[Top][All Lists]

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

[ft-devel] TrueType scan conversion rules and FreeType

From: Werner LEMBERG
Subject: [ft-devel] TrueType scan conversion rules and FreeType
Date: Mon, 09 Jun 2008 15:01:10 +0200 (CEST)

There's the following comment in ftraster.c:

  Drop-out mode 2 is hard-coded since this is the only mode used on
  Windows platforms.  Using other modes, as specified by the font,
  results in misplaced pixels.

[Note that the drop-out modes as numbered in ftraster.c differ from
 the modes described the documentation of the SCANTYPE command in
 ttinst1.doc, part of the OpenType specification.]

Here is the evidence that this statement is incorrect.  The attached
font contains five times the same glyph (a small Cyrillic `i', taken
from micross.ttf) at positions of the characters 0, 1, 2, 4, and 5,
respectively, selecting the corresponding scan conversion modes (which
I've added manually).  Alexander Morozov was so kind to provide the
attached image which shows the results on Windows (character 0 is the
leftmost glyph in each row; each glyph is printed three times).

In particular: The font micross.ttf needs rendering mode 4 (no removal
of stubs), while, say, verdana.ttf needs rendering mode 5 (with
removal of stubs).  Currently, FreeType always uses mode 5.

Have a look at Savannah bug #23310 report for more.

I'm going to work on this, but after 2.3.6 (which I'll release soon).


Attachment: micross-cyrillic-i.ttf
Description: Binary data

PNG image

reply via email to

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