freetype-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ft-devel] CJK rendering screenshots


From: mpsuzuki
Subject: Re: [ft-devel] CJK rendering screenshots
Date: Mon, 8 May 2006 10:48:09 +0900

Hi,

Sorry for lating.

Just I've attached 2 screenshots of ftview based on freetype
2.1.10 and CVS current. If following conditions are not
suitable, please let me know.

-------------------------------------------------------

freetype2 is built to use unpatented hinting algorithm,
by modifying devel/ftoption.h as:

        #define TT_CONFIG_OPTION_BYTECODE_INTERPRETER
        #define TT_CONFIG_OPTION_UNPATENTED_HINTING

screenshots are taken under the conditions:
        embedded bitmaps are ignored
        anti-alias is off
        glyph hinting is active

sample font is sazanami-gothic.ttf, it is volunteer-based
free font for Japanese script (not commercial quality).
Gothic (HeiTi, for Chinese) families are designed to use
fixed-width strokes, so I think this is the simplest example.
-------------------------------------------------------

The point size is 20. If the point size is less than 14,
it's almost impossible to render all strokes in complicated
glyphs (so most bitmap fonts omit strokes to keep readability
in such small size).

Comparing 2 results, you can find that the stroke width in
freetype-2.1.10 result (in most glyphs) is bolder (2 pixel)
than freetype2-current result (1 pixel). However, even in
freetype2-current, still some strokes are rendered by bolder
strokes.

Because freetype-2.1.10 often uses bolder strokes,
the compilicated glyphs rendered by freetype-2.1.10 show
many "blackout" boxes, in comparison with freetype2-current.
Thinking about the fact that most Japanese 20 pixel bitmap
fonts draws 1 pixel strokes, the result by freetype2-current
is reasonable.

Yet I've not tested the Mincho (SongTi, for Chinese) families
that uses different stroke widths for horizontal & vertical
strokes. I'm going to test Mincho style, but I wish if any
PRC resident evaluates for SongTi results (because the contrast
of stroke width in SongTi is more explicit in comparison with
Japanese Mincho).

Regards,
mpsuzuki

Attachment: ftview-2_1_10-sazanami-2.png
Description: Binary data

Attachment: ftview-current-sazanami-2.png
Description: Binary data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]