On 22 Jan 2003 00:16:28 GMT, contaxman@aol.comnospam (Lewis Lang) wrote:

>>Subject: Re: "learning" wide angle lenses? - L - "Perception Vs. Perspective"

>>From: d_ruether@hotmail.com  (Neuman - Ruether)

>>Date: Tue, Jan 21, 2003 8:35 PM

>>Message-id: <3e2fac31.4555389@newsstand.cit.cornell.edu>

>>On 21 Jan 2003 10:17:03 GMT, contaxman@aol.comnospam (Lewis

>>Lang) wrote:

 

>>[...]

>>>When you are viewing a scene your unaware of how your mind alters the total

>>>picture while it processes all these scans, so yes it does take some "training"

>>>to see photographically and see the difference between what we see/perceive

>>>w/ our eyes and what we will record w/ a camera, but in actuality, a camera

>>>placed at the same exact spot as one of our eyes should, in effect, see the same

>>>arrangement of shapes/compostion/perspective relationships between foreground,

>>>middleground and background see the exact same image relationships as the

>>>eye does.

>>[...most snipped...]

 

>>I guess I disagree with much of this post, though your

>>first line above may indicate why... For those who see

>>"narrowly", your explanation may make some sense - but

>>for those of us who see "widely", it is obvious that

>>the perspective type commonly thought of as "correct"

>>(rectangular perspective - in which all subject straight

>>lines are rendered straight in the image) is not as good

>>a representation of the way we see as the "wrong" one is

>>(spherical [or "fisheye"] perspective). See for more

>>www.David-Ruether-Photography.com/articles.html#perspective and

>>www.David-Ruether-Photography.com/perspective-correction.htm.

>>BTW, it is very easy to show in a couple of ways that

>>our eye-viewing perspective-type is not rectangular,

>>and that some of the effects you noted (verticals staying

>>parallel on tipped views of buildings, for instance)

>>are simply characteristics of spherical perspective...

>>  David Ruether

 

>Our eyes may see more spherically in perspective but our minds are trained to

>see or accept more rectangular perspective (or perspective conventions) as more

>normal.

[...]

 

Yes, this is true - but it is useful to say why extreme

wide-angle images often look "wrong", in addition to reasons related to learned conventions. If one uses only a 50mm (or

so) lens on a 35mm camera (particularly if it has no linear

distortion), this view may seem "normal", though it is FAR

more restricted in angle than what most people can easily

see. If one goes to a 15-20mm lens (which is closer to

the normal angle of view for many), it is obvious that

image characteristics that still follow the rules of rectangular perspective (the learned "correct" perspective...) do not look right, though the perspective type may follow the conventions precisely. (It was amusing to see an old painting once which presented a wide angle of

view of a courtyard with a row of columns with arches on top

on the right side. Most of the image followed the rules

of rectangular perspective, but the arches were progressively rendered more in spherical perspective as

the edge of view was approached. The painter obviously knew

how to draw the arches "correctly" in rectangular

perspective, but just as obviously, he decided to modify

the shapes that would have otherwise looked "wrong",

regardless of the conventions...) I think it is useful for

people to understand how they see, so that they can resolve

issues of "distortion" themselves, if they wish (or they can

just keep shooting with a 50mm, if they don't...;-). BTW, if

you place a full-frame fisheye lens on a camera in place of

a rectangular-perspective super-wide, you may note a few

things: the forground-to-background object relative

proportions change, favoring the fisheye for looking

"natural"; if you place rounded objects in the field that

can be viewed with both lens types, these will look more

normally rendered near the edges with the fisheye; if you

tilt the view upward and look at the renderings of

parallel vertical lines on normal-distance/normal-height

buildings, they are shown more nearly parallel in the

fisheye view; if you place very long straight lines

off-center in the fields of view, the

rectangular-perspective lens will look more natural

(though it isn't - we see these lines as curved, much

as people resist this idea...!;-); if you limit the coverage

of both lens types to that of a 50mm, the images from

both will look quite similar...