Fwd: [Moderator Action (size limit exceeded)] RE: SVG masks and luminance calculations

This is a forwarded message
From: Rik Cabanier <cabanier@adobe.com>
To: Erik Dahlstrom <ed@opera.com>, Boris Zbarsky <bzbarsky@mit.edu>, Tavmjong Bah <tavmjong@free.fr>
Date: Friday, February 4, 2011, 6:46:01 PM
Subject: [Moderator Action (size limit exceeded)] RE: SVG masks and luminance calculations

===8<==============Original message text===============
Assuming you're talking about luminosity masks, the formula for calculating them is defined in the PDF spec.
I've attached the relevant pages to this mail. (Page 523 has the actual math.)

Rik

> -----Original Message-----
> From: www-svg-request@w3.org [mailto:www-svg-request@w3.org] On
> Behalf Of Erik Dahlstrom
> Sent: Friday, February 04, 2011 4:05 AM
> To: Boris Zbarsky; Tavmjong Bah
> Cc: Dirk Schulze; www-svg@w3.org; Robert Longson
> Subject: Re: SVG masks and luminance calculations

> On Thu, 03 Feb 2011 16:57:45 +0100, Tavmjong Bah <tavmjong@free.fr>
> wrote:

> > On Thu, 2011-02-03 at 10:09 -0500, Boris Zbarsky wrote:
> >> On 2/3/11 4:39 AM, Dirk Schulze wrote:
> >> > The most(unsure about Safari/Win) ports of WebKit use linearRGB as
> >> default now
> >>
> >> As far as I can tell, Safari/Mac uses linearRGB and Safari/Windows
> >> uses sRGB (which certainly confused the person who reported a bug on
> us).
> >> Chrome uses linearRGB, in my experiments.  Gecko uses linearRGB.
> >>
> >> Also, Inkscape apparently uses sRGB while Batik uses linearRGB, for
> >> what that's worth.
> >
> > Inkscape does use sRGB at the moment but I was planning on fixing it
> > in the next release.

> Forwarding a message from Jos Hirth:

> color-interpolation="sRGB"-ish:

> - Inkscape
> - Opera
> - Safari 4/5 on Win32, Mobile Safari on iOS
> - IE9


> color-interpolation="linearRGB"-ish:

> - Firefox
> - Batik
> - Chrome/Chromium
> - Safari 5 on Mac (really odd)


> My use case for sRGB:

> http://kaioa.com/b/1102/svgjng/index.html (use Opera ;))

> I basically emulate JNG's functionality with SVG here. As you can see the
> quality/size ratio is pretty impressive. The advantage becomes even bigger if
> the images are larger and photo-based (PNG8 becomes quickly unfeasible).

> The SVG format is a lot more powerful than the JNG format. There are a few
> more things one could do to improve the size/quality ratio even further.
> There is no need to use a single image for color and a single image for the
> alpha channel. E.g. the upper 2/3 of the Jetpack mask could be more
> efficiently stored as PNG.


> I still don't know why one might want linearRGB. It got that psycho-visual
> stuff going for it (doesn't really help if it's not supported by Inkscape), but
> that's about it. The main downside I can see is that it negatively affects
> antialiasing (e.g. if the mask is a fully opaque rectangle with some path
> drawn onto it).

> I couldn't find out yet what Adobe Illustrator is doing.

> --
> Erik Dahlstrom, Core Technology Developer, Opera Software Co-Chair, W3C
> SVG Working Group Personal blog: http://my.opera.com/macdev_ed


===8<===========End of original message text===========


-- 
 Chris Lilley   Technical Director, Interaction Domain                 
 W3C Graphics Activity Lead, Fonts Activity Lead
 Co-Chair, W3C Hypertext CG
 Member, CSS, WebFonts, SVG Working Groups