distortion free wide angle lens
bacra
frederick
> http://fullygeek.com/2006/12/distortion-free-wide-angle-lens/
>
>
Unless they have invented new spherical paper to print on or a new
spherical monitor, then that's a load of crap.
Don Stauffer in Minnesota
150 degrees maybe. I have seen lenses with wider fields than that
which used software to correct the image from a really wierd format,
but these were used in dedicated cameras with the software built in so
they could ONLY use that lens.
150 degrees may be possible. It would be a challenge. My group at one
time was in the very wide angle lens business and it sure is a
difficult job. And we even were allowed the use of software correction
to eliminate residual distortion.
if
> http://fullygeek.com/2006/12/distortion-free-wide-angle-lens/
This would not be a distortion free image: what they are actually talking
about is introducing more rectilinear distortion, ie. unnaturally
stretching things at the edges of the image. Curved lines in ultra-wide
views are not really a distortion (except subjectively, due to how the
image is viewed). Rather they are a the product of a natural perspective
which reconciles the fact that different ends of the same line tend to go
off to different vanishing points (both above and below the viewer, or on
either side). Rectilinear perspective removes one set of vanishing points
by grossly distorting the parts of the image that would have gone off to
those vanishing points. The result is that objects at the corners of the
image are enlarged, and circles, faces, etc become ovals or ellipses: in no
way can such an image be called "distortion free"! All wideangle lenses
have to make a compromise between distortions of shape, size, and straight
lines, there is no single right answer.
--
_______________________________________________________
HELP! I'M BEING HELD PRISONER IN A .SIGNATURE FACTORY
_______________________________________________________
bugbear
>
> 150 degrees may be possible. It would be a challenge. My group at one
> time was in the very wide angle lens business and it sure is a
> difficult job. And we even were allowed the use of software correction
> to eliminate residual distortion.
>
It strikes me in the modern are there may be an interesting
design compromise where the lens only has corrections
that cannot be done in software.
Since most corrections in lens design (AFAIK) are a compromise
with each other, pushing off some of the correction
to a post process should allow superior correction
of the abberations that remain.
e.g. most of the geometric distortions can be handled
nicely in software.
BugBear
Kevin McMurtrie
"bacra" <ba...@noemailhere.com> wrote:
> http://fullygeek.com/2006/12/distortion-free-wide-angle-lens/
Of course there's no such thing. Mapping 3D to 2D always creates
distortion of some kind. Rectilinear lenses are only distortion free
when the subject is in a 2D plane. Fisheye lenses are only distortion
free when the distance of the subject is constant.
Making photographic use of a 150 degree field of view would be very
challenging. If this is for a security camera, then even major lens
defects would be trivial compared to the extreme natural distortion.
Don Stauffer in Minnesota
if wrote:
> "bacra" <ba...@noemailhere.com> wrote:
>
> > http://fullygeek.com/2006/12/distortion-free-wide-angle-lens/
>
> This would not be a distortion free image: what they are actually talking
> about is introducing more rectilinear distortion, ie. unnaturally
> stretching things at the edges of the image. Curved lines in ultra-wide
> views are not really a distortion (except subjectively, due to how the
> image is viewed). Rather they are a the product of a natural perspective
> which reconciles the fact that different ends of the same line tend to go
> off to different vanishing points (both above and below the viewer, or on
> either side). Rectilinear perspective removes one set of vanishing points
> by grossly distorting the parts of the image that would have gone off to
> those vanishing points. The result is that objects at the corners of the
> image are enlarged, and circles, faces, etc become ovals or ellipses: in no
> way can such an image be called "distortion free"! All wideangle lenses
> have to make a compromise between distortions of shape, size, and straight
> lines, there is no single right answer.
>
correct meaning of distortion. Aberrations, flare, curvature of field,
etc., are NOT distortion. It is only in the past five years or so that
I have heard of such image defects being called distortion. I am for
going back to the original meaning, and referring to others as
aberrations or whatever the correct optical term is.
Not Disclosed
This is a rehash of the "bird's eye" lens of the 1970s
Don Stauffer in Minnesota
Kevin McMurtrie wrote:
> In article <Ukwch.1795$HU....@news-server.bigpond.net.au>,
> "bacra" <ba...@noemailhere.com> wrote:
>
> > http://fullygeek.com/2006/12/distortion-free-wide-angle-lens/
>
> Of course there's no such thing. Mapping 3D to 2D always creates
> distortion of some kind. Rectilinear lenses are only distortion free
> when the subject is in a 2D plane. Fisheye lenses are only distortion
> free when the distance of the subject is constant.
>
plane in object space perpendicular to the lens axis. If the ratio of
the distance to any point from the OA at which a ray from a given point
in object space intercepts that plane to the distance from the OA to
image of that point in the image plane was the same ratio for all
conjugate points, then the lens is distortion free.
That is essentially saying that the ratio of the field angle of any
point in object space to the distance to the image of that object in
the image plane is constant.
> Making photographic use of a 150 degree field of view would be very
> challenging. If this is for a security camera, then even major lens
> defects would be trivial compared to the extreme natural distortion.
What do you mean by "natural distortion"?
Kevin McMurtrie
The perspective distortion. I have the Canon 10-22mm lens and it's very
difficult to use. Aim it at a room full of people at 10mm and it will
look like a funhouse mirror. The lens is best used for cases where you
want to amplify perspective for a controlled effect.
Don Stauffer in Minnesota
Kevin McMurtrie wrote:
> In article <1165503818.0...@f1g2000cwa.googlegroups.com>,
> "Don Stauffer in Minnesota" <stau...@usfamily.net> wrote:
>
> > > Making photographic use of a 150 degree field of view would be very
> > > challenging. If this is for a security camera, then even major lens
> > > defects would be trivial compared to the extreme natural distortion.
