Swing/Shift

[Craig Gerard]

I'm playing around with a Novoflex T/S bellows and found the following observation interesting. It won't come as news to most, but I thought it worthwhile uploading.

First image is with no adjustment applied to the bellows.

The second image is the result with swing/shift applied to the camera end of the bellows. Minor but noticable difference.

The third image was taken with swing/shift adjustments applied to the lens end of the bellows.

Canon 50D, SK Apo-Digitar 80/4

Image #1:


Image #2:


Image #3:


Craig
_________________
To use a classic quote from 'Antz' - "I almost know exactly what I'm doing!"

[enricosavazzi]

If I remember my geometry correctly, shift applied to the rear standard (camera end) is completely equivalent to shift applied in a reverse direction to the front standard (lens end), ignoring in this discussion the need to reposition the camera relative to the subject.

If one applies no shift, tilt of the rear standard of course produces different results from tilt of the front standard. This is probably what you are observing in the pictures. However, by combining tilt of the rear standard with shift of either standard one can completely duplicate the effects of tilt of the front standard, within the limits of the available amount of movements.

At least in principle, and in a geometric sense, it is not necessary to have tilt/shift capabilities at both standards, and with sufficient amounts of tilt and shift available at a single point between lens and camera (and equivalent movements available for repositioning the camera-lens complex relative to the subject) one can duplicate the same effects available in a camera with movements at both standards.

It is of course convenient to have movements at both standards - sometimes very convenient.
_________________
--ES

[Harold Gough]

Craig,

An impressive demonstration.

A question I have to ask myself, having recently invested in such bellows, is, in macro and close-up, as opposed to the scale of buildings, etc., how many times could a similar effect be obtained by a change in position of the camera and/or of angle of the long axis of the lens in relation to the subject?

harold
_________________
My images are a medium for sharing some of my experiences: they are not me.

[enricosavazzi]

[Harold Gough]

[elf]

[Harold Gough]

[rjlittlefield]

Harold, search on "Scheimpflug principle".

The trick is to get the subject plane, lens plane, and sensor plane to all intersect in a single line. If you can do that, then everything will be in focus.

Tilting the sensor plane will keep you in the sweet center of the lens field but will stretch & keystone the image compared to the non-tilted version.

Tilting the lens plane will move the sensor off-center in the lens field but will retain the original geometry (barring lens distortions). It can also offer a wider range of adjustment, especially at higher magnifications where the lens is relatively much closer to the subject. You can see in this example that the same subject tilt is handled with much less lens tilt (left) than sensor tilt (right)



If you're pushing the envelope of the lens then the best solution might involve tilts at both ends.

Craig, in #2 was the problem that you didn't have enough tilt range to accomplish the focus you wanted?

--Rik

[Harold Gough]

Thanks, Rik.

I'm going to have to find some very understanding subjects!

Harold
_________________
My images are a medium for sharing some of my experiences: they are not me.

[Craig Gerard]