Open gr5 opened 1 week ago
I'm guessing the gaussian blur is widening the image by the number of pixels in the convolution and this isn't getting cropped at the edge of the mirror so somehow the ronchi code is including this extra data in it's calculations.
Version 7.2.3 but I think it does it in many older versions as I've seen this before
To Reproduce I processed igrams from a real mirror. I'll include pictures of the issue and a picture of a ronchigram of the real mirror.
I smoothed the wavefront with "tools" "smooth current wavefront" although that's not required to see the problem.
I created a perfect simulated wavefront and the same issue happens: looks bad if gaussian smoothing is on.
GOOD:
BAD:
ACTUAL: