On Monday, July 23, 2007, 8:33:41 PM, Felipe wrote:

FS> I have been working on feDisplacementMap but I havent figured out FS> a way to test my first drafts of code since the displacementmap FS> test on the SVG 1,1 full test suite also depends on feImage. That FS> sucks!

Fair comment. feImage is the most usual way to give a displacement input to feDisplacementMap, but its true that its possible to test without, by making the object being filtered also be the source of the displacement.

So here is a test, which will likely be in the next release of the 1.1 test suite, which tests feDisplacementMap by itself. Also an image (from batik) which indicated the expected result.

FS> So, I stopped working on fedisplacementmap and started to FS> implement feImage. The specs say that feImage should behave FS> similarly to the image element. So, I imagine that it may be FS> possible to not duplicate code. But I tryied to look at FS> sp-image.cpp and I dont know exactly how to reuse it. Could anybody help on this?

I suggest checking that the image code handles gamma and color space conversions correctly. For the image element by itself, incorrect gamma will just make the picture a bit too light, or too dark; when used as feImage input to feDisplacementMap, incorrect gamma will be much more noticeable by giving the wrong geometry.