This is similar to the article http://root.cern.ch/phpBB2/viewtopic.php?t=2780&start=0&postdays=0&postorder=asc, but the problem seems to still exist.
Please see the attached script and images (test.ps is an EPS file). The plotted marker sizes relative to the histogram frame are obviously different between
the EPS and GIF images generated from the script. The EPS file is also strangely shrunk relative to the original canvas size.
I am using ROOT 5.25/01 on a Fedora-based Linux.
I would appreciate any help. test.ps (7.74 KB)
I disagree with you. For me the marker sizes are very similar. Realize that PS and EPS outputs are not a screen dump like gif output is (in interactive mode) but there are completely independent streams. The coordinates systems are different, one is pixel oriented and the other is vector, etc … We try to be as close as possible but tiny differences are still possible and in that particular case I do not think we can do better.
Okay, then I will opt to convert GIF to EPS.
It seems that the difference between GIF and EPS becomes obvious when graphical objects are drawn on a sub-pad.
Even with subpads it is fine seems to me. Look the attached plot. Specially the blue marker which is on the axis. You can better guess its size compared to the axis’ tick marks. c.ps (13.5 KB)
Thank you for your example, but “c.ps” seems a wrong file…
By the way, the difference in the first example still looks obvious for me.
Although the size of the EPS image is much smaller than the GIF image, the marker sizes look almost the same or bigger in the EPS image.
I agree that the difference in your example is not serious although the markers in the EPS image are little bigger.
I attached a bit complicated images as the third example.
If you could print out them, I am sure that you will understand that the difference is nontrivial.
Both images were generated from the same canvas, but the EPS one is not suitable for printing as it is.
