[Painter]

Recently someone asked me if I had used the composition adjustment to frame a scene where I took two images from the same location years apart. Because so much of getting the scene right was about camera heights and angles I did not use it. It did get me thinking about it's purpose and how it differed from just moving the tripod though and so I put together this post on my blog that explains it. I hope that someone may find it useful.

Pentax K-7 Sensor shift (Composition adjust) | Wallace Koopmans Artlog



I realize I'm a bit of a newbie here at Pentaxforums but I have been taking pictures for more than half my life and own more camera's than I actually know.

[Adam]

Should I copy this into the article section? Pentax Articles - PentaxForums.com

Thanks for posting this!

[Painter]

Sorry if I posted to the wrong place, as I said I haven't posted much here so if it should be moved or placed somewhere else in addition, by all means do. Thank you.

Should I copy this into the article section? Pentax Articles - PentaxForums.com

Thanks for posting this!

[Eruditass]

Wow, now I understand why the shifting in tilt/shift photography is different than simply stretching my arms out upward.

[ragingtornado]

I always wondered why you wouldn't just move the tripod. Now i understand! Thanks for the link

[panoguy]

Well, as long as your full explanation goes to the right place. And truly, it was your images with the ball and blue block that were quite instructive... but alas, my K20D does not have this function, so I'll deal with a shift lens.

BTW, excellent work on your artblog, Painter! Love scratch n' print.

[max889]

Very Interesting! So this, coupled with bigger image circle lenses (which every 35mm lens is on an APS format sensor), would provide the possibility of shooting two perfectly stitchable photos, the way a t/s lens does. Any idea about the area differential got from the sensor displacement? It could add some lateral megapixels for very good landscape scenes, hardly enough to call it a panorama, though, but great to take advantage of high resolution full frame lenses.
THis could also provide some perspective correction for converging lines in buildings or am I lossing it here?

[Painter]

Pentax claims the sensor can be moved about 1mm in each direction (I believe it's slightly more when the sensor isn't rotated) so a quick calculation says that a 23.4mm sensor would stitch together as a 25.4mm sensor. That's only an 8% increase not much I'm afraid. In my test I used the DFA 100 Macro WR so it covers a larger image circle

The action of moving the sensor directly relates to sliding the rear of a view camera and has no tilt action. Unfortunately this cannot be used to effect diverging lines, also shift of the lens while leaving the sensor stationary is something different again and requires a Shift capable lens.

In practice this feature is limited, I would love to see the ability to tilt the sensor but then there would be all kinds of post saying the bottom half of all my images are blurry ect.

Now I'm going to have to go back and see just how big a file you can produce, likely with four images stitched

Very Interesting! So this, coupled with bigger image circle lenses (which every 35mm lens is on an APS format sensor), would provide the possibility of shooting two perfectly stitchable photos, the way a t/s lens does. Any idea about the area differential got from the sensor displacement? It could add some lateral megapixels for very good landscape scenes, hardly enough to call it a panorama, though, but great to take advantage of high resolution full frame lenses.
THis could also provide some perspective correction for converging lines in buildings or am I lossing it here?

[siva.ss.kumar]

Can this technique be used to take 3-D pictures? If this shift can kind of simulate a left and right eye view, with proper filters applied, I think we can take 3D pictures. Or am I missing something big time ?

[Painter]

Can this technique be used to take 3-D pictures? If this shift can kind of simulate a left and right eye view, with proper filters applied, I think we can take 3D pictures. Or am I missing something big time ?

Unfortunately the answer is no you cannot make a stereo pair by moving the sensor. It is actually the opposite effect. Part of our visual system works such that we perceive depth because of the disparity between the left and right eyes. The sensor shift is as if you had one eye and could move the receptors behind it. It's funny you should bring that up, because the images I took where I moved the camera do make a convincing 3D image. I use a device called a Pokescope that allows you to view stereoscopic pairs easily and the form of the ball convincingly stands out. While the sensor shift image just looks flat (although it appears to be a flat image behind the screen) as if the edges of the image were a window.