Thread: How can I do a limited/masked gradient fill in GIMP?

@sambrookjm: Unfortunately I think I am doing something wrong as your script just makes a large circular gadient centred in my image.

@RobA: That is actually the approach I have been working on, but I am none to keen about fiddling with the nodes as Inkscape has a tnedency to crash when I start working on things like that. The other thing I noticed is that I would need to create a 'mid-way' guide contour if one of my shapes is especially complex. In the test image I provided the inner shape has a curved 'indent' that needs to be guided to the outer shape, else one gets high contrast edges/ridges forming. I hope that makes sense.

Recently I have been looking into using Wilbur. I just wish Wilbur allowed the use of layers, so far I've been exporting as greyscale bmp's and combining them in Gimp as layers.

Anyways, back to it for me.

Originally Posted by Thurlor

@sambrookjm: Unfortunately I think I am doing something wrong as your script just makes a large circular gadient centred in my image.

Try making the white part transparent (color to alpha). That way, you'll only get the shape to have the gradient.

If you make the white background transparent, it will look like ths:



As I said, it's not a perfect script, but it is a starting point. Maybe a select by color instead of a select by alpha would be better?

@sambrookjm: Now I see where I was going wrong. After the month winds down and I'm not messing with a challenge entry I'll look into again.

Originally Posted by Thurlor

@sambrookjm: Now I see where I was going wrong. After the month winds down and I'm not messing with a challenge entry I'll look into again.

There were a couple of errors in the first version of the script, mainly when the layers weren't the same size as the image. I've fixed that, and also given the choice of either doing the entire alpha channel of the layer, or just selecting a specifc color. Use this new version of the script when you get around to it.

JMS-ContourLines.zip

@sambrookjm: Once again thank you.