[XeTeX] Work around for Italic correction; +kern doesn't work(?)

[Adam Lindsay]

Hello Robert, 

I don't mean to be pedantic here, but I am trying to clarify things to be
sure bigger TeXperts can get to the heart of the problem. I'm not using
XeLaTeX, so I tried debugging using just plain XeTeX.

Robert Voogdgeert said this at Tue, 28 Dec 2004 23:26:39 +0100:

>I recall there being no italic-correction proper in XeTeX. As I'm facing
>a problem now that preferably needs some kerning, I wondered whether any
>of you has come up with some work-around yet.
>My problem is the italicized German 'Begriff' set with Minion Pro
>Regular, but and this is problematic, within parentheses: (Begriff). The
>last `f' running into the left top of ) quite a lot.

Most likely, you mean the 'ff' ligature that's running into the parenthesis.

I tried the example in XeTeX, and it gave great kerning, so I dug further...

>Even in absence of Italic correction, the Minion Pro (and other otf
>fonts) have good kerning. In InDesign for example the use of Minion Pro
>Italic with the kerning option set to metric, gives nice results,
>slightly better than with optical kerning. 

I do wonder if you tried the exact same sample text in both. I suspect
that in XeLaTeX you had an italic 'Begriff' inside regular (upright)
italics. You're right, that looks awful. (Okay, re-reading your statement
above, you sort of say that's what you're doing. Sorry.)

But I suspect that in InDesign you had the parens in italics as well. Is
this true?

>I tried +kern in Jon's minion fd for Italic, but this didn't seem to
>alter anything. Is this tag not supported by XeTeX, or is the problem
>more complex?

If my guesses above are correct, then automatic kerning from the font
won't help you here: you're switching fonts, and kerning tables don't
span across fonts. You hit on it: you want italic correction.

If it's failing for you within XeTeX, then you can always add a kern manually:
(\emph{Begriff\kern0.167em})

Where you use whatever italics command you like and adjust the kerning
manually. The 0.167em figure (16.7% of the font size; equal to 2pt in a
12pt font) was deduced with the help of Eddie Kohler's otftotfm:

  otftotfm -p --no-type1 -f kern -f liga MinionPro-It.otf

...and looking at the resulting MinionPro-It--Fkern--Fliga.pl file:

(CHARACTER D 1 (COMMENT ff)
   (CHARWD R 530)
   (CHARHT R 712)
   (CHARDP R 257)
   (CHARIC R 167)
   )

I just started looking at PL files recently, though, so I could be wrong.

Looks like there's a workaround, and some more info for further
investigation, but I don't have a general solution for you!

adam
-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Adam T. Lindsay, Computing Dept.     atl at comp.lancs.ac.uk
 Lancaster University, InfoLab21        +44(0)1524/510.514
 Lancaster, LA1 4WA, UK             Fax:+44(0)1524/510.492
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-