gsautter
commented
4 years ago Font decoding options for born-digital PDFs:
- "Do Not Decode Fonts": simply use the Unicode mapping (makes sense mostly when conversion goal is either testing or extraction of figures and graphics, or PDF does not use embedded fonts at all)
- "Render Glyphs Only": draw the character images (glyphs) encoded in embedded fonts (e.g. for later correction), but use the Unicode mapping as is (makes sense mostly when conversion goal is either testing or extraction of figures and graphics, or PDF is known to come with comprehensive Unicode mapping)
- "Decode Unmapped": draw the character images (glyphs) encoded in embedded fonts, use the Unicode mapping as is where available, and decode character images that lack a Unicode mapping via visual comparison to reference fonts (makes sense mostly when PDF is known to come with faithful, but sometimes incomplete Unicode mapping)
- "Verify Mapped": draw the character images (glyphs) encoded in embedded fonts, verify available Unicode mappings via visual comparison to reference fonts, and decode character images that lack a Unicode mapping via latter as well (makes sense mostly when PDF is known or suspected to come with incomplete and partially erroneous Unicode mapping)
- "Decode All Glyphs": draw the character images (glyphs) encoded in embedded fonts and decode character images via visual comparison to reference fonts regardless of any available Unicode mappings (makes sense mostly when PDF is known to come with incomplete, erroneous, or outright malicious Unicode mapping - the latter happens when Unicode the mapping is abused as a protection against copy&paste from the PDF in a viewer)
Computational effort increases from top to bottom, as does the level of scrutiny the embedded fonts are subjected to. "Decode Unmapped" usually is a good starting point, with the option to escalate to "Verify Mapped" or "Decode All Glyphs" and re-convert a PDF in case the conversion result has too many character errors with "Decode Unmapped". "Render Glyphs Only", and especially "Do Not Decode Fonts", a quick decode options useful mostly for testing some PDF, or for extracting figure images and graphics with little regard for the text, even though the former can (and often does) produce good results for PDFs whose embedded fonts come with faithful and complete Unicode mappings. If the latter is known to be the case, "Render Glyphs Only" is a viable option.
@gsautter can you please explain all the options to encounter in the opening menu of GGI and what when should be used so we can add this to the manual.