Different rendering between XeTeX, LuaTeX and MS Word

[YuanshengZhao]

The rendering of the three engines are different (Currently, the font is optimized for and tested on XeTeX with unicode-math).

• LuaTeX ignores math kern (if there are any).
• normal kern will be taken into account only in inline math of LuaTeX.
• MS Word not add italic-correction to the width of glyphs in juxtaposition. This now makes the font unusable in Word (too cramped)

[YuanshengZhao]

There is two kinds of builds: Garamond-Math.otf(the position of script is controled by: upright->math kern; italic->italic correction) and Garamond-Math-Word.otf(the position of script is controled math kern only). For best rendition, use Garamond-Math.otf in *TeX; and Garamond-Math-Word.otf in MS Word / XeTeX

[callegar]

Kerning in LuaTeX is indeed rather problematic. Out of curiosity, is there any particular reason why it is different in XeLaTeX and LuaLaTeX? Other math fonts manage to score well in both.

[YuanshengZhao]

With the update of the engines, there have been large improvements in the rederning. I believe that for the latest version, both lua and xe have reasonable output.

The main reason for these problems, I think it is partly the result of the large italic angle and small x-height of the font, which can cause many problems in math environment where the metric and kerning is different from text.

[YuanshengZhao]

For the newest font and engine, the metric should be similar. Though there still some differences:

• LuaTeX does not support nesting of substitutions. this is the reason of https://github.com/YuanshengZhao/Garamond-Math/issues/14
• Usual kerning not supported in XeTeX
• Left mkern: has no effect in XeTeX for V_{AA}, etc. (V_{AA} works with both engines)