Re: [cairo] issue with blend modes in pixman
Rik Cabanier <[email protected]>
| Newsgroups | gmane.comp.graphics.pixman,gmane.comp.lib.cairo |
|---|---|
| Message-ID | <CAGN7qDCteBdMiGEB9-ciWsN389++Y8aYU6+7vmPy0UsQkDnyYw@mail.gmail.com> |
On Sat, Jan 5, 2013 at 11:07 PM, Adrian Johnson <[email protected]>wrote: > On 01/01/13 07:35, Rik Cabanier wrote: > > Some of the blend modes are not producing correct results. > > You can find a diff with the incorrect results > > here: https://bugs.webkit.org/show_bug.cgi?id=105074 > > > > Looking at the formulas, I can see what's wrong but I don't know who to > > contact. > > Could you tell us what specifically is wrong with the formula? > I was actually seeing bugs with blending in Core Graphics and not in pixman. I got pulled away from looking at blending in cairo. Hopefully, I will get back to it later this week. > > The webkit bug suggests there is a difference between the SVG and PDF > formulas. At the time the blend modes were added to pixman the formulas > were based on the PDF specification and the SVG working group said the > SVG formulas should match PDF [1]. Are the PDF formulas wrong? Or do we > need CAIRO_OPERATOR_COLOR_BURN and CAIRO_OPERATOR_COLOR_BURN_SVG if we > are to support both PDF and SVG renderers? > No, the SVG operators should match the PDF ones. It would be a mistake to have 2 different formulas for the same blend mode. The confusing part if that the PDF (and CSS) blend modes are done in non-premultiplied space and produce a color before you composite. > > [1] http://lists.cairographics.org/archives/cairo/2008-October/015362.html > > _______________________________________________ Pixman mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/pixman