You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm interested in implementing a Lanczos resampling algorithm here. This appears to be simply the convolution of the Lanczos Kernel by the samples.
I created a fork and started developing, but I have become completely overwhelmed by the complexity inside the library. I have no idea if such an interpolator should be its own interpolation type, or if it should be a Gridded type. I also have no clue how to interface with, well, anything here.
Can i get some advice on how/where to implement this algorithm (and whether you want it at all!)?
The text was updated successfully, but these errors were encountered:
First of all I don't now much about the internals, but after consulting the DevDocs (part of #365) I think that it's not easy to do this with the Gridded directly, because it somehow built to work with polynomials rather then with convolutions.
In terms of implementation it's probably a use case for FFT, at least when the window gets large (which is difficult in case of non-equidistant data). And actually it would be good to have some ConvolutionInterpolation, where you can define arbitrary window functions. But again, I'am not an expert regarding this package (which is looking for a new maintainer #365)
I am definitely interested in adding FFT based interpolation schemes to this library. I'm getting up to speed on the library itself. I will get back to you.
I'm interested in implementing a Lanczos resampling algorithm here. This appears to be simply the convolution of the Lanczos Kernel by the samples.
I created a fork and started developing, but I have become completely overwhelmed by the complexity inside the library. I have no idea if such an interpolator should be its own interpolation type, or if it should be a
Gridded
type. I also have no clue how to interface with, well, anything here.Can i get some advice on how/where to implement this algorithm (and whether you want it at all!)?
The text was updated successfully, but these errors were encountered: