399133 - svg feColorMatrix in saturation mode has a math error on the blue channel

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Drew Perttula]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.3) Gecko/20070309 Firefox/2.0.0.3
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a9pre) Gecko/2007100704 Minefield/3.0a9pre

http://lxr.mozilla.org/seamonkey/source/content/svg/content/src/nsSVGFilters.cpp#1346
looks wrong:

1346     colorMatrix[12] = 0.715f - 0.715f * s;

I think that should be  0.072+0.928s  according to
http://www.w3.org/TR/SVG11/filters.html#feColorMatrix

The error result is that saturate with values="1" looks way yellow, and
values="0" looks somewhat blue.

Here a screenshot of values="0" in minefield 3.0a9pre 2007100704:
http://bigasterisk.com/post/bugzilla363909-compare.jpg
(The lower image is in batik and looks right to me)


Reproducible: Always

Steps to Reproduce:
1.
2.
3.



I initially put this as a note at https://bugzilla.mozilla.org/show_bug.cgi?id=363909#c13

Comment 1

[Robert Longson [:longsonr]]

Could you attach the svg file you used to generate the jpg please so that we have a testcase.

Comment 2

[Drew Perttula]

Attachment: test case- three saturation levels

Comment 3

[Robert Longson [:longsonr]]

Attachment: patch

Comment 4

[Drew Perttula]

Attachment: output of sat-test.svg in firefox vs batik

Comment 5

[Drew Perttula]

Attachment: demo image used in sat-test.svg

Comment 6

[Robert Longson [:longsonr]]

Attachment: correct testcase to work in bugzilla

Comment 7

[Robert Longson [:longsonr]]

With the patch the output looks like the batik version.

Comment 8

[Robert Longson [:longsonr]]

checked in