-
Notifications
You must be signed in to change notification settings - Fork 27
Open
Labels
featuregood first issueIndicates a good issue for first time contributorsIndicates a good issue for first time contributors
Description
We want to integrate the "exponential of semicircle" window proposed in this paper: https://arxiv.org/abs/1808.06736
I am not completely sure but my FINUFFT wrapper indicates that it is slightly more accurate than Kaiser-Bessel.
Remarks:
- The window is completely simple
- Its Fourier transform seem to have no analytical form and therefore it is approximated using Gauss-Legendre-Quadrature (equation 3.10 from the paper)
- The implementation in FINUFFT is here: https://github.com/flatironinstitute/finufft/blob/master/src/finufft.cpp#L227. But if I get it right they applied some caching tricks that result in this https://github.com/flatironinstitute/finufft/blob/master/src/finufft.cpp#L167 But these implementations should only be used for debugging purpose, the mathematics is straight forward and permits a cleanroom implementation.
- Both functions require the Gauss-Legendre nodes, which are calculated here: https://github.com/flatironinstitute/finufft/tree/master/contrib Of course we would not use that but instead go for a Julia package. A quick search pointed me at https://github.com/JuliaApproximation/FastGaussQuadrature.jl
This would be a nice "good first time contributor" project if somebody want to have look at it. Otherwise I will five it a go when time permits.
ChenZhu-Xie
Metadata
Metadata
Assignees
Labels
featuregood first issueIndicates a good issue for first time contributorsIndicates a good issue for first time contributors