The only problem I know of is a cosmetic one in certain (older?) Adobe apps. I seem to recall it was just in Illustrator. The inverted selection area that you get when you select text will be displayed at the wrong height. If it's a PS or CFF font, it assumes 1000 units and displays the selection area at that height even when the UPM is different. So, if your UPM is 2048, the selection area will be about half the expected height. The font will still work fine otherwise.
We had some problems with CFF fonts with an UPM bigger than ~5000. It looks fine on screen and print but you can’t convert it to outlines in Indesign and Illustrator (CS3+4). It will distort all paths above 5000 units. So if you go 4000 you are save, you do not need to use the TTF numbers.
Or, as @mekkablue said, use fractional positioning in Glyphs. Disable the auto hinting as the adobe autohinter will do rounding. I have to rewire most of the adobe code to improve that.
Fonts with PS outlines (Type 1 and OpenType CFF) have a maximum coordinate value of +/- 4096. This is the biggest, albeit indirect, limitation on unusually large em squares for OpenType CFF (OTF) fonts.
If you go beyond the 4096-unit limit, you are taking your chances. Lots of things will doubtless be fine, but without inspecting source code or experimenting, I don't know how one would predict what will fail.
Anything that has its own unique native rendering capability for fonts would need to be tested. A few obvious ones:
- OS X Quartz - Windows GDI - Windows DirectWrite - OpenType Sanitizer (in Firefox and Chrome) - FreeType - Adobe CoolType - Adobe PostScript (assorted version) - each different PostScript clone on the market - other printer drivers that have their own font rendering engine - etcetera
Just for a start, I will bet (lunch at the next type conference we're both at) that the OpenType Sanitizer will reject an OTF with > 4096 upm. If it doesn't reject such a font today, and we tell them about it, they will make sure it does in the future. They are pretty eager to restrict fonts to confirm to published specs... even if Mac OS and Windows are fine with the fonts in question.
Comments
Or, as @mekkablue said, use fractional positioning in Glyphs. Disable the auto hinting as the adobe autohinter will do rounding. I have to rewire most of the adobe code to improve that.
Anything that has its own unique native rendering capability for fonts would need to be tested. A few obvious ones:
- OS X Quartz
- Windows GDI
- Windows DirectWrite
- OpenType Sanitizer (in Firefox and Chrome)
- FreeType
- Adobe CoolType
- Adobe PostScript (assorted version)
- each different PostScript clone on the market
- other printer drivers that have their own font rendering engine
- etcetera
Just for a start, I will bet (lunch at the next type conference we're both at) that the OpenType Sanitizer will reject an OTF with > 4096 upm. If it doesn't reject such a font today, and we tell them about it, they will make sure it does in the future. They are pretty eager to restrict fonts to confirm to published specs... even if Mac OS and Windows are fine with the fonts in question.
Cheers,
T