Originally posted by Oakland Rob Maybe I'm misinterpreting what you're saying, but if you look at the slider image at the bottom you can see that the DCU version does eliminate the screen artifacts (good description). But it also has far less detail in other areas, which is one of the reasons some of this the algorithm is simply "when you detect artifacts over X threshold use first frame only." I.e., it doesn't eliminate motion artifacts without chucking the whole high res baby out with the artifact bath water.
Thank you, I had missed that. I took a look at the large images and re-compared. What is very interesting, is that if you look at the building in the background, the PDCU images are sharper and contain more detail versus the Lightroom version. Definate pixel shifting effect as it shows improvement over the non-PS version. However, like you pointed out, the foreground rocks are much LESS sharp than the Lightroom version.
Couple possible explanations. One, the focus point or aperture was different between shots. I am going to give the photographer the benefit of the doubt here, and say this is probably unlikely as they would know it would screw up the test. My other thought was that maybe PDCU is applying diffraction correction. So I researched that and Diffraction Correction was not introduced until 5.4.2. There test used 5.4.1. So, that squashes that idea.
With all that said, I spent a couple hours working on a PS image I took over the weekend. What I found is that the movement correcting algorithm ONLY corrected artifacts in the moving water. No other areas were touched. I will post some image crops, at the latest, by the weekend. I am using the newer PDCU (5.4.2) though I see no mention of PS related enhancements/bug fixes.