Originally posted by hjoseph7 Too bad Pentax didn't give the sensor shift a little more leeway 11mm instead of the measly 8mm, it's a serious tool as we can see from these images.,
Do we actually know how much physical space the sensor is capable of moving? The sensor (K5) is 23.7 x 15.7mm in size. An 8mm displacement in either axis would be extremely large. 8mm in the X axis would be 33%, with the Y axis yielding on the order of 50%.
The old K 28mm/f3.5 Shift lens was capable of 11mm in all directions (Y, Y and X+Y). That
amount of movement translates (see the shift calculator) to ~70% in the vertical and ~45% in the horizontal. Using the images of the Tokyo TV Tower -
I would guess that the amount of additional space is probably a bit less than 10%. Regardless, of what the actual numbers are, or the absolute amount of the shift, the bottom line is - is it useful? Clearly it is. Its not a substitute for a shift lens, but it certainly helps, and it has a positive affect. If you know that you 1) have the tool available; and 2) a rough idea as to how much it can aid you; you can apply it for real useful purposes.
A couple of months ago, I was down at the state capital. I wanted to take some night pictures. I plopped my tripod down and mounted my camera, turned it on and went into live view. I remember putting the grid center line down the middle of the building, and thought - that wow, I should have brought the shift lens. That building is really leaning backwards - in danger of tipping over. I had not leveled camera all the way. So, I started adjusting with the geared head - axis by axis. It was a real visual learning experience, to just dial in the axis corrections by watching the Liveview as the building automagically corrected its framing, up against the Liveview's grid and h/v level indicators. The building's perspective was corrected with just adjusting the physical geometry of the camera as it related to the building. I have done this previously, just using a ballhead. But making the adjustments with a ballhead are not nearly as noticeable or as easily associated with each individual axis.
Even though the sensor's shift amount is not a large amount, as has been shown, it can have a reasonable impact on the image. Coupled with the relationship of how the camera addresses the structure (tripod, framing, alignment, etc.), it certainly is an additional tool that can be used, with out hauling around specialized lenses.
_____
There is another useful purpose for shifting the sensor. There are times when you really do not want to introduce any rotation in to a stitch. Normally, you just twist slightly to the left, take a frame, then back to the center, take a frame and then to the right. When you stitch them together, you get some geometric distortion as the images are overlaid. I do like to use the shift lens to "shift" over and grab some extra view. This is a lot more work - moving the sensor, but it you need it, it is nice to have available to you. It is not going to be a tremendous amount of additional area, but sometimes it will probably be just enough.
_____
All of this use to be painfully apparent to the folks who use/used view cameras. They have all of these adjustments available to them. The physical knobs and adjustments are sitting there right in front of them - all the time. With the advent of the brownie box camera up through the dSLR, we have lost a lot of this to varying degrees - only to rediscover it when technology puts it back into the available mix (or the photographer goes searching it out to re-acquire it, i.e., buying a tilt/shift lens, or getting a set of bellows).