>GraphicsMagick FAQ

  1. I received the following message, "?????? delegation failed ...". What does it mean?

  2. How can I improve the appearance of 24-bit images displayed on my colormapped X server?

  3. How do I set the transparency index in a GIF image so it displays properly within Mozilla?

  4. How can I stop the filenames from changing in the title bar of the animate(1) image window?

  5. The image grabbed by import(1) does not look like the image on my X server. What's wrong

  6. How do I animate a digital YUV image sequence?

  7. How do I view only the red channel of an RGB image?

  8. How do I change the default PostScript page size?

  9. When I display or convert an image, I get Memory allocation error. What can I do?

  10. How do I concatenate three images left-to-right with no borders, frames, or text?

  11. How do I create a GIF animation sequence to display within Netscape?

  12. When I display a PostScript image it appears to be trimmed. I would like to display the entire page, including any white space.

  13. What are visual image directories? How do I use them?

  14. I have an HP. How do I take advantage of the Color Recovery option?

  15. I use the window ID reported by xwininfo(1) with import and it does include the window manager frame as expected. How can I save the window with its frame?

  16. I displayed an image and it appears as one solid color. What did I do wrong?

  17. I received the following message, "???? library is not available ...". What does it mean?

  18. I want to inspect the values of the matte channel within my image. How can I do that?

  19. Convert in.gif out.png" says "Invalid Parameter - out.png"

  20. How can I add one of those cool bevels to my image that I see used on the Web?

  21. I try to launch display from my window manager and it fails. What's up?

  22. When I convert Postscript to another format, how can I improve the appearance of the text?

  23. How can I annotate an image with text that is 2 to 3 inches tall?

  24. How can I convert my GIF animation sequence to individual image files?

  25. How can I remove the background that prints around my image when I display it with Netscape?

  26. How do I create a GIF or PNG image with Web safe colors?

  27. How come Adobe Acrobat 2.1 can't read GraphicsMagick's PDF format?

  28. How can I add a matte layer to my image?

  29. How can I draw with text using 'convert' under Windows NT?

  30. Why are my GIF files larger than expected?

  31. Why are my JPEG files larger than expected?

  32. How do I extract a single image from a multi-image file?



1. I received the following message, "?????? delegation failed ...". What does it mean?
GraphicsMagick uses several freely available packages to perform the translation of certain image formats (PostScript, MPEG, etc.). Make sure these packages are available as described in the README file. Also verify that you have plenty of temporary disk space. If not, set the TMPDIR environment variable to an area where space is available. Finally, for PostScript, verify that Ghostscript supports the pnmraw or ppmraw device (gs -h) and that the document contains valid PostScript statements (gs image.ps).

2. How can I improve the appearance of 24-bit images displayed on my colormapped X server?

If you have a colormapped X server, the number of unique colors in an image must first be reduced to match the requirements of your server before it can be displayed. Display tries to make the color reduced image closely resemble the original. However, there are some options that may improve the appearance.

You can use a private colormap. By default, your image must share colors with existing applications on your X server. The colors in your image may be very different than the shared colors and the results may be visually unsatisfactory. Try

 display -colormap private image.miff

The disadvantage is the annoying colormap loading and unloading as you move your pointer in and out of the image window.

The YCbCr colorspace may give better results when color reducing images of people faces. Try

 display -colorspace YCbCr image.miff

If you intend to view an image frequently, reduce the number of colors to match the resolution of your X server once and write to a new image. Then display this new image. For example,

 convert -colors 256 -dither image.jpeg image.miff
 display image.miff

This forces the time-consuming process of color reducing the image to one-time, rather than each time the image is displayed.


3. How do I set the transparency index in a GIF image so it displays properly within Mozilla?
Display your GIF image with display. Choose Matte from the Image Edit command menu and identify a pixel that has the transparency color. Press Method and select a matte edit method from a pop-up menu. Choose from these methods:
 point
 replace
 floodfill

The point method changes the matte value of any pixel selected with the pointer until the button is is released. The replace method changes the matte value of any pixel that matches the color of the pixel you select with a button press. Floodfill changes the matte value of any pixel that matches the color of the pixel you select with a button press and is a neighbor.

Select your transparent pixel with the pointer and press a button. The image is redisplayed with any transparent pixels recolored to the background color. You can select other pixels or areas to force to transparent. When you are satisfied, press Return.

Finally, choose Save from the command menu and write your GIF image to a file. Note that setting transparency works best on a TrueColor or DirectColor visual. If your server only exports colormapped visuals you will need to use a Standard Colormap to set transparency.

 xstdcmap -best
 display -map list image.gif
 choose Matte Edit
 select your transparent pixel then press Return
 choose Save

If you do not have the xstdcmap(1) program, try

 display -visual TrueColor image.gif

4. How can I stop the filenames from changing in the title bar of the animate(1) image window?
Animate updates the image file name in the title bar of the image window as each image is displayed from the image sequence. To display just a single name that will not change, use -title:
 animate -title "My Image Sequence" images.

