Leonardo Uieda
Leonardo Uieda
The problem is that padding needs to be done before FFT, right?
> Yes, but the prism Fourier is computed directly in the Fourier domain, and just needs to know dx/dy and nx/ny. So really you just need to generate a padded...
> Do you think you send pad_arrays the same inputs as gridder.regular, and it returns a padded xp,yp,zp and additionally padx, pady (calling gridder.regular internally)? Or would it be better...
@drandykass wow that's way more than I expected! Great! You can open the PR whenever you want. No need to wait for things to be finished. That way TravisCI will...
Yes, that would be great!
Hi @luofangood, what do you mean by "minimum grid precision"? Regarding the ratio, please refer to [Uieda et al 2016](http://www.leouieda.com/papers/paper-tesseroids-2016.html). A higher ratio doesn't necessarily imply higher accuracy.
@birocoles, @eusoubrasileiro what do you think? This is something that we'll have to discuss going forward with the project.
@eusoubrasileiro > A Python package for modeling and inversion in geophysics need to expand from this. Well at least I think this is the core of fatiando, isn't it? Are...
> @leouieda you are the owner! :+1: I needed to think more before writting. > I will try to help... but your points are much better. The whole point of...