Weird Moiré pattern in a registered image

[bretm]

I am trying to process some DSLR problems I took this weekend, and the output has a weird pattern in them.  I normally use PI with CCD images.  This is my first attempt with a DSLR, so I suspect I'm doing something wrong, but don't know what.

The simplest repro case is:
- convert 2 CR2 images to FITs using BatchFormatConversion
- Register 1 of those images to the other (use 4196 as the reference, 4171 as the image to be registered)
- look at the registered version of 4171 to see the problem
- optionally, for a more psychedelic version, debayer the image.

A real workflow would include bias, darks, flats, etc. but since the problem shows up without them, so I'm leaving them out here.

Two CR2 files, and a sample of the bad output can be found here:

https://www.dropbox.com/sh/u1n8zaa4dag4vsm/S-oeM9GtaG

If it matters, these images were taken with a Canon EOS 5D MkII + 24mm lens on a non-tracking tripod.

Thanks,

Bret

[pfile]

you need to debayer before registration, it sounds like you did not do that...?

normally the flow is to calibrate the images in RAW mode, then debayer, then register, then integrate.

[bretm]

Thanks for the suggestion -- that was my problem.