Re: [Freetype] Unexpected advance values

[Werner LEMBERG]

> I am running my trial with a string of 50 "i".
> 
> - Font is Century Schoolbook L Roman
> - Both GhostScript and Freetype2 load the font from the file
>   c059013l.pfb
> - I print the kerning, it's always 0
> - xfig sets the string width to 3000, which makes an advance of 60
>   per i
> - Freetype2 reports a width of 55 and an advance.x of 63 (which is
>   5% too much).

Hmm, looking into the metrics file c059013l.afm (which FreeType
doesn't use if not explicitly loaded with FT_Attach_File), I see that
there is indeed no kerning between pairs of `i', and that the
horizontal advance value of `i' is 315 font units -- the same value is
actually used in the PFB file.  With 12pt at 1200dpi and 1000 font
units per em (as the default in Type 1 fonts), this makes

  12 * 1200
  --------- / 1000 * 315 = 63
     72

Having 50 `i' glyphs, this makes 50 * 63 = 3150.  In the unlikely case
that I have to remove the left side bearing from the first `i' (18
font units) and the right side bearing from the last `i' (26 font
units), I get 3150 - 8.8 = 3141.2.

Testing with gs, I get the following:

  > gs
  AFPL Ghostscript 8.14 (2004-02-20)
  Copyright (C) 2004 artofcode LLC, Benicia, CA.  All rights reserved.
  This software comes with NO WARRANTY: see the file PUBLIC for details.
  GS>(CenturySchL-Roma) findfont
  Loading CenturySchL-Roma font from
    /usr/local/share/ghostscript/fonts/c059013l.pfb...
    2235788 901660 1442288 145674 1 done.
  GS<1>12 1200 72 div mul scalefont
  GS<1>setfont
  GS>(i) stringwidth
  GS<2>stack
  -0.0
  62.9991

As you can see, `gs' also returns 63 as the string width of the
glyph `i'.  For 50 `i', I get 3149.96 as the result of the
`stringwidth' operator.

I have never used xfig so I can't comment.


    Werner