Post #1793965
2026-04-24 16:31 UTC
Fortunately, the /Font object specifies the font name: Open-Sans, which is an extremely common font distributed by Google.
So we can make sure the font is installed on our machine (on Linux: sudo apt install fonts-open-sans) then run Ghostscript to re-write the PDF and embed the font:
gs -o fixed.pdf \
-sDEVICE=pdfwrite \
-dEmbedAllFonts=true \
-dSubsetFonts=true \
-sFONTPATH="/usr/share/fonts/truetype/open-sans" \
corrupted.pdf
Replies (1)
-
@BertrandCaron@digipres.club 2026-04-24 16:39
The result is not perfect - still some overlapping characters - but at least it can be read in Adobe Reader! I will probably file a new issue for #JHOVE. I'm always wondering what benefit we get from a new test, but here it seems justified for two main reasons: a) It has a severe impact on rendition,: viewers, esp. Adobe Reader, do not cope well with, b) There is a possible remediation.