Tilt shift questions. Planning 28mm f/3.5 mods

[CardBoardBoxProcessor]

I'm thinking of turning the lens into something more like jborn's(sp?) And giving you the ability to have full tilt shift ability sort of like a view camera but smaller. Inspired by his Toadstool pictures I like that sort of thing with the x1.5 macro and even able to get things in the far distance InFocus because of the lenses ability to tilt so far.

My question is what would be better for this setup to have front tilt and shift of the lens and the camera is stationary or if possible make it so that the camera moves and the lens stays still on tilt and shift.

Or if possible should I make it so that the lens and Camera can both shift until separately.

What are the benefits of each? Is there any difference between tilting the camera versus tilting the lens.

[Bjørn Rørslett]

You will need to combine tilt front/rear and shift probably mainly in front to get the full range of movements.

If the front is not tilted and the rear (ie. camera) is, the tilted sensor will  be unevenly exposed. In the film days, deviations of more than 2/3 EV was generally not acceptable, whilst in the digital era you could get away with significantly more unless the camera's mount and/or mirror box hit the frame with heavy vignetting first. Thus, shifts are often need to bring the projected cone of light within acceptable boundaries to avoid mechanical vignetting.

[CardBoardBoxProcessor]

Well I figure I could have the camera or lens move as much as I need as it'd be custom equipment. Just wonder if I should

Option A is a lens that does all tilt and shift

Option B is the lens is always static and camera does it all. I could have it have enough movement with shift and vtilt I can always put it into the light cone.

Jborn yours seems to move the front element (and the whole optical path) around in the v front.

Is that "easy" to work with or would have moving the camera behind the lens be "easier"

[Bjørn Rørslett]

As *I clearly stated initially, movements have be be carried out on either side.

[CardBoardBoxProcessor]

Oh. My apologies. I am just having issue trying to figure out why exactly both are needed is all. sorry. not that I doubt you. I just do not know enough to properly understand.

In my brain it would seem that if the front has enough shift motion it should be able to shift enough to get the image circle over the sensor. and vice versa. if the sensor can shift enough to get into the light circle how would a front shift benefit.

for example I would understand if you need to shfit 50mm to the left but the front can only shift 25mm and the back 25mm so you shit both opposite directions to obtain 50mm. But if the rig can shift 50mm in just the front is back shift actually needed?

for example these seem equal to me -but it would seem I am missing some knowledge of how this works to properly simulate the results?



And then i think, how is tilting the front different than tilting the back? Drawn out it looks the same. But perhaps i am not thinking correctly.

[Jack Dahlgren]

Oh. My apologies. I am just having issue trying to figure out why exactly both are needed is all. sorry. not that I doubt you. I just do not know enough to properly understand.

In my brain it would seem that if the front has enough shift motion it should be able to shift enough to get the image circle over the sensor. and vice versa. if the sensor can shift enough to get into the light circle how would a front shift benefit.

for example I would understand if you need to shfit 50mm to the left but the front can only shift 25mm and the back 25mm so you shit both opposite directions to obtain 50mm. But if the rig can shift 50mm in just the front is back shift actually needed?

And then i think, how is tilting the front different than tilting the back? Drawn out it looks the same. But perhaps i am not thinking correctly.

You are mostly correct.
Rear tilt (assuming top of rear tilting toward photographer) is the same as:
front tilt + front drop + body rotation
Or
front tilt + rear rise + body rotation

As long as the geometry between lens axis and center of focal plane imaging area remain the same, the same image will be formed regardless of how the camera is bent to achieve it.

Trouble arises when front drop is limited (ok, invert the camera) or when mechanical vignetting or bellows elasticity stop you.

[CardBoardBoxProcessor]

Ah I see. Thank you.

I think I'll attempt to move the body then as I dislike parallax errors and love shifted panorama. It is just a 28mm DSLR shift lens so The image circle can't be as huge as a view cameras. So ill jsut give body movment more than is needed and copy wide angle bellows style designs.