View Single Post
  #29  
Old August 2nd 04, 02:36 PM
Nostrobino
external usenet poster
 
Posts: n/a
Default perspective w/ 35mm lenses?


"Jeremy Nixon" wrote in message
...
Nostrobino wrote:

Yes. However, a wide-angle lens includes more objects and therefore has

more
and different relationships, than a long lens.


No; it has more relationships, but the ones it shares are exactly the

same.

It is the relationships NOT SHARED that make the difference.


Wide-angle lenses tend to exaggerate differences in distance, while
telephoto (or more correctly, long-focus lenses whether they are true
telephotos or not) produce the effect of spatial compression. These are
clearly differences in perspective, as it is perceived by the viewer.


Except that the effect is not in any way the result of the focal length of
the lens, but of the magnification.


The magnification IS a direct result of the focal length used.



If that were true, wide-angle photos and long-lens photos would appear

to
have the same perspective. They do not. I know you know this as well as

I
do.


Except that they quite clearly do.


You are saying that you really cannot see any difference in perspective
between a shot taken with a 24mm lens and one taken with a 200mm lens?
Remarkable.



Wide-angle photos taken from the same distance do not have a "telephoto
look," do they?


There is no such thing as a "telephoto look".


You honestly BELIEVE this? Looking at photos taken with 200mm and 300mm
lenses, you would have no clue from their appearance that they'd be taken
with long lenses? Remarkable.


The "look" you are talking
about is a product of magnification,


Take a 24mm shot and magnify it all you like, it will never (when viewed in
its entirety) look like a 300mm shot.


and the relationship between objects
in the picture is exactly the same as it would be from a wide angle lens.
You can easily prove this to yourself by taking two pictures and comparing
them.

If I shoot buildings with an ultra-wide lens with the camera tilted

upward,
the sides of those buildings will converge toward the top in a way that
appears very distorted, very spatially exaggerated. This is clearly a

matter
of perspective, and meets every ordinary definition for perspective. If

I
shoot the same buildings from the same position with a long lens, there

will
be no such effect; on the contrary there will be a flattening and

spatial
compression as verticals are made more parallel and distance differences

are
made to appear less.


Except that this just plain won't happen. You will merely see less of the
buildings,


Of course.


and thus not notice that the parallel lines are doing exactly
the same thing in both pictures.


There is nothing to notice or not notice; many parallel lines in the
wide-angle shot do not even exist in the long-lens shot.


Changing the field of view (from the same position) CHANGES the

perspective,
is what I am saying.


And this is incorrect.

Just remember that perspective is something that involves THE WHOLE

PICTURE.

No, it's not.


Of course it is. This is really the sticking point, as I have indicated
before. When one speaks of any picture as having perspective, it is the
whole picture that one is talking about. If you start zeroing in on smaller
and smaller components of the picture, you not only change the perspective
as you do so but could eventually reach a point where there is no
perspective at all.



Those things aren't what matters as much as perspective. With 35mm for
example, why does anyone use a 105mm or so lens for portraiture? Because

a
longish lens gives a more flattering perspective.


No, that's not why. It's because standing farther away from the subject
gives a more flattering perspective, and the telephoto lens lets you fill

the
frame with the subject from farther away. The long lens absolutely,

clearly,
provably does *not* flatten anything.

You could use a 28mm lens and move in to fill the frame just the same,
couldn't you? But the results would be horrid. Perspective is what makes
the difference.


Yes, and as you said, you've moved in to fill the frame, changing the
perspective. Try it from the same place, and the perspective will be
exactly the same as the telephoto shot; the features of the person's
face will have the same relationship to each other and to their
surroundings.

If you used the 28mm from the original 105mm position would the

perspective
be the same (this is what you're claiming, right)?


Yes, it would.

No, it would not. The 28 would produce not only a smaller image of the
subject, but also more convergence in parallel lines outside of the

subject

It would produce a smaller, but identical image of the subject,


"Smaller but identical" is a contradiction in terms.


and the
parallel lines would be the same except that you'd see more of them,


Then they would not be "the same."


and
thus be fooled into thinking there is more convergence when in fact there
is not.


The convergence is not really there anyway, i.e. parallel lines do not
converge in a three-dimensional world. It is only the APPEARANCE of
convergence that lends any picture its perspective. It is the exaggerated
convergence of parallels in a wide-angle shot that give it the familiar and
easily seen (no matter how strenuously denied) wide-angle perspective.



How often do you have to see a certain look with your own eyes before

you
admit that that look does, in fact, exist?


The wide-angle look is a product of the larger field of view, not of the
perspective. The reason things seem to distort at the edges of a wide-
angle image is because you are projecting a spherical image onto a flat
plane,


No, not a spherical image. What is projected onto the flat plane is the
two-dimensional representation of a three-dimensional world. That's
perspective.


and the larger the area of the sphere you use, the more that will
result in "distortion" from what you expect to see (but in fact it's not
distorted, it's just one possible projection; a fisheye lens produces
another, equally valid, projection, but one that differs from the way
we assemble images in our brain and therefore one that looks weird).


The fisheye lens does project a spherical image (more correctly, an
inside-the-hemisphere image) onto a flat plane. But I don't believe we have
any real disagreement about that, and it's really off the subject of
perspective anyway. Perspective, at least in the classical sense, is only
obtained with (reasonably) rectilinear lenses.

We could probably get into a discussion of non-classical perspective with
fisheye lenses, but it would make my teeth hurt.