Re: [Freetype] Opening one font causes distortion of another

[Wenlin Institute]

At 9:24 AM +0200 6/3/03, Werner LEMBERG wrote:
> Is the problem still there if you use the current CVS?  I don't have
> those fonts so I can't debug it myself.  Have you been able to verify
> whether autohinting is indeed the problem?

With 2.1.4, the discrepency disappeared when FT_LOAD_NO_HINTING was 
specified for FT_Load_Glyph(). Interestingly, the bitmap produced 
with hinting turned off was different from both of the bitmaps 
produced with hinting turned on.

With the current code (June 5: unstable/freetype2-current.tar.gz), 
the discrepency has disappeared regardless of FT_LOAD_NO_HINTING. 
And, the hinted bitmap is identical to the unhinted bitmap.

Unfortunately, there's a decrease in rasterizing quality between 
2.1.4 and the current code. For comparison, I've put a small PDF with 
samples at this URL:

   http://www.wenlin.com/ft_diff.pdf

2.1.4 did a better job with both the Chinese and the roman text. The 
current code makes the Chinese blurry regardless of autohinting, and 
it makes the roman terrible with autohinting on. With autohinting 
off, the roman isn't so bad, but it's still not as good as with 2.1.4 
and autohinting on. (These are both TrueType fonts -- the Chinese one 
is mssong.ttf, from Microsoft, and the roman one is modified from the 
S-Garamond from www.rubicon.ca. Patented TrueType hinting is not 
enabled.)

By the way, in the current code, ftmac.c won't compile due to an 
undeclared variable "rlen" in the function FT_New_Face_From_SFNT(). 
Maybe it should be "sfnt_size", but I simply replaced ftmac.c with 
the 2.1.4 version in order to compile the library.

Tom Bishop

--

Wenlin Institute, Inc.        Software for Learning Chinese
E-mail: address@hidden     Web: http://www.wenlin.com
Telephone: 1-877-4-WENLIN (1-877-493-6546)