That elusive term: "Perspective"

[cowanw]

Uh-oh...

Perspective controls on a stereo view camera...

This could get complicated...

"The drawing you present is 2D, it looks fine, parallax has not changed.

We need a 3D scene though, I'm going to imagine it on top of that 2D scene."


I woke up this morning thinking this is significant. the world is only seen as 3D because we have 2 eyes. If we close one Eye (and then we see 2D) we are then looking as a camera looks.
Regards
Bill

[Mark Barendt]

Side note for clarity.

In a painting or drawing a viewer's perspective and the artist's perspective are normally the same. Their common view is looking "at" or "through" a canvass.

When painting or drawing an image the perspective does not exist except, in the artists imagination, until it is on the canvass.

An artist painting or drawing on a canvas can choose the line of view (the perspective they want to portray) without any regard for his or her actual line of view. A skilled artist could paint any perspective (line of view) they choose without error from anywhere they please.

In photography the image actually has to exist in the real world for a given perspective to be captured in a camera.

With a view camera we can try to mimic a different point of view/line of view by skewing the geometry of the image around the center of the film, but we are limited to real points of view, perspectives, that we can actually get our cameras to.

[Mark Barendt]

"The drawing you present is 2D, it looks fine, parallax has not changed.

We need a 3D scene though, I'm going to imagine it on top of that 2D scene."


I woke up this morning thinking this is significant. the world is only seen as 3D because we have 2 eyes. If we close one Eye (and then we see 2D) we are then looking as a camera looks.
Regards
Bill

Good morning Bill,

You are right.

[Emmanuel BIGLER]

Quoting Mark Barendt:
An artist painting or drawing on a canvas can choose the line of view ..

May be, at this point of the discussion, should we all re-built Brunelleschi's peephole-mirror device before setting up a view camera !

http://www.visiblespace.com/phdtext/...ng%20space.htm
http://www.webexhibits.org/arrowinth...elleschi1.html

[Rakesh Malik]

Rakesh,

The scene I'm speaking of is the world not in the camera.

The illustration cowanw provided is 2D, the scenes we normally photograph are 3D.

The last time I checked, the image we're recording on film is 2-d.

[Rakesh Malik]

(Rakesh, if you're still using the Ebony sv45u, doesn't it have back rise? If so, you can turn your camera 90 degrees on a sturdy tripod + head, and its back rise will "convert" into back shift. Just a quick reminder I've offered before, since one's camera is often better than one thinks! )

Good point -- I do forget that I could do that, I'm so used to just rotating the back.

Of course, rise and fall on the rear have the same effect as shift as far as parallax goes, which is to say, none.

[Mark Barendt]

The last time I checked, the image we're recording on film is 2-d.

True Rakesh,

The world/subjects we normally photograph are in 3D.

cowanw was using a 2D subject.

[Mark Barendt]

Well, I do think of the ground glass as a 2D picture and thus think as I do, but you raise a good point of whether the camera will see around real 3d objects. May be some one can get out and do a set of real world examples. If I posted a set from my 35mm sony digital would I be sent to LFF hell?
Regards
Bill

I found an example or two here on the Large Format Photography Forum. (Link below)

Both shots in the first example show the effect of using rear tilt to square-up parallax in the image.

Both shots in the first post are successful, they don't look weird or distorted, they look like we expect the scene to look like from an eye level perspective.

It is very apparent in both shots that the camera was at about eye level.

The geometric give away to the original perspective is what we can see under. i.e. in the motel shot we can see, albeit underexposed, under the parking area roof.

The lens in both cases has been raised and the line of view slopes upwards. Find the exact center of the frame and you can see which way the camera was pointed.

http://www.largeformatphotography.in...6&postcount=68

Another example, http://www.largeformatphotography.in...&postcount=255

In this shot parallax "error" has been exaggerated after the fact, the top edge of the negative is shorter that the bottom edge. Regardless of the skew that has been added, there is no way to mistake that the original perspective was pointing down into the scene from maybe the 7th or 8th or 9th story.

[Rakesh Malik]

True Rakesh,

The world/subjects we normally photograph are in 3D.

I've noticed this also.

It does not however change the fact that the film plane is, in fact, a plane. Moving the film on that plane will have no effect on parallax.

Have you tried it yet? Can you point out any reason that shifting the rear standard should be any different from moving a crop rectangle over the image in Photoshop?

I've done it, and seen the results myself... though since I tend to forget that I can flop my camera 90 degrees and turn my rise into shift, I've only viewed on my own camera with rise.

Note that I've been specifically referring to SHIFTs, that is to say, PLANAR movements on the rear standard, because they are not the same as rotations for obvious reasons.

[Mark Barendt]

... It does not however change the fact that the film plane is, in fact, a plane. Moving the film on that plane will have no effect on parallax.

True. It does though change the line of view, and therefore, the perspective.

Have you tried it yet?

I have tried it but can't scan it. See post 109 for examples I've seen. These examples show rise and drop.

Can you point out any reason that shifting the rear standard should be any different from moving a crop rectangle over the image in Photoshop?

The answer is absolutely yes.

First, the camera's perspective on the world is defined by the center of the lens and the center of the film. Shift changes the line of view for the camera, the camera ends up pointing a new direction and you have a new composition.

Have you tried this? Just set up your camera aimed at a door dead center in the ground glass, now add shift, rise, or fall. The door will move off center. The film and lens will define a new center around which everything balances or at least pivots.

Any view you can capture on the ground glass or film has all the info we need to understand the perspective it was shot from. In a sense the perspective pivots around the center point of the film. Parallax also pivots on this center point.

If we crop off center, we can lose that balance (the pivot point won't seem real) and the parallax you started with may not make sense any more.

Second, since shifting, in contrast to any crop (PS or enlarger), changes only the line of view of the camera, there is no crop. No crop means less grain and better detail and better enlargements.

Third, the basic purpose of a shift, IMHO, is to see around something.

If there is a tree in the way you can move the camera and tri-pod sideways a few feet and adjust so that your view is clear and your chosen parallax remains.

This is not without cost though, the angles that the rays of light come through the lens at and reach the film at get progressively steeper as we move to the edge of the cone the lens is providing.

Real world lenses are normally best when the camera is zeroed and square straying off center makes it tougher to get a perfect film. Straying way off center can introduce many problems.

Straying off center does beat a PS crop in my mind any day though.