Re: How can we improve quality of text content rotated using CSS transform?

Pascal Germroth wrote:
> On 26/07/09 01:27 Biju wrote:
>> How can we improve quality of text content rotated using CSS transform.

I don't see how this falls within the scope of the list.  It seems to me 
that it is a purely implementation issue.

> 
>> In this case cant we re-use CSS property  [...]
>>    text-rendering: optimizeQuality

I would assume that this property is just a hint to the rederer to be 
more concerned about quality than time, not something that forces the 
use of some special algorithm.

> 
> Especially when using animations, speed can be improved by rendering the 
> text to a texture and rotating it (especially when using 3D-transforms), 
> opposed to transforming the glyph paths and rendering them using 
> sub-pixel anti-aliasing/ClearType.

Interestingly, when W3C's Amaya editor/browser started using 
anti-aliasing, there were a lot of calls to turn it off because people 
objected to the "fuzzy" characters.  This also comes down to an 
implementation issue, in that it assumes hardware that can represents 
multiple intensity levels.

I wonder if the difference here is between looking good and being easy 
to read.
> 
> Now however, it seems as if rotation angle had some influence on 
> kerning, in Firefox, I've seen letters jump by some amount around their 
> supposed positions when rotating, but I guess this would be a bug in the 
> browser where it didn't use sub-pixel positions for the glyphs.

That's what one would expect for a hinted font implementation. Hinting 
tries to optimise placement with respect to pixels, which would cause 
characters to move to achieve that.

With regard to Visio, it is quite possible that it uses a patent 
encumbered algorithm.

Although the Amaya complaints were for unrotated text, I think, given 
that designers like using 7x5 and 6x5 fonts, rotated text is either 
going to be significantly distorted or is going to lose a lot of contrast.

-- 
David Woolley
Emails are not formal business letters, whatever businesses may want.
RFC1855 says there should be an address here, but, in a world of spam,
that is no longer good advice, as archive address hiding may not work.

Received on Sunday, 26 July 2009 08:44:58 UTC