2D Azimuthal integration vs Polar transformation

[Marcraven]

I have been playing around with the 2D AzimuthalTransformation of pyFai just to be extremely suprised by the staggering difference in high frequency preservation with respect to e.g. skimage polar transform.
pyFAI 2D AI:

skimage warp_polar:

Is this behavior expected? I know the units are different (yes the axis names are wrong) but the 2D integration erased all the texture.

Cheers

[kif]

I have to admit this is puzzling ... can you provide the code and the image ?
Your azimuthal range looks really strange, how many azimuthal bins are you using ? Which pixel splitting scheme are you using ?

In this notbook, there is an example with peaks of less than 4 degrees of azimuthal extent which are properly reduced.
Untitled2.pdf