4D output should crop center

[apryor6]

The current 4D output results in many 2D images, each of which includes the anti-aliasing aperture. The result is that much of the array is wasted space (zeros). Only the central region should be returned.

[matkraj]

4D output also has swapped quadrants

[matkraj]

I am not 100% sure how to do this, but here is a suggestion for change to crop and swap quadrants for 4D STEM - I am not sure this is not breaking any other data saves
matkraj@d7c0a98#diff-d87e66171dfb2a89f41ff2713110710f

[apryor6]

The quadrant swapping functionality is already in Prismatic in the function fftshift2 within include/utility.h. Also, the Prismatic array class has a to_mrc method to write out the data, so all you should need to do is create a new Prismatic array, fftshift the large one, and then copy the center into the smaller array before writing

[matkraj]

my code was changing toMRC_f method for this. Thanks for fftshift2 suggestion

[cophus]

I prefer the default FFT quadrant positions because the origin pixel is always at (0,0). I agree though that at some point we should definitely change the 4D output to only save the portions inside the anti aliasing aperture.

[apryor6]

I think the solution would be to have only inside the anti aliasing aperture saved always, and then to add an additional option that will shift the quadrants if the user desires

[thomasaarholt]

Anyone opening these files in python can achieve the shift with np.fft.fftshift(pr.fileio.readMRC(filename)[0]). The [0] is required because the readMRC function assumes that the mrc file can be a stack, I assume, and the file only contains a single image.

[lerandc]

Outputs are now cropped and fftshifted by default.