bugThe FreeType Project - Bugs: bug #15785, character "7" in verdana...

 
 

bug #15785: character "7" in verdana regular 10px with hinting without antialiasing is bad rendered.

Submitted by:  None
Submitted on:  Fri 17 Feb 2006 11:20:52 PM UTC  
 
Severity: 3 - NormalItem Group: Hack required
Status: FixedPrivacy: Public
Assigned to: Werner LEMBERG <wl>Open/Closed: Closed
Planned Release: 2.3.10

Add a New Comment(Rich Markup)
   

You are not logged in

Please log in, so followups can be emailed to you.

 

(Jump to the original submission Jump to the original submission)

Tue 16 Jun 2009 01:41:07 PM UTC, comment #11:

Finally fixed, together with many other B/W rendering issues in various well-known TT fonts like tahoma. Please test the current git.

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Wed 08 Apr 2009 01:41:10 PM UTC, comment #10:

We have this bug again, however, for other reasons: It's now
related to a border case of dropout control which is very hard
to understand not clearly explained in the OpenType
specification.

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Fri 16 Feb 2007 08:22:11 AM UTC, comment #9:

This seems to be fixed now in the current CVS --
David has updated the TT interpreter code.

Please test!

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Tue 21 Feb 2006 12:29:00 PM UTC, comment #8:

That's not true. I misread rule 4.

Chia-I Wu <olvaffe>
Project Member
Tue 21 Feb 2006 10:14:00 AM UTC, comment #7:

This potential dropout (the lowest pixel of the foot of `7') does not become an actual dropout because y=0.5 is the lowest scanline intersecting the foot and that pixel is thus treated wrongly as a stub. (line 2227 of ftraster.c).

My understanding of rule 4 of the scan converter is that, we should extend the two line segments forming the foot to see if they intersect each other before the next scanline (y=-0.5). It is a stub only when this holds and this is not the case (the line segments are parallel!).

Chia-I Wu <olvaffe>
Project Member
Sun 19 Feb 2006 04:33:47 PM UTC, comment #6:

If none was able to, then the chance is high that I was wrong. I retrieved the outline through GetGlyphOutline with GGO_NATIVE. The coordinates of the points are the same as what freetype says, for the several nominal sizes I tried.

Chia-I Wu <olvaffe>
Project Member
Sun 19 Feb 2006 04:19:49 PM UTC, comment #5:

Interesting! This is very good news since until now apparently
noone was able to retrieve the hinted outline on the Windows
platform.

My method was to run the TrueType debugger within FontForge
(this currently fails for me for various reasons).

Of course it can be a problem with the monochrome rasterizer!
Please investigate.

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Sun 19 Feb 2006 03:55:25 PM UTC, comment #4:

I tried to dump the outline of the hinted glyph under windows and it is exactly the same as what we get from freetype. Is this how you know that they are different bugs?
(It's my first win32 app and I'm not sure whether it works correctly...)

So, could it be a bug of the monochrone raster?

Chia-I Wu <olvaffe>
Project Member
Sun 19 Feb 2006 02:38:50 PM UTC, comment #3:

In case it is the problem I suspect, the answer is no. You might
refer to the freetype-devel list where this has been discussed
already (w.r.t. bad pixels in tahoma.ttf).

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Sun 19 Feb 2006 12:19:44 PM UTC, comment #2:

It isn't posible to do same (additional?) rounding, what Windows does?

--Lefti

Anonymous
Sun 19 Feb 2006 08:53:19 AM UTC, comment #1:

I can confirm this. Later on I will check whether this problem
belongs to the category `buggy windows truetype engine', this is,
whether the Windows engine does (additional?) rounding, not
documented in the TrueType specs. If this is the case (and I
suspect this) we can't do anything against this.

Werner LEMBERG <wl>
Project AdministratorIn charge of this item.
Fri 17 Feb 2006 11:20:52 PM UTC, original submission:

Hi,
There is bug in freetype: look at http://mbox.troja.mff.cuni.cz/~peak/ftdemo/index.cgi and set hinting :on, antialiasing:off,font:verdana.ttf,size:10 and text to 0123456789 or something else what containts character 7 and look at "7". The downest pixel is missing.

screenshots:
on my linux box: http://sekec.net/7.png
on windows: http://sekec.net/7w.png
expected (handly edited): http://sekec.net/7e.png

PS: Size of font is 10px in ftdemo,12px in "my" gtk2 and 13px in "my" firefox

Anonymous

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by wl
  • -unavailable- added by wl (Posted a comment)
  •  

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

    Please enter the title of George Orwell's famous dystopian book (it's a date):

     

     

    Follow 15 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Tue 16 Jun 2009 01:49:10 PM UTCwlCarbon-Copy-=>Added -unavailable-
    Tue 16 Jun 2009 01:48:12 PM UTCwlCarbon-CopyRemoved pocek=>-
    Tue 16 Jun 2009 01:41:07 PM UTCwlStatusConfirmed=>Fixed
      Open/ClosedOpen=>Closed
      Planned Release2.3.2=>2.3.10
      Carbon-Copy-=>Added pocek
    Wed 08 Apr 2009 01:41:10 PM UTCwlStatusFixed=>Confirmed
      Open/ClosedClosed=>Open
    Tue 12 Jun 2007 12:28:27 PM UTCwlStatusConfirmed=>Fixed
    Fri 16 Feb 2007 08:22:11 AM UTCwlOpen/ClosedOpen=>Closed
      Planned ReleaseNone=>2.3.2
    Tue 21 Feb 2006 01:31:40 PM UTCwlCarbon-Copy-=>Added freetype
    Sun 19 Feb 2006 08:53:19 AM UTCwlStatusNone=>Confirmed
      Assigned toNone=>wl
      Carbon-Copy-=>Added olvaffe

    Back to the top


    Powered by Savane 3.1-cleanup1