Enblend and verdandi comparaison and issues

Currently verdandi has no other options.
Verdandi is using a hard seam (no blending). So I assume the differences in the sky exist already in the original images. Maybe a better photometric optimization could help to reduce the problem.

When you see bigger color differences in the fast preview window then enblend is the better choice. But when the fast preview window looks already (nearly) good with no big color differences between the images, verdandi does a good job.

Thanks for your answer. This was the case, difference already exist in the source files.

I want to experiment with some better blending algorithms. Can you provide the source images for testing?
But I don't promise anything.

I'm curently on bolides. I'll Provide pictures when i'll come back home.

I'm working currently on an algorithm and would need some more test cases for testing and optimization.
Can you still provide the test images?

Yes i can,

You can download an archive here (part of the full pano) : http://pmllc.free.fr/panophoto/F.7z

Thanks. It helped to fix some little bugs in my code.
The default repository contains now code to blend the seam. It can be selected on the stitching tab, under options of internal blender.
So the seam becomes less visible (ideally in-visible), but at cost of longer processing times.