View Single Post
Old 08-12-2014, 10:12 AM   #882
willus
Fuzzball, the purple cat
willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.willus ought to be getting tired of karma fortunes by now.
 
willus's Avatar
 
Posts: 1,303
Karma: 11087488
Join Date: Jun 2011
Location: California
Device: iPad
Quote:
Originally Posted by pmarty View Post
Here are my results with another file.

k2pdfopt -col 1 -cbox 0,0,21cm,26.7cm -odpi 100 -sm 2.pdf

k2pdfopt is apparently confused with the text box in the middle. It detects 3 red regions there which are oddly numbered (58, 99, 99) in the marked source. I tried using -evl 2 but it didn't help.

Any ideas?
Interesting case. Thanks again for the thorough information and attachments. The 58 and 99 regions are not actually "red"--they're orange (the color I use for diagnostics involving finding rows of text), and IIRC the values are related to a row-splitting figure of merit. k2pdfopt initially doesn't split the center box into text rows but then goes back and looks for where it might have missed row splits and this is where the issue is. See the -gtr and -rsf options and maybe try tweaking them. I'm surprised -evl 2 didn't at least change the behavior. I'll investigate further when I have more time.

Last edited by willus; 08-12-2014 at 10:44 AM. Reason: Added link
willus is offline   Reply With Quote