Spaghettification

[Maris Rusis]

Marcus, what you saw is a manifestation of LATERAL MAGNIFICATION. This is a shape distortion that is produced off-axis in rectilinear imaging systems. The classic example is where a sphere imaged at the edge of a wide angle view becomes a stretched oval shape!

Not only are shapes stretched but image movement is also accelerated as the edge of the format is approached. Again, this is very conspicuous with wide or ultra-wide images.

Lateral magnification is a purely geometric effect and happens to both lenses and pinholes. There are lenses that avoid lateral magnification by embracing spherical perspective rather than rectilinear rendition. Unfortunately these lenses have other distortion problems. They are known as fish-eyes.

[Marcus Carlsson]

Was the camera's back parallel to the subject or parallel to a divergent universe? If it was a divergent universe, then the effect should be seen with any lens.

Yes, both the model, the front and back was parallell. The camera was parallell with the ground (so it was not tilted) and I dropped the front to make sure that Ellen's feet was shown in the picture.

/ Marcus

[Struan Gray]

If you have he patience, or a roll film back, it is interesting to take a series of portraits with the camera at the same position but at different heights, using vertical shift to frame the shot the 'same' way each time. I once saw an extremely effective 'group' portrait of the same child taken this way, digitally assembled from the individual shots. Combined with the child's changing pose, the flaring outwards at the extremes gave a fantastic sense of gesture to the series of portraits. Few people know the optics behind LF movements, but most can appreciate the changing perspective.

[Marcus Carlsson]

In this photo you may see it a bit. Don't you think that Ellen's hand look a bit to long (I know it's out of focus, that shouldn't matter, right)?



/ Marcus

[Patrik Roseen]

Marcus,
I think you need to consider the effect of perspective. The hand is much closer to the lens than the rest of the person...

[Marcus Carlsson]

Marcus,
I think you need to consider the effect of perspective. The hand is much closer to the lens than the rest of the person...

Ok, I'm still a newbie (aren't we all), but I just hopped this would be a good example, but maybe not. I still get spaghetti all over me when Ellen is standing up though.

But thanks for pointing that out, because I just might do some more still-life and really learn what happends with different objects depending where they are.

/ Marcus

[Struan Gray]

I think of it this way: imagine a portrait of someone with their arms stretched out sideways taken with a 40 mm lens on 6x6 rollfilm. If the portrait is framed so that the person's head is in the center and their fingertips are at the edges it seems obvious that their hands will be elongated by use of such a wide angle lens. If you took the same shot on 35 mm, also with a 40 mm lens, but with enough sideways shift to place the subject's nose on the right edge of the frame you would see the same elongation. The 35 mm shifted shot is equivalent to a crop from the 6x6 negative, and the elongation occurs for the same reason, but now it looks odd because it is not centered in the frame.

All rectilinear lenses do this, but most of those made for smaller formats will vingette before you get to the point where the elongation is noticeable. LF lenses are deliberately made with image circles larger than the intended film format, precisely so that you can play this kind of game.

Mathematically, the elongation is because a 'rectilinear' lens focusses an object Ø degrees off axis in front of the lens at a position equal to f*tan(&#216 away from the center of the focal plane. (f is the focal length). tan(&#216 runs away to infinity at larger angles, pushing objects further and further away from the optic axis and thus stretching them radially.

If you have a wide angle lens and a willing model (or a clothes horse, or dressmaker's dummy) it can be fun to see the distortion move from head to feet as you do the exercise I mentioned above.

PS: this thread at acutance.net has a nice example, which includes near-far effects and the wierd perspective you get if you shift by a reasonable fraction of the focal length:

http://www.acutance.net/phpBB2/viewtopic.php?t=1277

[Struan Gray]

Incidentally, you need 3D shapes for the elongation to be apparent. A sphere looks more and more egg-like as you shift away from the center axis, but a circular disk parallel to the film plane always appears as a circular disk, no matter how much you shift the lens (this is why 2D copy work can be done with Biogons and Topogons). It is the three dimensional nature of the object and the different stretching of different bits of it that make things look wierd.