I have a question regarding the export options in Glyphs. My understanding: ttfautohint works better for low resolution environments, Windows and has generally an advantage where the rasterizer respects the font’s embedded hinting instructions. I tested in Edge on Windows and found a ttfautohinted woff2 to work better at small sizes compared to a postscript/cff autohinted woff2 especially small shapes like superior figures etc.
In macOS and other environments where font hints are mostly ignored it shouldn’t matter much whether it’s ttfautohinted or postscript/cff autohinted, is that right?
Given that use on websites puts a font on an unpredictable mix of platforms, I’m leaning toward ttfautohint as the safer default. Is that reasoning sound, or is there more to consider?
What Sebastian said. But still test on all platforms you can get your hands on. Some fonts, e.g. display fonts, fonts that are intended for bigger sizes might look better as otf. And the file size can be different too.
Thanks guys, yes makes sense! I tested it and i found them to work better with ttfautohint i just wanted to see if theres any other opinions on that matter. You say display fonts because the original shapes of the font are better preserved / less distorted with postscript autohinting right?
Yes, TT requires outline conversion (if the original is drawn in cubic curves). This outline conversion distorts the original outlines somewhat and can introduce unwanted artefacts, so it’s better to use CFF fonts (.otf), which preserve the original cubic outlines.
Furthermore, for display fonts, hinting will be irrelevant anyway, since they will be used at big sizes only. Hinting only matters at small sizes.
The outline conversion is usually not the problem. But CFF curves are rendered a bit more truthfully (with and without hints). And TrueType can look a bit edgy around 20–30 pt because of stronger grid fitting. look at a waterfall from very small to reasonably big (40pt) and compare.