Bug in ImageMagick

This page reports a bug I found in ImageMagick 

Version: @(#)ImageMagick 5.3.5 07/01/01 Q:16 http://www.imagemagick.org
running on Sun Solaris 8. Please do not release this page on the web. The image will be the cover of a Scientific Journal in December 2001.

I have a PostScript image file baseflipcover.ps that I needed to convert to tiff format at 300dpi or higher.

This is the script I used to create the files: 

echo time is about 2 minutes to do both
echo converting ...
echo ... do 300 dpi ...

date

convert \
-density 300 -geometry 24% \
-rotate 90 \
baseflipcover.ps \
baseflipcover300.tiff

echo ... now do 600 dpi ...

convert \
-density 600 -geometry 12% \
-rotate 90 \
baseflipcover.ps \
baseflipcover600.tiff

date

echo ... done

chmod a+r *

The resulting files are:

I gave these to the journal and was told that they were 72dpi! I was able to confirm this using Gimp. I then used Gimp to do the conversion and that worked to produce the final image, baseflipcover.tiff.

Apparently ImageMagick reports a incorrect value using identify -verbose. One can see this in the report, which claims that baseflipcover300.tiff and baseflipcover600.tiff are correct (ie 300 and 600 dpi respectively) whereas they are both really 72 dpi. A clear giveaway that there is a problem is the file sizes: 

-rw-r--r--   1 toms     delila   23192054 Oct 24 13:13 baseflipcover.tiff
-rw-r--r--   1 toms     delila    1337119 Oct 23 15:44 baseflipcover600.tiff
-rw-r--r--   1 toms     delila    1333079 Oct 23 15:43 baseflipcover300.tiff
which shows that the correct image is quite a bit larger!

Let's see if it is as big as one would predict. The baseflipcover300.tiff and baseflipcover600.tiff images are 72 dpi. Per square inch, thats 72*72 = 5184 dots. The baseflipcover.tiff image is 300 dpi. Per squqare inch thats 300*300 = 90000. The ratio is the expansion, 90000 / 5184 = 17.36. So the baseflipcover300.tiff image starting at 1333079 bytes should become 1333079 * 17.36 = 23143733 bytes. PRETTY DARN CLOSE. It is only 23143733 /23192054 = 0.99792 off, less than 1%.

SUMMARY:
convert -density 300 -geometry 24%
did not set the density to 300 when converting from PostScript to tiff format.

Tom 

  Dr. Thomas D. Schneider
  toms@alum.mit.edu
  permanent email: toms@alum.mit.edu
  https://alum.mit.edu/www/toms/

Notes on reporting bugs:

color bar Small icon for Theory of Molecular Machines: physics, chemistry, biology, molecular biology, evolutionary theory, genetic engineering, sequence logos, information theory, electrical engineering, thermodynamics, statistical mechanics, hypersphere packing, gumball machines, Maxwell's Daemon, limits of computers


Schneider Lab
origin: 2001 Oct 28
updated: 2001 Oct 28
color bar