RE: [ft] Bytecode enabled rendering

[Turner, David]

Sorry, I jumped too quickly in the discussion.

If I recall correctly, the Tahoma fonts are hinted to
appear perfectly on the Windows font rasterizer.

Analysis has shown that in some cases, the Windows
engine produces incorrect results, regarding the vectors
it was passed as input.

By invalid, I mean that it cleary contradicts the TrueType
rasterizations rules, which are very strict. We believe it
is due to some rounding errors, probably the consequence
of different algorithms, say like the one used to compute
the intersections of bezier arcs and scanlines.

We have not the time to study this any further, nor even
provide patches to our bytecode interpreter or rasterizer.

If you can provide a patch, we'll accept it, but only after
careful testing, because it'd better not ruin the rendering
of other fonts for the sake of "optimizing" Tahoma and
Times New Roman.

Hope this helps,

- David Turner
- The FreeType Project  (www.freetype.org)


> -----Message d'origine-----
> De : address@hidden
> [mailto:address@hidden la
> part de Anton
> Danilov
> Envoyé : mercredi 16 novembre 2005 13:53
> À : address@hidden
> Objet : Re: [ft] Bytecode enabled rendering
>
>
> I am not speaking about anti-aliased rendering in any way and I
> completely agree on your position about it. That all is quite
> reasonable.
>
> What I do speak about is incorrect rendering of
> non-anti-aliased text.
> Is this case closed as well and there is no hope Tahoma,
> Times New Roman
> and friends will look better when rendered with Free Type?
>
> Anton
>
> Turner, David wrote:
> > I've already spoken about this in a previous e-mail.
> >
> > Microsoft holds a large number of patents covering specific
> > parts of the ClearType technology, see the bottom of this
> > page for a non-exhaustive list:
> >
> >   http://www.microsoft.com/mscorp/ip/tech/cleartype.asp
> >
> > For the moment, we're not going to implement Microsoft's
> > rendering techniques (which are also documented in a document
> > named "Avalon Text", that you'll find with a search engine).
> >
> > Don't expect to have identical rendering in the case of
> > anti-aliased text with FreeType on Linux.
> >
> > We could implement it, just like bytecode interpretation,
> > by disabling it by default in the sources. But very frankly,
> > I don't care enough about this to spend my time on it.
> >
> > If you happen to take the time to provide patches to FreeType
> > and/or libXft to support that, we may even *not* be interested
> > in integrating them to the official FreeType sources.
> >
> > You could still distribute a patch however, that distributions
> > could pick if they want to.
> >
> > Case closed.
> > 
> > - David Turner
> > - The FreeType Project  (www.freetype.org)
> >
> >
> >
> >
> >
> >>-----Message d'origine-----
> >>De : address@hidden
> >>[mailto:address@hidden la
> >>part de Anton
> >>Danilov
> >>Envoyé : mercredi 16 novembre 2005 11:00
> >>À : address@hidden
> >>Objet : Re: [ft] Bytecode enabled rendering
> >>
> >>
> >>Thanks for the answer, Werner!
> >>
> >>I see your point. But what really makes this situation strange and
> >>unnatural is that Windows makes "undocumented and
> >>unpredictable" things
> >>which produce well-documented and predictable results, am I
> not right?
> >>
> >>It's clear that these fonts were designed by Monotype to look
> >>exactly as
> >>they look under Windows.
> >>
> >>You're speaking about the patch for "these two symbols". I have got
> >>problems with several symbols on Tahoma 8 pt, some other symbols on
> >>Tahoma 10 pt, Times New Roman 14 pt.
> >>
> >>Maybe, there is some way to fix these symbols on my own if
> you do not
> >>plan to create this patch soon?
> >>
> >>Anton
> >>
> >>Werner LEMBERG wrote:
> >>
> >>>>I am experiencing problems with aliasing-disabled
> rendering, and if
> >>>>you look at Kaya's screenshot (my system, and any other will
> >>>>probably do, shows just the same behaviour)
> >>>>
> >>>>http://kayalabs.com/images/tahoma.png
> >>>>
> >>>>you will see the difference in rendering Tahoma's 8 and v -- in
> >>>>freetype's version you have just one more pixel and it's really
> >>>>disturbing and annoying.
> >>>
> >>>
> >>>This has been discussed at great length on the freetype-devel list,
> >>>IIRC.  The `correct' rendering result of Tahoma on Windows
> >>
> >>is tightly
> >>
> >>>bound to anomalies in the Windows rendering engine which internally
> >>>rounds certain values differently, namely in an undocumented and
> >>>unpredictable way.  With other words, Windows is `wrong'.
> >>
> >>The correct
> >>
> >>>solution is to fix the bytecode instructions for those two
> >>
> >>characters
> >>
> >>>-- this breaks the digital signature, but FreeType ignores
> >>
> >>it anyway.
> >>
> >>>A longer time ago I announced to provide a patch, but
> >>
> >>unfortunately I
> >>
> >>>haven't found enough time yet to do that.
> >>>
> >>>
> >>>    Werner
> >>
> >
> **************************************************************
> *********************
> > Information contained in this email message is confidential
> and may be privileged, and is intended only for use of the
> individual or entity named above. If the reader of this
> message is not the intended recipient, or the employee or
> agent responsible to deliver it to the intended recipient,
> you are hereby notified that any dissemination, distribution
> or copying of this communication is strictly prohibited. If
> you have received this communication in error, please
> immediately notify the address@hidden and destroy the 
> original message.
> >
> **************************************************************
> *********************
> >
> >
> >
> > _______________________________________________
> > Freetype mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/freetype
>
***********************************************************************************
Information contained in this email message is confidential and may be 
privileged, and is intended only for use of the individual or entity named 
above. If the reader of this message is not the intended recipient, or the 
employee or agent responsible to deliver it to the intended recipient, you are 
hereby notified that any dissemination, distribution or copying of this 
communication is strictly prohibited. If you have received this communication 
in error, please immediately notify the address@hidden and destroy the original 
message.
***********************************************************************************