W3C home > Mailing lists > Public > www-style@w3.org > April 2007

Re: CSS Opacity

From: David Woolley <forums@david-woolley.me.uk>
Date: Sun, 29 Apr 2007 10:26:34 +0100
Message-ID: <463464CA.4060205@david-woolley.me.uk>
To: www-style@w3.org

Daniel Beardsmore wrote:
> the alpha channel bloats the format more. (The Web image formats leave 

A PNG alpha channel ought to be highly compressible, as one would expect 
long runs of the same value.

> so much to be desired. Imagine, for example, if you could assign 

SVG is a web format that supports alpha and can include JPEG images.
Unfortunately, probably for marketing reasons, it is not supported by IE 
and plugins are not supported as img elements.

> multiple GIF image colours to varying transparency levels to get, say, 

That wouldn't of course, be GIF any longer, just an LZW compressed 
palettised image.  In any case, PNG *already* supports this mode of 
working <http://www.w3.org/TR/PNG/#6AlphaRepresentation>, although it 
might not be well supported by image manipulation tools and I don't know 
which if any browsers support it.  An indexed RGBA PNG ought to compress 
better than a hypothetical GIF derived one. (I don't think it is 
particularly well known that PNG supports palletised images.)

> 250 colours plus alpha in a mere 8 bpp for logo edge anti-aliasing? Or 
> DCT lossless alpha channels on JPEGs, or all sorts :)

My understanding was that DCT was always lossy (cosine is a 
transcendental function, and compression ratios were poor if one 
operated as close to lossless as reasonable.
Received on Sunday, 29 April 2007 09:27:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:50 GMT