5. The image grabbed by import(1) does not look like the image on my X server. What's wrong?
Use the -descend option:
 import -descend image.miff

or set this X resource:

 import.descend: True

By default, import quickly grabs the image from the X server. However, it may not always have the correct colors in some areas. This can happen when a subwindow has a different colormap than its parent. With -descend, import descends the window hierarchy. Descending involves grabbing the image and colormap of each window or subwindow associated with the window you select and compositing it on a blank canvas. This can be significantly slower than just grabbing the top-level window but ensures the correct image.


6. How do I animate a digital YUV image sequence?
Suppose your sequence is 72 352x240 frames titled frame0.Y, frame0.U, frame0.V, frame1.Y, frame1.U, etc. Use this command:
 animate -geometry 352x240 -scene 0-71 yuv3:frame%d

7. How do I view only the red channel of an RGB image?
Use the -gamma option, for example,
 display -gamma 1.0,0.0,0.0 image.miff

8. How do I change the default PostScript page size?
The default dimensions of a PostScript page is 612x792. If you prefer another default, change the page geometries (PSPageGeometry) in magick/image.h and recompile.
9. When I display or convert an image, I get Memory allocation error. What can I do?
Memory allocation is a complex topic in GraphicsMagick and image processing requires a lot of memory. GraphicsMagick tries to take best advantage of the resources available by optimizing its use of virtual memory. Virtual memory is normally the amount of RAM (Random Access Memory) available to the process plus the amount of free space in the system paging area (known as "swap" under Unix, or a "page file" under Windows). In addition to the RAM and the paging area, GraphicsMagick is able to expand the amount of virtual memory available by using memory-mapped files which cause the file to be treated as extra memory. A piece of hardware known as the Memory Management Unit (MMU) performs the magic which allows the system paging area and memory-mapped files to be treated as more memory by your system's CPU.
Usually when a memory allocation error occurs, it is because the system's paging area is full, or the programs memory (heap) is fragmented badly enough that there is no large-enough block of memory available to satisfy the request. In some cases the operating system may artificially limit the memory available to the program.
Failures of small memory allocations are generally considered fatal by GraphicsMagick. If it is impossible to allocate even a little bit of memory, then there is no point in continuing. When large memory allocations associated with allocating memory for image pixels fail, GraphicsMagick uses that as a key to know that it should start using memory-mapped temporary files to increase virtual memory. Under Unix and other POSIX-compliant systems, these memory-mapped temporary files are created in either the system default location (e.g. /var/tmp) or the directory set by the TMPDIR environment variable. Under Windows, temporary files are created in the directory set by the TEMP or TMP environment variable.
If the temporary file directory is too small, or is itself mapped to the system's paging area (no win!), then GraphicsMagick will fail to allocate more virtual memory via the temporary file and will return an error.
The single biggest factor in how much memory is required by GraphicsMagick is the QuantumDepth setting when it was compiled. The amount of memory (in bytes) required to store a single image in memory is may be calculated via the equation (QuantumDepth*Rows*Columns*5)/8. As a means of example, the following table shows the amount of memory consumed by a single 1024x768 image using the supported QuantumDepth settings:

QuantumDepth

Virtual Memory

8

3MB

16

8MB

32

15MB

Performing an image processing operation may require that several images be in memory at one time. In the case of animations, hundreds of images may be in memory at one time.

GraphicsMagick is designed to be general purpose. It can display many image storage formats (Monochrome, PseudoColor, or TrueColor) on many different types of X visuals (StaticGray, StaticColor, PseudoColor, GrayScale, DirectColor, or TrueColor). To support all these combinations of image storage formats and X visuals, extra memory is required. Additionally, animate and montage store an entire image sequence in memory.

It is recommended that systems used to run GraphicsMagick have at least 96MB of RAM and 140MB free in their paging area. Systems used to process large images (do the math!) may require considerably more memory to operate efficiently. RAM is typically 1000 times faster than disk, so it is important to keep as much data in RAM as possible (buy lots of RAM). The temporary file area should have plenty of free space. The size of the temporary file area is usually the deciding factor as to whether GraphicsMagick is able to complete an operation. With sufficient free disk space, GraphicsMagick is primarily limited by your CPU and operating system's ability to address memory. GraphicsMagick has been used to process images 20K by 20K pixels in size!

10. How do I concatenate three images left-to-right with no borders, frames, or text?
Assume your three images are called image1.ppm, image2.ppm, and image3.ppm. Type
 montage -mode concatenate -tile 3x1 image1.ppm image2.ppm \
 image3.ppm concatenated.miff

To concatenate the images top-to-bottom, use -tile 1x3.

