W3C home > Mailing lists > Public > www-style@w3.org > October 2012

Re: Text anti-aliasing on the Mac

From: Andrew Fedoniouk <news@terrainformatica.com>
Date: Tue, 2 Oct 2012 20:10:20 -0700
Message-ID: <CALRQH79Y7y34hr_n+SnrwhANCf8-y7tJ6dmMMdoHn=vb3q4Atw@mail.gmail.com>
To: John Hudson <tiro@tiro.com>
Cc: www-style@w3.org
On Tue, Oct 2, 2012 at 6:38 PM, John Hudson <tiro@tiro.com> wrote:
> On 02/10/12 4:49 PM, Andrew Fedoniouk wrote:
>
>> But I agree that in some cases some configuration is needed.
>> For example I was forced to introduce property
>>
>>    font-rendering-mode: snap-pixel | sub-pixel;
>>
>> for my engine running Direct2D backend on Windows.
>
>
>> font-rendering-mode:snap-pixel; is what Windows GDI uses for rendering
>> fonts in desktop UI - glyphs snap to pixel grid. That is good for e.g.
>> input elements especially edit elements.
>> and text having relatively small font-sizes (but not always).
>> font-rendering-mode: sub-pixel; is when glyph allowed to start
>> anywhere inside single pixel.
>
>
> Your terminology may be confusing for some people. It sounds like you are
> describing what Microsoft's rendering folk call full- vs sub-pixel
> *positioning*. Sub-pixel positioning is a particular kind of (usually*)
> sub-pixel rendering in which, as you say, glyphs outlines may begin anywhere
> inside the pixel (this is DirectWrite's default rendering). GDI ClearType
> does sub-pixel rendering but positioned on a full pixel grid, so your term
> 'font-rendering-mode: sub-pixel' sounds ambiguous even though you seem to
> mean specifically sub-pixel positioning.
>

Thanks for the clarification.
Yes, I know that these names are technically not that perfect.
But they are short  :)

>
> * This all got more complicated in the Windows 8 Metro environment, in which
> greyscale 8x4 asymmetric antialiasing is applied with sub-pixel positioning.
> And if that's a surprise to you imagine how I felt when told of this change
> two weeks before code lockdown on the new UI fonts we were delivering. :0
>

I can imagine. If window manager supports device rotation then I think some
isotropic AA algorithm is needed so no surprise.

-- 
Andrew Fedoniouk.

http://terrainformatica.com
Received on Wednesday, 3 October 2012 03:10:48 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:21:01 GMT