I have a font typeface I’ve started that has a handful of non-UNICODE characters that are normally entered in a document as combined keystrokes. Normally I would just use pair adjustments, but some of these non-UNICODE characters contain 3 or 4 combined keystrokes to make the character (for example: a unicode character with 2 separate combining diacritics).
I’ve used a Ligature Sustitution method for taking the combining keystrokes and substituting in the final character the way I’d like to see it. This appears to work just fine in Microsoft products, but when converted to PDF, the Ligature Substitutions appear to be ignored by Adobe. Any ideas?
Thank you in advance for any help you can provide.
This looks like a problem with the PDF export from Word.
Look at any of my Typeface Samples PDF files (produced from Serif PagePlus), and the ligature substitutions all seem to be working just fine.
Did you have to turn on any other settings in the font properties or elsewhere for the PDF to recognize your substitutions?
I found that if I go into the PDF and select the text that shows the combined keystrokes incorrectly and paste them in the same PDF as a new text string, the character shows up correctly. I am confused beyond belief now.
The text string in the PDF should be exactly what is typed in the publishing application (Word or PagePlus). If I copy some text from the PDF, I get the plain text before the substitution:
kunststoffflugzeug to plastic aircraft ff fi
If I paste that into PagePlus, and enable ligatures using a different font with the same liga and dlig features, I get the substitutions as in the PDF.
That is how glyph substitutions work. They are performed by the font. If you change the font to one that does not include the OpenType features, you get the plain text string. Perhaps the PDF is substituting a fall-back font?
I don’t think you need the kerning(1) lookup at all. The StandardLigatures lookup does the business.
Some PDF producers and/or viewers rely on glyph names, so you could let FontCreator generate proper glyph names, and then make sure you also export them.
In your case after generating glyph names, the ligature should have this name:
idotless_dotaccentcomb_macroncomb
Let us know if that solves the problem.
Also let us know if you are exporting with with TrueType or CFF based outlines, as I think Word does a better job with with TrueType based outlines.
I don’t think this is a Word issue. I think it is in how Adobe is interpreting the text after it is exported to PDF. All our characters look good in Word. All substitutions are working.
I did try updating the Glyph Names for the ligatures, however, the results were the same.
I exported to a different font family name too just to rule out any reference to the older iterations of the font I had installed. Still same results only with the PDF.