Fourier Transforms

rascil.processing_components.fourier_transforms.fft_coordinates Module

Support for coordinates in FFTs

All grids and images are considered quadratic and centered around npixel//2, where npixel is the pixel width/height. This means that npixel//2 is the zero frequency for FFT purposes, as is convention. Note that this means that for even npixel the grid is not symmetrical, which means that e.g. for convolution kernels odd image sizes are preferred.

Functions

`w_beam`(npixel, field_of_view, w[, cx, cy, ...])	W beam, the fresnel diffraction pattern arising from non-coplanar baselines
`grdsf`(nu)	Calculate PSWF using an old SDE routine re-written in Python
`coordinates`(npixel)	1D array which spans [-.5,.5[ with 0 at position npixel/2

rascil.processing_components.fourier_transforms.fft_support Module

FFT support functions

Functions

`ifft`(a)	Fourier transformation from grid to image space
`fft`(a)	Fourier transformation from image to grid space
`pad_mid`(ff, npixel)	Pad a far field image with zeroes to make it the given size.
`extract_mid`(a, npixel)	Extract a section from middle of a map
`extract_oversampled`(a, xf, yf, ...)	Extract the (xf-th,yf-th) w-kernel from the oversampled parent