explicit image format must be specified. For example, suppose our image is named image and contains raw red, green, and blue intensity values. ImageMagick has no way to automagically determine the image format so we explicitly set one:

$magick> convert -size 640x480 -depth 8 rgb:image image.png

ImageMagick has a number of built-in images and patterns. To utilize the checkerboard pattern, for example, use:

$magick> convert -size 640x480 pattern:checkerboard checkerboard.png

Unix and Windows permit the output of one command to be piped to the input of another. ImageMagick permits image data to be read and written from the standard streams STDIN (standard in) and STDOUT (standard out), respectively, using a pseudo-filename of -. In this example we pipe the output of convert to the display program:

$magick> convert logo: gif:- | display gif:-

The second explicit format "gif:" is optional in the preceding example. The GIF image format has a unique signature within the image so ImageMagick's display command can readily recognize the format as GIF. The convert program also accepts STDIN as input in this way:

$magick> convert rose: gif:- | convert - -resize "200%" bigrose.jpg

Other pipes can be accessed via their file descriptors (as of version 6.4.9-3). The file descriptors 0, 1, and 2 are reserved for the standard streams STDIN, STDOUT, and STDERR, respectively, but a pipe associated with a file descriptor number N>2 can be accessed using the pseudonym fd:N. (The pseudonyms fd:0 and fd:1 can be used for STDIN and STDOUT.) The next example shows how to append image data piped from files with descriptors 3 and 4 and direct the result to the file with descriptor number 5.

$magick> convert fd:3 fd:4 -append fd:5

When needed, explicit image formats can be given as mentioned earlier, as in the following.

$magick> convert gif:fd:3 jpg:fd:4 -append tif:fd:5

Some images formats contain more than one image frame. Perhaps you only want the first image, or the last, or some number of images in-between. You can specify which image frames to read by appending the image filename with the frame range enclosed in brackets. Here our image (an animated GIF) contains more than one frame but we only want the first:

$magick> convert 'images.gif[0]' image.png

[Unix shells generally interpret brackets so we enclosed the filename in quotes above. In a Windows command shell the brackets are not interpreted but using quotes doesn't hurt. However, in most cases the roles of single-quotes and double-quotes are reversed with respect to Unix and Windows, so Windows users should usually try double-quotes where we display single-quotes, and vice versa.]

You can read more than one image from a sequence with a frame range. For example, you can extract the first four frames of an image sequence:

$magick> convert 'images.gif[0-3]' images.mng

Finally, you can read more than one image from a sequence, out-of-order. The next command gets the third image in the sequence, followed by the second, and then the fourth:

$magick> convert 'images.gif[3,2,4]' images.mng

Notice that in the last two commands, a single image is written. The output in this case, where the image type is MNG, is a multi-frame file because the MNG format supports multiple frames. Had the output format been JPG, which only supports single frames, the output would have consisted of separate frames. More about that below, in the section about the Output Filename.

Raw images are a sequence of color intensities without additional meta information such as width, height, or image signature. With raw image formats, you must specify the image width and height but you can also specify a region of the image to read. In our example, the image is in the raw 8-bit RGB format and is 6000 pixels wide and 4000 pixels high. However, we only want a region of 600 by 400 near the center of the image:

$magick> convert -size 6000x4000 -depth 8 \
'rgb:image[600x400+1900+2900]' image.jpg

You can get the same results with the ‑extract option:

$magick> convert -size 6000x4000 -depth 8 \
-extract 600x400+1900+2900 rgb:image image.jpg

It is sometimes convenient to resize an image as they are read. Suppose you have hundreds of large JPEG images you want to convert to a sequence of PNG thumbails:

$magick> convert '*.jpg' -resize 120x120 thumbnail%03d.png

Here all the images are read and subsequently resized. It is faster and less resource intensive to resize each image it is read:

$magick> convert '*.jpg[120x120]' thumbnail%03d.png

It is sometimes convenient to crop an image as they are read. Suppose you have hundreds of large JPEG images you want to convert to a sequence of PNG thumbails:

$magick> convert '*.jpg' -crop 120x120+10+5 thumbnail%03d.png

Here all the images are read and subsequently cropped. It is faster and less resource-intensive to crop each image as it is read:

$magick> convert '*.jpg[120x120+10+5]' thumbnail%03d.png

There are two methods to use a filename to reference other image filenames. The first is with '@' which reads image filenames separated by white space from the specified file. Assume the file myimages.txt consists of a list of filenames, like so:

  frame001.jpg
  frame002.jpg
  frame003.jpg

We then expect this command:

$magick> convert @myimages.txt mymovie.gif

to read the images frame001.jpg, frame002.jpg, and frame003.jpg and convert them to a GIF image sequence.

