Re: [ft] diff Pixel height

[Werner LEMBERG]

Samaram,


I've now found some time to answer this your question:

> I am using Arial.ttf file [at 12px/72dpi], [...] I am getting x min
> for few characters as -1 [after a call to compute the CBox].  Is it
> correct?

Yes.  Have a look at the attached image, from a debugging session with
FontForge.  It shows the outline of glyph `A' from arial.ttf at 12px
with 72dpi, after hinting has been applied.  The hinting engine moves
point 0 to the left so that it eventually gets a negative x value.
After rounding, this becomes -1.  Similarly, point 3 gets moved to the
right, causing the xmax value to be 8.  In other words, the CBox of
the hinted glyph `A' is two pixels wider than necessary.

Always bear in mind the following warning from the description of the
`bbox' field in FT_FaceRec:

  Note that the bounding box might be off by (at least) one pixel for
  hinted fonts.  See @FT_Size_Metrics for further discussion.

In the description of FT_Size_Metrics:

  [...] Indeed, the only way to get the exact metrics is to render
  _all_ glyphs.  As this would be a definite performance hit, it is up
  to client applications to perform such computations.


    Werner