1

I have a 100-page PDF master file exported from LuaLaTeX. Several of the pages need to be trivially edited in Inkscape. So I separated them from the master file and opened them one by one in Inkscape. But, in Inkscape itself — and also when I export them back to PDF — the font of these separate pages is ruined: on my computer, the pages' font looks thinned; and, in print, the very same pages' font looks bloated and lower quality. (The unedited pages from the original 100-page master file look in print as they should.)

Why is this happening? From Inkscape, I've tried exporting them at 1000, 600, 96, etc., DPI, yet the problem persists.

In Inkscape, I'm importing the pages via Poppler/cairo (no fiddling with these settings helps).

Here's an example of the font of the digital master file:

enter image description here

Here's an example of the font of the digital "thinned" Inkscape file (please ignore the size difference; notice the weight difference):

enter image description here

Here's an example of the printed master file:

enter image description here

Here's an example of the printed Inkscape file (please ignore the size difference; notice the weight difference):

enter image description here

6
  • 1
    The problem is known. Poppler-Cairo import converts texts to vector paths and that removes all possibilities to utilize hinting (included to font files) to place letters optimally to the pixel grid in displays and in printing. Rendering engines make their own optimization attempts with varying results. See this graphicdesign.stackexchange.com/questions/153869/…
    – user82991
    Jul 21, 2022 at 15:53
  • Okay. Is there a way in Inkscape, after Poppler-Cairo importation, to convert text that has been converted to paths, back to text? Whenever I use Internal import, my entire file is messed up: all the letters become, essentially, "hyroglyphics."
    – Lijishe
    Jul 21, 2022 at 16:57
  • No practical and reliable way to convert paths automatically back to text is included to Inkscape nor other common graphics programs. Some experimental projects may exist, but I do not know them. Rasterized text can be interpreted by OCR but that's not 100% reliable. proofreading is a must. OCR can be very succesful if one has text in English and with no typographic specialities. But to use it you must practically recreate the layout. OCR gives the text content, not the original look.
    – user82991
    Jul 21, 2022 at 17:10
  • 1
    (Continued) Try Acrobat Pro or Foxit Phantom to edit the PDFs They allow text editing to some degree, but you must have the used font installed in your computer to avoid font substitutions. That's because those programs avoid helping font piracy. Interpreting outlined text for text changes is advertised at least here esko.com/en/products/deskpack/plugins/text-recognition All mentioned stuff is quite high cost commercial software.
    – user82991
    Jul 21, 2022 at 17:15
  • 1
    Maybe interesting: An Inkscape document made by doing the Poppler-Cairo import of a PDF seems to still know which letter shapes are made of the same font glyph, The letters are not independent paths but clones. A competent programmer probably could restore the text. Unfortunately I am not one.
    – user82991
    Jul 21, 2022 at 19:35

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.