Originally posted by dlh Well, I see your point, dudes, but to me, real pixel shifting involves the acquisition of new and different data. Interpolation doesn't do that, it simply averages the pre-existing data.
Lightroom and Raw Therapee as well as Pentax's own DCU combine the four images captured, assuming the pixel shift as specified by the camera. This works superbly on all of these platforms -- at least as well as it does when processed in camera.
The difference come in how the software packages deal with movement within the scene. Lightroom doesn't. It instead creates artifacts in every one of those spots. DCU masks in the first image of any artifacted areas it detects automatically when you select the "motion correction" option. Raw Therapee allows you more options for the motioned areas -- an average feature and the ability to choose which of the four images you want as the "base" image. You can also see which areas are being masked and which ones there is no motion in.
Long story short, for images where there is much movement, the majority of the image is going to be whichever base image you choose and pixel shift gives you no benefit over a standard image, but for images where things are relatively still, it gives quite a bit of improvement over a standard image.
Dynamic pixel shift is a different story and as far as I know can only be processed in camera.