Pixel-shift and superresolution imaging

[kaweis]

There is currently quite a bit of excitement around the sensor-shift technology that will be implemented in the new Pentax K-1. Based on the posted samples (also including the ones for the K3, e.g see diglloyd.com or http://www.imaging-resource.com/PRODS/pentax-k3-ii/pentax-k3-iiTECH2.HTM) this looks indeed quite promising and might significantly enhance the resolution at least for reproduction or landscape work.
In principle, such a superresolution approach could also work for other cameras if the entire camera is slightly shifted between exposures (e.g. see http://petapixel.com/2015/02/21/a-practical-guide-to-creating-superresolution-photos-with-photoshop/). I gave this poor man’s pixel-shift a try and mounted my D800E on a tripod equipped with a rail and shifted between each exposure by very small amounts (< 1mm) to the left or right. When I combined up to 15 exposures in Photoshop (using a median filter as described on petapixel.com) very little improvement could be seen. By contrast, using a trial version of Photoacute gave a nice resolution enhancement with as little as 4 exposures (in fact adding more exposures did not really seem to help for my test images). So far so good. Unfortunately, Photoacute seems to be no longer supported or further developed, and no new camera/lens profiles will be added according to their website. Is there anything else out there that would produce high quality supperresolution images from multiple shifted exposures? Google gave me some interesting research papers on this topic but most of the math is quite a bit beyond me, and I am definitely not in a position to write codes that would implement these ideas. Any other ideas?

[Tim Lookingbill]

I was investigating that technology online a while back since I've been looking to upgrade my Pentax K100D, but I have some reservations going by the A/B image comparisons in that linked article which I do remember coming across in my research. It appears in the before samples the camera system doesn't provide very sharp results with the increased resolution unless that soft looking before image was intentionally made to appear blurred due to poor focus metering. That's what it looks like to me.

I can't be sure but I know that I can get the same sharp results shown in the after image sample with my 6MP Pentax K100D. I don't know if all those image samples are 100% crops though. At a certain distance my 6MP DSLR can't resolve fine fabric threading textures and it has not been made clear how close those shots were taken in those image resource examples.

Something isn't adding up with regard to the relationship of higher resolution and sharpness.