Originally posted by wkraus If you have a file like this it is perfectly obvious if all subframes are used or not, although this does not answer the question about motion correction.
Pixelshift (left) vs. no Pixelshift (right)
If you want me to I can send you the .dng to play with. This can be interesting – in another thread we discussed the observation that with the K-3 III the Image Transmitter 2 software for tethered shooting combined with Pixelshift produced files with only three of the four subframes readable, which were then processed without PS in ACR and RawTherapee. To my knowledge this is still not fixed.
Sure why not, I can do some experiments with DxO. Meanwhile, here's a high-iso pixelshift that I edited with DxO. Compared o other non-pixel shifted images I took, it seems to make a difference in how effectively I can denoise. It's an unusual use-case, nevertheless it looks like PS can be useful for some odd type of low-light/night shooting.
Even if, maybe in terms of noise reduction I would have got the same result by stacking frames; but it's handier in a single file.