Re: [Freetype] News from the field

[Roberto Peon]

Hmm... I'm having issues with certain fonts (it is especially noticable 
with numbers) rendering distorted with hinting turned on...
Specifically, numbers such as 3 or 7 (random examples depending on 
font size) may appear 1 pixel larger (very noticable, even on TV!) than
the other numbers.

With hinting  disabled, they render fine. I'm willing to send the font 
files (ttf) to interested parties.

This 'bug' effects much of the 2.1.x series for me.

And Congrats David! (from a longtime list lurker =) )

-Roberto JP
address@hidden

---------- Original Message ----------------------------------
From: Anthony Fok <address@hidden>
Reply-To: address@hidden
Date:  Thu, 13 Mar 2003 01:49:27 +0800

>Hello David,
>
>On Wed, Mar 12, 2003 at 11:51:29AM +0100, David Turner wrote:
>> Hello everyone,
>>   I'd like to announce you that I'm now the happy father of a little
>>   girl named Lucie, who was born on February 26th 2003 !
>>   [...]
>>   And guess what, I'm so happy to be "dad" ;-)
>
>Congratulations!  :-)
>
>>      - still these crazy TrueType native rendering glitches. ah ah !!
>
>For what it's worth: I finally tried some experimenting last night, and
>my troubles -- FreeType-2.1.4rc2 with bytecode interpreter turned on
>caused distorted glyphs at certain sizes -- were gone when I _reversed_
>your recent fix for a rounding "bug" in ttfdriver.c:
>
>===========================================================================
>2003-02-25  David Turner  <address@hidden>
>
>    * src/truetype/ttdriver.c (Set_Char_Sizes): fixed a rounding bug when
>    computing the scale factors for a given character size in points with
>    resolution.
>
>diff -u -r1.60 -r1.61
>--- freetype-2.1.4rc2~/src/truetype/ttdriver.c 2003/02/18 22:25:22     1.60
>+++ freetype-2.1.4rc2/src/truetype/ttdriver.c  2003/02/25 20:37:50     1.61
>@@ -208,8 +208,8 @@
>      /* we need to use rounding in the following computations. Otherwise,
>       * the resulting hinted outlines will be very slightly distorted
>       */
>-      dim_x = ( ( ( char_width  * horz_resolution ) / 72 ) + 32 ) & -64;
>-      dim_y = ( ( ( char_height * vert_resolution ) / 72 ) + 32 ) & -64;
>+      dim_x = ( char_width  * horz_resolution + 36 ) / 72;
>+      dim_y = ( char_height * vert_resolution + 36 ) / 72;
> 
>       metrics2->x_ppem  = (FT_UShort)( dim_x >> 6 );
>       metrics2->y_ppem  = (FT_UShort)( dim_y >> 6 );
>===========================================================================
>
>So I am starting to wonder which whether ((x+36)/72) or (((x/72)+32)&-64)
>is correct.  Is the rounding "bug" fix in 1.61 necessary for the
>bluescale patch?  Or perhaps the former is good for the autohinter, and
>the latter is good for bytecode-interpreter hinting instruction?
>
>Cheers,
>
>Anthony
>
>-- 
>Anthony Fok Tung-Ling
>ThizLinux Laboratory   <address@hidden> http://www.thizlinux.com/
>Debian Chinese Project <address@hidden>       http://www.debian.org/intl/zh/
>Come visit Our Lady of Victory Camp!           http://www.olvc.ab.ca/
>
>_______________________________________________
>Freetype mailing list
>address@hidden
>http://www.freetype.org/mailman/listinfo/freetype
>