Hi there, 

I'm in a process of doing some png crushing on a whole bunch of images. To be 
extra sure the crushing is lossless, I thought I could run compare on the 
original versus optimized. That works great in most cases, except I am hitting 
a few false negatives, could anyone confirm with me if it is a bug or if I am 
doing something wrong?

Two such images are here (very small 12x12 icons):
http://timotheegroleau.com/tmp/external.png (~46kb)
http://timotheegroleau.com/tmp/external-crushed.png (114b)

To check that my images are identical or not, I run compare (from IM-6.5.8-8 
under gentoo) like so:
compare -dissimilarity-threshold 1 -metric AE original.png crushed.png 
/dev/null 2>&1

and read the result. If it is 0, images are identical; not zero, some pixels 
differ.

For the 2 pics above, I get a result of 84, but as far as I can tell both 
images are identical. The difference image that compare generates shows a 
whole bunch of pixels supposedly different, but I just cannot see them in the 
original versus optimized (I'm using gimp to zoom in and check).

The original png is grossly uncompressed, the crushed image is super optimized 
with a 2-color palette.

The false negative makes me wonder if there are false positives too, so I am 
worried on the reliability of my validation with compare. Any help would be 
much appreciated.

Many thanks!
Tim.
_______________________________________________
Magick-users mailing list
[email protected]
http://studio.imagemagick.org/mailman/listinfo/magick-users
  • ... Timothee Groleau
    • ... da . mihi . sis . bubulae . frustrum . assae . solana . tuberosa . in . modo . gallico . fricta . ac . quassum . lactatum . coagulatum . crassum
      • ... Timothee Groleau
    • ... Fred Weinhaus
      • ... Timothee Groleau
    • ... Fred Weinhaus
    • ... Fred Weinhaus

Reply via email to