freetype-devel
[Top][All Lists]
Advanced

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

Re: Odp: Re: The TrueType rendering accuracy nightmare (Re: Freetype-dev


From: Hin-Tak Leung
Subject: Re: Odp: Re: The TrueType rendering accuracy nightmare (Re: Freetype-devel Digest, Vol 184, Issue 24)
Date: Fri, 15 May 2020 14:38:01 +0000 (UTC)

Please include freetype-devel in all following discussions. Or stop posting.

There are official demos which allows you to change modes. They are in the ft2-demos repo. Many linux distros bundle them as freetype-devel or freetype-tools. Fontforge is not freetype.

I am not sure what you are trying to say here or trying to do. Your own program is buggy - so go fix it.  As I mentioned twice already - we are on reasonable friendly terms with the Microsoft folks, and they are happy to confirm undocumented behaviors, and afaik from your screenshots, freetype /fontforge matches Microsoft's rendering quite well, and your program is way off and buggy. If you want to ask for help on debugging, ask for help properly.



On Friday, 15 May 2020, 05:39:13 GMT+1, address@hidden <address@hidden> wrote:



About my 2nd point - you did not get it: fontforge is not freetype. If you want freetype to imitate Microsoft's bilevel rendering, please do as I suggested. Fontforge does not have a mode of "use freetype to imitate Microsoft's rendering", afaik. Use v35 and set rendering mode to mono.
FreeType by itself doesn't layout text or select foreground or background colors, an external program must be used to render with FreeType. And I cannot expect a change of mode to somehow fix the bugs.


As for the initial paragraph - it is clear from your own png posted that freetype (in fontforge) matches Microsoft's rendering well enough, but your own rendering does not. You need to show a lot more evidence and proof to claim freetype is glitchy - even our Microsoft friends don't make that kind of claim!

What? If anything, it's evident that neither of the two open-source rasterizers are accurate. If both rasterizers are inaccurate you can't just arbitrarily mark one the better. The test can only pass if the rendering is pixel by pixel correct. Also, FreeType's rasterizer probably had much more time to get developed than mine; FreeType may have had multiple updates while it is still the first version of TD rasterizer, so it is yet to improve.



reply via email to

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