Archived posting to the Leica Users Group, 2008/01/21
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]Michiel, you are completely correct.
In photography there are three common ways to deal with distortion.
One, try to eliminate it by having the lens draw straight lines as
straight lines. This works for most lenses of about 35 or 50mm (in
35mm film terms) on up. This is done by having the lens have the same
magnification over the whole field. When you get to shorter lenses,
using this design philosophy causes three dimensional things to go
funny in the corners, ie, elongate into the corners. As H. Ball Arche
noted, people don't look the way we want them to when shooting with a
28 and placing their heads in a corner. Go wider and it gets worse.
The second way to deal with distortion is to have a rotating lens
camera, or stitch panoramas. This allows straight lines in the
subject that are off the horizon to bow, and also works for very wide
shots with people in them, because the only elongation that can occur
is at the extreme edge top and bottom, and even there there will not
be much elongation because the distance off the horizontal is not
that much, really. A camera that does 140 degree panoramas typically
has a lens equivalent to a 28 or more. Also, as there is obvious
linear distortion we readily accept other distortion.
The last way is with fisheye lenses. This is the only way we can
instantaneously image 180 degrees in all directions (aside from very
esoteric means). If we have a person's head near the edge of a
fisheye image, it will be distorted, but it won't be as elongated as
if it were near the edge of a similarly wide rectilinear lens. If we
crop out a center section of a fisheye image , we have a 'normal'
lens with severe barrel distortion, and things like elongation
wouldn't happen. There will still be distortion, but for three
dimensional objects the distortion will appear less than that of a
(wide) rectilinear lens. For two dimensional objects the distortion
will of course appear quite severe.
What this all means is that if we have wide angle images, there will
be distortion. Period. What type is most acceptable to us depends on
ourselves and the subject.
For retrofocus wideangle lenses, two types of correction are used.
Try to keep the distortion as low as possible over the whole field,
and you get slight 'moustache' distortion. That's the Summicron - and
the distortion levels of the Summicron are _very_ low compared to
those of other fast 28's. Or try for more basic correction, which
will produce classic barrel distorion (Ultron) and let the corners
have more distortion. Usually, this is a less desirable solution,
even though is easier. The Ultron still doesn't have high levels of
distortion, but they are greater than those of the Summicron.
In the end, you'll have distortion of faces in the corners with a 28.
That's why the 'standard' rule is to use longer lenses for people.
At 12:47 AM +0100 1/21/08, Michiel Fokkema wrote:
>If you are shooting architecture you don't want barrel distortion.
>A good lens is is supposed to have as less barrel distortion as possible.
>I bought a elmarit 19 fist version lately and I noticed the same
>effect you mention. Not too much elongation of peoples head. I did
>shoot my birthday party with it. Must have some barrel distortion
>but unless you shoot serious architecture or brick walls you won't
>notice it.
>yes, you have to pay to read Sean's reviews. I did, and he sure
>makes nice and thorough reviews.
>
>Cheers,
>
>Michiel Fokkema
>
>H. Ball Arche wrote:
>> Huh. Well. Jeez.
>> Thanks for that Michiel, but now I'm really confused.
>> I'd like to read that explanation, but I guess I'd
>> have to buy into Sean Reid's site to get to it. What's
>> the point of eliminating barrel distortion if by doing
>> so a grotesque interpretation of human physiognomy is
>> introduced?
>> Practically, how does the elimination of barrel
>> distortion apply - is it a matter of acheiving a clear
>> straight line interpretation of linear, architectural,
>> perspective? I can't see the point if screws up the
>> expression of volume or biomorphic form.
>> Is this a case a case of engineering to bench testing
>> parameters as opposed to real-world shooting? Or a
>> display of the difference between Japanese design
>> esthtetic (the folks who gave us the concept of bokeh)
>> and a Western attention to mathematics?
>> Since the Ultron has been discontinued, I guess I'd
>> better get a hold of a spare from Gandy while he still
>> has some.
>>
>> --- Michiel Fokkema <michiel.fokkema@wanadoo.nl>
>> wrote:
>>
>>> Hi,
>>>
>>> As I understood from Sean Reid, barrel distortion
>>> helps to prevent the elongation of the peoples heads. I can imagine the
>>> cv 28 has more barrel distortion than the Leica. The elmarit asph. is
>>> said
>>> to have the least barrel distortion of all the Leica 28's. This would
>>> mean that it will not suit you.
>>> All hearsay from Sean and Erwin.
>>>
>>> Cheers,
>>>
>>> Michiel Fokkema
>>
>>
>>____________________________________________________________________________________
>> Be a better friend, newshound, and know-it-all with Yahoo! Mobile.
>>Try it now.
>>http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
>> _______________________________________________
>> Leica Users Group.
>> See http://leica-users.org/mailman/listinfo/lug for more information
>>
>
>
>_______________________________________________
>Leica Users Group.
>See http://leica-users.org/mailman/listinfo/lug for more information
--
* Henning J. Wulff
/|\ Wulff Photography & Design
/###\ mailto:henningw@archiphoto.com
|[ ]| http://www.archiphoto.com