Re: [ft] Bytecode hinter producing bad results, seems to be a regression

[Werner LEMBERG]

> It seems the bytecode hinter is defective, and the autohinter is
> working.  Check out this image:
> 
> http://img524.imageshack.us/img524/1057/freetypehinting.png
> 
> The only difference in the output, is the use of
> FT_LOAD_FORCE_AUTOHINT.

You haven't told us which FreeType version you are using...  For bug
reports, please always try the recent one, this is, 2.4.4.

> I'm using a professional font, Helvetica Neue Com LT Md 65 from
> Linotype.

Assuming that you are using normal hinting, the eight lines starting
from the fourth line in the left part of the image are bad.  This
looks very suspicious, namely that the font only contains hints for
non-antialiasing in this range (as specified in the `gasp' table),
something which you apparently ignore.  Please use fontforge or ttx to
check your font whether my assumption is correct.  I only have
`Helvetica Neue LT Pro 65 Medium' version 001.000 (from 2003), but
this is an OpenType font with CFF outlines.

To fix this you should apply the proper rendering flags based on the
values returned by FT_Get_Gasp.

BTW, I'm not sure that `professional' fonts are intended for screen
display, working at small ppem values.  I would rather assume that
such fonts are for professional printing, and for that you essentially
don't need hinting at all.

> In the older version of Freetype (I forget which one), activating
> the bytecode hinter did not produce those sickly thin and sometimes
> non-existent stems.

I doubt that, but maybe there was or is a bug in FreeType.


    Werner