35mm, wide angle tilt lens, and bowing of foreground effect.

[newmoon2night]

I know this isn't LF, but I'm looking to understand this effect, that I don't get when using a moderate wide angle lens on 4x5.
Maybe it's just more noticeable with the more elongated 35mm format.
When looking at shots taken by people using a 17mm tilt and shift lens, where the foreground is relatively flat (a beach for example) there seems to be an exagerated effect of looking down to the immediate foreground, and a slight bowing from edge to centre at the foreground (the centre appears slightly higher than the edges). However the horizon looks as straight as a rule.
Now I've noticed this effect, I can see it in a shot of limestone pavements I've taken with a 24mm shift lens on 35mm format. In the shots I took, the camera was tripod mounted about five feet from the ground, there was no shift, and one degree of tilt ... and I wasn't on a slight hill!
The effect is only really noticeable when the foreground is flat and uncluttered.
Assuming this isn't an illusion on my part (my wife, an artist, noticed it immediately) is there a work around? I haven't had a chance to experiment, but am wondering whether getting nearer the ground would improve matters, though when using this lens I use tilt tables, so the closer to the ground I get the more degrees of tilt I would use, so maybe the effect will be the same. The alternative is not to use tilt in this sort of flat foreground situation, or change the way I compose a shot.
Any advice on this much appreciated. Thanks. David.

I have just added that as far as I can tell from test data, neither the 17mm or 24mm tilt and shift lenses show any significant distortion without tilt, and as stated previously, the horizons are absolutely straight.

[speedfreak]

I would venture to guess it has to do with the angle of the film plane in relation to the lens and subject plane. Just as when you use a view camera and tilt the rear standard back or forward it will distort objects within the frame, and will be especially apparent in your foreground subjects or any sort of straight lines. Often if I want to enhance the presence of my foreground object with a wide angle shot I'll use a bit of rear tilt with my front tilt to achieve this effect (the front tilt would be used to get the depth of field that I wanted, but not necessary to get this effect).

[newmoon2night]

I would venture to guess it has to do with the angle of the film plane in relation to the lens and subject plane. Just as when you use a view camera and tilt the rear standard back or forward it will distort objects within the frame, and will be especially apparent in your foreground subjects or any sort of straight lines. Often if I want to enhance the presence of my foreground object with a wide angle shot I'll use a bit of rear tilt with my front tilt to achieve this effect (the front tilt would be used to get the depth of field that I wanted, but not necessary to get this effect).

That's a decent point, though I'm not sure why that effect should happen, unless the sensor (film) plane isn't quite vertical. I was using a spirit level, but maybe a very slight backwards tilt in 35mm format wouldn't look very far off on the spirit level, and the smaller the format the more exagerated the effect.
It could also be that this is more visible with what would be considered ultra wide angle lenses in LF use (17mm x 3= 51mm LF equivalent, 24mm = 72mm LF equivalent)
It could also be with other people's shots where I've noticed this, that they are hand holding the camera (perfectly possible and I think often done with a 35mm DSLR and Tilt and Shift lens - people just set the tilt at one degree if standing, or two degrees if kneeling, and away they go! I do it myself sometimes), so backward tilt in sensor plane would be very easy to happen.

Anyone else with any other thoughts on this please?

[Emmanuel BIGLER]

Hello from France

I am not sure that I understand precisely what you are explaining, a link to an image would help to understand the situation, but I can right now tell for sure that retrofocus lenses behave differently from quasi-symmetric view camera lenses in certain specific 'tilted' shooting situations.
I do not know whether what follows is relevant or not to your point, but at least the information is interesting to all users of retrofocus tilting lenses.
Those lenses do not exist for the large format since large format reflex cameras were no longer in vogue when engineers begun to develop top-class wide-angle retrofocus lenses in the sixties. However retrofocus wide-angle tilting lenses for reflex cameras are actually in vogue, in addition to Canon's and Nikon's, Schneider-Kreuznach presented nothing less than 3 lenses at the last 2010 Photokina (2 for the 35 mm format, one for the 645 medium format)

With a quasi symmetric view camera lens, imagine that you take a classical picture of a tall building, and that as usual obeying the good old rules, you keep your film plane vertical. Then you decide to slightly infringe the rules by tilting the lens board upwards, in order to get more of the higher part of the buidling within your frame. Doing so, you infringe Scheimpflug's rule and get defocusing in some parts of you image, but you do not get any keystone distorsion.

Now take a retrofocus lens and place it exactly at the same point of view, so that the perspective of the scene will be exactly the same for the two shots. Keep the camera precisely levelled, with the sensor plane perfectly vertical. Now slightly tilt the lens upwards. The image, for the same reason, will be slightly de-focused as before ; however, you'll get an unusual keystone distorsion that does not exist with quasi symmetric view camera lenses.

The reason for this phenomenon is that de-focused images are created by projections of the exit pupil on the sensor, and retrofocus lenses do not have their pupils located in the principal (or nodal) planes, like quasi-symmetric view camera lenses.
This induces a specific keystone-like distorsion unknown with view camera lenses.

It is not easy to expain the phenomenon in a few words, but if I have time I'll prepare some diagrams with simulations of the effect (based on the OSLO optical design software).