For more control over the placement of an image, use composite. First create a matte image and position your images onto the matte. For example,

 convert -size 350x500 xc:black composite.miff
 composite -geometry +0+0 composite.miff image1.gif composite.miff
 composite -geometry "+1"00+0 composite.miff image2.gif composite.miff
 composite -geometry +0+300 composite.miff image3.gif composite.miff
 composite -geometry +0+375 composite.miff image4.gif composite.miff

11. How do I create a GIF animation sequence to display within Netscape?
Use convert with the -delay and -page options. The -delay option is used to specify the delay in 1/100ths of a second between the display of each frame of the animation. For example,
 convert -delay 20 frame*.gif animation.gif

You can also declare specific delays for each frame of the image sequence. For example, if the delay was 20, 10, and 5, use

 convert -delay 20 frame1.gif -delay 10 frame2.gif \
 -delay 5 frame3.gif animation.gif

Use -page to specify the left and top locations of the image frame:

 convert frame1.gif -page +50"+1"00 frame2.gif -page +0"+1"00 \
 frame3.gif animation.gif

Finally, if you want the image to loop within Netscape, use -loop:

 convert -loop 50 frame*.gif animation.gif

For further information about GIF animation, see GIF Animation on the WWW.

Note, that all the images are composited into a single multi-image GIF animation. If you want a single image produced for each frame, use +adjoin:

 convert +adjoin images.* frames%d.gif

12. When I display a PostScript image it appears to be trimmed. I would like to display the entire page, including any white space.
GraphicsMagick automatically trims any PostScript image as defined by the bounding box. To preempt this behavior, remove the bounding box statement from the Postscrpt or explicitly set the page size. For example,
 display -page letter image.ps

13. What are visual image directories? How do I use them?
A visual image directory (VID) is an image that contains thumbnails of one or more images in a file directory. Rather than displaying each individual image at its full resolution, you can browse the visual image directory and choose an image to display. You can create a VID with either of these commands:
 montage *.jpg directory.vid
 convert 'vid:*.jpg' directory.vid

Of course you can substitute any filenames you desire. Montage has many relevant command line options. You can exercise more control over the appearance of the VID than with convert.

Next display the directory:

 display directory.vid

Finally browse and select an image to display. Move the pointer to the image and press button 3.

You can create the VID directory with this command:

 display 'vid:*.jpg'

You can also select Visual Image... from the File menu of the command widget.

Note, that creating a VID is time consuming. Creating them on-the-fly within display may be less convenient than using montage or convert . Also, if you create them with montage. or convert, you can reuse them as often as necessary.

Note that a visual image directory is useful for looking at individual frames of an image sequence:

 display vid:movie.mpg

14. I have an HP. How do I take advantage of the Color Recovery option?
Type
 display -map list image.miff

If you want to use Color Recovery all the time, put this in your X defaults file:

 display.map: list

Now if you need to turn off Color Recovery temporarily, type

 display +dither image.miff

15. I use the window ID reported by xwininfo(1) with import and it does include the window manager frame as expected. How can I save the window with its frame?
By default, xwininfo(1) returns the ID of the window you click on. Use the -frame option to get the reparented window ID:
 xwininfo -frame

You can then use the returned window ID with import:

 import -frame -window ID window.miff

16. I displayed an image and it appears as one solid color. What did I do wrong?
A blank image generally means that the image is either corrupt or it has a matte channel and the matte values are all zero. GraphicsMagick treats a matte value of zero as completely transparent. To determine if this is the problem, try
 display +matte image.miff

17. I received the following message, "???? library is not available ...". What does it mean?
GraphicsMagick requires source libraries not included with the distribution to view or convert certain image formats such as JPEG or TIFF. The above message means you did not compile the required library and link with the GraphicsMagick utilities. See README for the location of these libraries and compiling instructions.

18. I want to inspect the values of the matte channel within my image. How can I do that?
View the matte image as a gray scale image. Suppose you have a TIFF image that has a matte channel and is 640 pixels in width and 480 in height. Type:
 convert image.tiff image.matte
 display -size 640x480 gray:image.matte

19. "Convert in.gif out.png" says "Invalid Parameter - out.png"
You have inadvertently run the Microsoft Windows program called convert.exe, rather than the GraphicsMagick program of the same name. The Windows convert.exe program is used to convert a system boot disk from MS-DOS "FAT" format to NTFS format. As such it can only be used one time in the life of a system.
You can

20. How can I add one of those cool bevels to my image that I see used on the Web?
There are four types of ornamental borders you can add to your image with GraphicsMagick. Each is listed below with the procedure to use them with your image.

21. I try to launch display from my window manager and it fails. What's up?
Display determines if it is executing interactively and behaves differently depending on the result. To convince display you are running in an interactive environment when launching from a window manager, use either of
 display logo:Untitled
 display < /dev/console

22. When I convert Postscript to another format, how can I improve the appearance of the text?
Simple. Increase the dots-per-inch when converting and sub-sample:
 convert -density 288 -geometry 25% image.ps image.gif

Change the density to 144 and geometry to 50% if the above command fails due to insufficient memory. Alternatively, see the Ghostscript documentation about using high-quality fonts.

The -density option increases the number of pixels (or dots) generated by Ghostscript when processing the input postscript file. However as all other images formats are generally displayed on screens which are typically about 72 to 100 dots per inch, the output image will be larger.

The -geometry option reduces the large image output of ghostscript image back to a normal 72 dpi resolution (25% of 288 dpi gives 72 dpi) but in the process anti-aliases (or smooths) the fonts and lines of the image so as to remove the jaggies you would otherwise get from a normal postscript to image conversion.


23. How can I annotate an image with text that is 2 to 3 inches tall?
If you do not access to a particular named font that is large, try scalable fonts. First see if you have any scalable fonts. Type
 xlsfonts -fn '*-0-0-0-0-*'
Or if you are using display, use the font pattern above within the Font Browser (see Image Edit->Annotate). Next substitute the appropriate resolution. Keep in mind that a scalable font must be fully qualified to work. That is, all 14 fields must be specified. Here is one example where we annotate an image with large Helvetica text:
 convert -font '-*-helvetica-*-*-*--300-300-*-*-*-*-iso8859-1' \
 -fill green -draw 'text 50,300 Magick' image.gif annotated.gif

If you have the FreeType support built into GraphicsMagick, just increase your pointsize and/or density:

 convert -font Helvetica -pointsize 100 -density 300 ...

24. How can I convert my GIF animation sequence to individual image files?
Use the scene embedded file format with convert:
    convert animation.gif frame%02d.gif
The resulting image files are titled frame01.gif, frame02.gif, frame03.gif, etc.

25. How can I remove the background that prints around my image when I display it with Netscape?
Use the +page option of the convert command:
    convert +page alpha.gif beta.gif
GIF allows for a page offset relative to some background. The page offset information may have been in your GIF image already or it could have been introduced by GraphicsMagick. Either way, +page removes the unwanted page offset and Netscape should behave as expected.

26. How do I create a GIF or PNG image with Web safe colors?
Use the -map option of the convert command:
    convert -map netscape: alpha.gif beta.gif
Netscape predefines 216 colors for colormapped displays. Use the above command to ensure only these predefined colors are used. Otherwise Netscape dithers your image with varying degrees of image fidelity.

27. How come Adobe Acrobat 2.1 can't read GraphicsMagick's PDF format?
The default PDF compression is Zip. You need Acrobat 3.0 and above to read Zip compressed PDF. Instead use no compression or LZW compression when you create the PDF file:
    convert +compress images.tiff image.pdf

28. How can I add a matte layer to my image?
One way is to use a bitmap as your transparency mask
First, use the -matte option to add an all-opaque opacity channel, then use the composite utility to copy the graylevel samples from the mask file into your new opacity channel:
      convert image.gif -matte temp.miff
      composite -compose CopyOpacity mask.xbm temp.miff transparent.gif
Note, GIF is limited to one transparent color. If your mask has variable opacity, use a format like MIFF, TIFF, or PNG as your output image format.

29. How can I draw with text using 'convert' under Windows NT?
The problem is that NT interprets the command line differently than Unix does, causing the documented command to fail. The following command has been reported to work correctly (all on one line):
      convert -font Arial -fill blue -draw "text 10,10 'your text here'" d:\test.tif png:d:\test.png
and here is another example which is reported to work (which relies on Ghostscript's 'gs' program to installed):
      convert.exe -pointsize 18 -draw "text 0,0 "This is my text!"" C:\blank.gif c:\text.gif

30. Why are my GIF files larger than expected?
LZW compression is disabled by default for GIF (also TIFF) formats, because of the UniSys patent. You can rebuild GraphicsMagick with it enabled, after you have obtained a license from UniSys. Or you can use a third-party commercial UniSys-licensed application to recompress them.

31. Why are my JPEG files larger than expected?
Your JPEG files may contain embedded "profiles" such as Exif or IPTC, or they may contain uncompressed thumbnails. You can use the +profile "*" commandline option or an equivalent API method to remove them.

32. How do I extract a single image from a multi-image file?
Use a square-bracket syntax to indicate which frame or frames you want. For example,
 convert "Image.gif[0]" first.gif

Will extract the first image (scene 0) from a GIF animation. Be sure to surround the file specification with quotation marks, to prevent the shell from interpreting the square brackets.


Copyright © GraphicsMagick Group 2002, 2003, 2004