If the image path includes one or more spaces, enclose the path in quotes:

  'my title.jpg'

Some ImageMagick command-line options may exceed the capabilities of your command line processor. Windows, for example, limits command lines to 8192 characters. If, for example, you have a draw option with polygon points that exceed the command-line length limit, put the draw option instead in a file and reference the file with the @ (e.g. @mypoly.txt).

Another method of referring to other image files is by embedding a formatting character in the filename with a scene range. Consider the filename image-%d.jpg[1-5]. The command

$magick> convert image-%d.jpg[1-5]

causes ImageMagick to attempt to read images with these filenames:

  image-1.jpg
  image-2.jpg
  image-3.jpg
  image-4.jpg
  image-5.jpg

Here, just below, are a few simple examples of geometry, showing how it might be used as an argument to the ‑resize option. We'll use the internal image logo: for our input image. This fine image is 640 pixels wide and 480 pixels high. We say its dimensions are 640x480. When we give dimensions of an image, the width (the horizontal dimension) always precedes the height (the vertical dimension). This will be true when we speak of coordinates or offsets into an image, which will always be x–value followed by y. Just think of your high school algebra classes and the xy–plane. (Well, almost: our y–axis will always go downward!)

$magick> convert logo: -resize '200%' bigWiz.png

$magick> convert logo: -resize '200x50%' longShortWiz.png

$magick> convert logo: -resize '100x200' notThinWiz.png

$magick> convert logo: -resize '100x200^' biggerNotThinWiz.png

$magick> convert logo: -resize '100x200!' dochThinWiz.png

The first of the four commands is simple—it stretches both the width and height of the input image by 200% in each direction; it magnifies the whole thing by a factor of two. The second command specifies different percentages for each direction, stretching the width to 200% and squashing the height to 50%. The resulting image (in this example) has dimensions 1280x240. Notice that the percent symbol needn't be repeated; the following are equivalent: 200x50%, 200%x50, 200%x50%.

By default, the width and height given in a geometry argument are maximum values unless a percentage is specified. That is, the image is expanded or contracted to fit the specified width and height value while maintaining the aspect ratio (the ratio of its height to its width) of the image. For instance, the third command above "tries" to set the dimensions to 100x200. Imagine gradually shrinking the original image (which is 640x480), keeping is aspect ratio constant, until it just fits into a 100x200 rectangle. Since the image is longer than it is tall, it will fit when its width shrinks to 100 pixels. To preserve the aspect ratio, the height will therefore have to be (480/640)×100 pixels=75 pixels, so the final dimensions will be 100x75.

Notice that in the previous example, at least one of the specified dimensions will be attained (in this case, the width, 100 pixels). The resulting image fits snugly within the original. One can do just the opposite of this by invoking the ^ operator, as in the fourth example above. In that case, when 100x200^ is given as the argument, again at least one of the dimensions will be attained, but in this case the resulting image can snugly contain the original. Here the geometry argument gives minimum values. In our example, the height will become 200 and the width will be scaled to preserve the aspect ratio, becoming (640/480)×200 pixels=267 pixels. With the ^ operator, one of those dimensions will match the requested size, but the image will likely overflow the dimensions requested to preserve its aspect ratio. (The ^ feature is new as of IM 6.3.8-2.)

We see that ImageMagick is very good about preserving aspect ratios of images, to prevent distortion of your favorite photos and images. But you might really want the dimensions to be 100x200, thereby stretching the image. In this case just tell ImageMagick you really mean it (!) by appending an exclamation operator to the geometry. This will force the image size to exactly what you specify. So, for example, if you specify 100x200! the dimensions will become exactly 100x200 (giving a small, vertically elongated wizard).

Here are a few more examples:

$magick> convert logo: -resize '100' wiz1.png

$magick> convert logo: -resize 'x200' wiz2.png

$magick> convert logo: -resize '100x200>' wiz3.png

$magick> convert logo: -resize '100x200<' wiz4.png

If only one dimension is given it is taken to be the width. When only the width is specified, as in the first example above, the width is accepted as given and the height is chosen to maintain the aspect ratio of the input image. Similarly, if only the height is specified, as in the second example above, the height is accepted and the width is chosen to maintain the aspect ratio.

