ppSub non-convolving option request

[Michael Wood-Vasey]

I would be useful to be able to run ppSub with an option for no convolution and a separate option for no flux matching (this second depends on whether or not the convolution kernels themselves are supposed to preserve flux).

The practical use would be to compare images taken in near succession, both at the same dither position and different to explore the focal plane and warping.

But the main motivation is to allow for simple tests and comparison within the framework of the IPP. I.e., all of the stuff gets written out right and in the same convention. While most of the operation of ppSub is the kernel and convolution, most of the code and thus most of what is hard for people not on the IPP team to do, is doing all of the bookkeeping.

Tests

• Verify that subtracting an image for itself result in zero.
• Verify that subtracting two images taken sequentially in stable conditions subtract well.
• Verify that a self-stack of 4 copies of the same image subtracts perfectly from that image
• As a self-check, make sure that using the standard fully-enabled ppSub doesn't do a worse job than using a non-convolving, flux-matching ppSub.

You can never have too many (motivated) tests.

[Paul Price]

Flux normalisation is done as a part of the PSF-matching, and cannot easily be disentangled.

All these tests can be performed using ppArith. Once the kernel calculation and application are disabled, ppSub is effectively like ppArith.