Originally posted by Wasp This is just pixel shifting on a bigger scale. There is not enough movement in the SR mechanism to cover the entire full frame but images will be bigger and better. That is a Good ThingTM. No movement allowed of course, but landscapes don't move around much.
A few months after I got my K-3 I did that sort of thing manually with the image composition editor function and would shoot 5 to 9 shots and combine in Photoshop and in some cases I could see it being worthwhile but a lot of the time the marginal expanded field of view just wasn't worth the effort. I went and looked things up and the max movement on the K-3 sensor appears to be 3mm stop to stop in either the horizontal or vertical direction (+/-1.5mm from center) so that would give about 15% to 20% more in each direction so no where near the 50% more needed for full frame. If it were automated I could see myself possibly using it but seemed like a feature that never did it for me and I haven't really used it since. There is one shot I want to use it on but that is one where I need about 5% more than I can get with my zoom that go down to 70mm so would probably use the 77 ltd instead and composition editor for that shot.
I was still at the stage of playing with features and really big into computational photography at the time so thought it might be worthwhile exploring. As such I played, explored, experimented and then got bored because it just didn't give a big improvement. Same thing with playing with super resolution images. They both have their place but it is far rarer than you think, especially stitching composition editor images. In both cases I find that going to a different lens size and doing a properly stitched panorama to be a better solution. Like using my 50mm instead of the wide or ultra wide, or using the 35 instead of the 28 + image composition.