Use > to change the dimensions of the image only if the original width or height is exeeded by one of the geometry specifications. Use < to resize the image only if the original dimensions are more than the geometry specifications. In either case, if a change is made, the result is as if the > or > operator was not present. So, in the third example above, we specified 100x200> and the original image size is 640x480, so the image size is reduced as if we had specified 100x200. However, in the fourth example above, there will be no change to its size.

Finally, use @ to specify the maximum area in pixels of an image, again while attempting to preserve aspect ratio. (Pixels take only integer values, so some approximation is always at work.) In the following example, an area of 10000 pixels is requested. The resulting file has dimensions 115x86, which has 9890 pixels.

$magick> convert logo: -resize '@10000' wiz10000.png

Another word about the quotation marks: In all the examples above and below, we have enclosed the geometry arguments within quotation marks. Doing so is optional in many cases, but not always. We must enclose the geometry specifications in quotation marks when using < or > to prevent these characters from being interpreted by the shell as file redirection. On Windows systems, the carat ^  needs to be within quotes, else it is ignored. To be safe, one should probably maintain a habit of enclosing all geometry arguments in quotes, as we have here.

Here are some examples to illustrate the use of offsets in geometry arguments. One typical use of offsets is in conjunction with the ‑region option. This option allows many other options to modify the pixels within a specified rectangular subregion of an image. As such, it needs to be given the width and height of that region, and also an offset into the image, which is a pair of coordinates that indicate the location of the region within the larger image. Below, in the first example, we specify a region of size 100x200 to be located at the xy–coordinates x=10, y=20. Let's use the usual algebraic notation (x,y)=(10,20), for convenience.

$magick> convert logo: -region '100x200+10+20' -negate wizNeg1.png

$magick> convert logo: -region '100x200-10+20' -negate wizNeg2.png

$magick> convert logo: -gravity center -region '100x200-10+20' \
-negate wizNeg3.png

Note that offsets always require +/− signs. The offset is not actually a true location within the image; its coordinates must be added to some other location. Let's refer to that as the current location. In the first two examples above, though, that location is the upper-left hand corner of the image, which has coordinates (0,0). (That is the default situation when there are no other directives given to change it.) The first example above puts the 100x200 rectangle's own upper-left corner at (10,20).

A negative offset can make sense in many cases. In the second example above, the offset is (-10,20), specified by -10+20. In that case, only the portion of the (virtual) rectangle obtained that lies within the image can be negated; here it is equivalent to specifying the geometry as 90x200+0+20.

In the third example above, the ‑gravity setting precedes the others and sets the current location within the image at the very center of the image. In this case that is at pixel (320,240), since the size of the image is 640x480. This means that the offsets apply to that location, which thereby gets moved, in this case, to (320-10,240+20)=(310,260). But the 100x200 region itself is affected by the ‑gravity setting, so instead of affecting its upper-left corner, the region's own center (at (+50,+100) within it) is determined. Therefore the center of the 100x200 rectangle is moved to (310,260). The negated rectangle's upper-left corner is now at (310-50,260-100)=(260,160).

Images can be stored in a mryiad of image formats including the better known JPEG, PNG, TIFF and others. ImageMagick must know the desired format of the image before it is written. ImageMagick leverages the filename extension to determine the format. For example, image.jpg tells ImageMagick to write the image in the JPEG format. In some cases the filename does not identify the image format. In these cases, the image is written in the format it was originally read unless an explicit image format is specified. For example, suppose we want to write our image to a filename of image in the raw red, green, and blue intensity format:

$magick> convert image.jpg rgb:image

Unix permits the output of one command to be piped to another. ImageMagick permits piping one command to another with a filename of -. In this example we pipe the output of convert to the display program:

$magick> convert logo: gif:- | display gif:-

Here the explicit format is optional. The GIF image format has a signature that uniquely identifies it so ImageMagick can readily recognize the format as GIF.

Optionally, use an embedded formatting character to write a sequential image list. Suppose our output filename is image-%d.jpg and our image list includes 3 images. You can expect these images files to be written:

  image-0.jpg
  image-1.jpg
  image-2.jpg

Or retrieve image properties to modify the image filename. For example, the command

$magick> convert rose: -set filename:area '%wx%h' \
'rose-%[filename:area].png'

writes an image with this filename:

  rose-70x46.png

Finally to convert multiple JPEG images to individual PDF pages, use:

  convert *.jpg +adjoin page-%d.pdf