> >
> > What do you mean by "natural distortion"?
>
> The perspective distortion. I have the Canon 10-22mm lens and it's very
> difficult to use. Aim it at a room full of people at 10mm and it will
> look like a funhouse mirror. The lens is best used for cases where you
> want to amplify perspective for a controlled effect.
The problem is that the human eye does not (in spite of what so many
books say) work exactly like camera and vis versa. The "distortion"
or perspective you are talking about is not due to the lens, per se,
but strictly the viewpoint, or distance from the scene at which the
picture is taken. If you somehow managed to fasten down a human eye at
the location from which you take a close-up picture, and take it with a
low geometric distortion lens, you would also see the same problem.
However, when we are close to something (or somethings) we see it or
them by moving our eye and scanning the picture.
Then the brain sort of "scales" the images because it KNOWS that
people's heads are all about the same size, and makes the scene look
like what your brain expects to see. Very hard for even a digital
camera with a computer on board to do this. The problem is NOT in the
lenses, it is in the way humans see
David Ruether
"Don Stauffer in Minnesota" <stau...@usfamily.net> wrote in message news:1166026419.6...@73g2000cwn.googlegroups.com...
> Kevin McMurtrie wrote:
>> The perspective distortion. I have the Canon 10-22mm lens and it's very
>> difficult to use. Aim it at a room full of people at 10mm and it will
>> look like a funhouse mirror. The lens is best used for cases where you
>> want to amplify perspective for a controlled effect.
> The problem is that the human eye does not (in spite of what so many
> books say) work exactly like camera and vis versa. The "distortion"
> or perspective you are talking about is not due to the lens, per se,
> but strictly the viewpoint, or distance from the scene at which the
> picture is taken. If you somehow managed to fasten down a human eye at
> the location from which you take a close-up picture, and take it with a
> low geometric distortion lens, you would also see the same problem.
> However, when we are close to something (or somethings) we see it or
> them by moving our eye and scanning the picture.
>
> Then the brain sort of "scales" the images because it KNOWS that
> people's heads are all about the same size, and makes the scene look
> like what your brain expects to see. Very hard for even a digital
> camera with a computer on board to do this. The problem is NOT in the
> lenses, it is in the way humans see
I think it is different from this. We actually see in spherical (or "fisheye")
perspective due to the curved structure of the eye. Since most people
attend only to the center of their vision (and move their eyes around to
see other things), they miss this and believe that vision uses rectangular
perspective (which it does approximate over a narrow angle). With
extreme angles of coverage the rectangular camera perspective looks
wrong with 3-D rounded objects, as does the (oddly...) unfamiliar
perspective of a lens with great barrel "distortion" (the fisheye, though
that is actually more correct in its representation of our vision with very
wide angles of view). With a little training it is possible to see the
curvature of straight lines away from the center of vision (characteristic
of spherical perspective) fairly easily. For more on this, see
www.ferrario.com/ruether/articles.html#perspective
--
David Ruether
DRue...@twcny.rr.com
rp...@cornell.edu
http://www.ferrario.com/ruether
Kevin McMurtrie
The brain almost refuses to see distortion from your eyes. You have to
really concentrate to see it. The trick is that you have two eyes and
they never hold still while open. You brain can extrapolate a mostly
distortion free image from that. The remaining distortion and defects
are ignored - you can not see your own blind spot even though you can
easily find it by watching your wiggling finger tip vanish at a point.
Unfortunately, a steady image from a camera won't make the visual cues
needed for your brain to ignore perspective distortion. Telephoto
images look flat and wide images look stretched.
David Ruether
"Kevin McMurtrie" <mcmu...@dslextreme.com> wrote in message news:mcmurtri-330E71...@sn-radius.vsrv-sjc.supernews.net...
[...]
> The brain almost refuses to see distortion from your eyes. You have to
> really concentrate to see it. The trick is that you have two eyes and
> they never hold still while open. You brain can extrapolate a mostly
> distortion free image from that. The remaining distortion and defects
> are ignored - you can not see your own blind spot even though you can
> easily find it by watching your wiggling finger tip vanish at a point.
>
> Unfortunately, a steady image from a camera won't make the visual cues
> needed for your brain to ignore perspective distortion. Telephoto
> images look flat and wide images look stretched.
And (not disagreeing with you), we actually see in spherical (or "fisheye")
perspective due to the curved structure of the eye. Since most people
attend only to the center of their vision (and move their eyes around to
see other things), they miss this and believe that vision uses rectangular
perspective (which it does approximate over a narrow angle). With
extreme angles of coverage the rectangular camera perspective looks
wrong with 3-D rounded objects, as does the (oddly...) unfamiliar
perspective of a lens with great barrel "distortion" (the fisheye - though
David Ruether
"David Ruether" <drue...@no-junk.twcny.rr.com> wrote in message news:qOAgh.4834$nq5....@twister.nyroc.rr.com...
> --DR
To add some things -- it seems to me that for shooting super wide angle
photos including people and other rounded objects (and some landscapes,
since more natural fore-to-distant size relationships are had with fisheyes)
that an as yet not offered lens type would be useful - one with some very
pronounced barrel distortion, but well short the amount offered by the fisheye.
Rectangular-perspective super-wide lenses are more suited to architecture,
due to the conventions of architectural representations in drawings, paintings,
and past photos - and due to people's conventional thought about how they
see it, wrong as that may be. Ever wonder why the sides of tall buildings
look parallel and straight up if we look up a moderate amount, but look
like they are converging when looking up sharply from the same viewing
location? Both can't be right unless the straight lines curve - and, sure enough,
both can be nearly true in a single fisheye photo of a tall building, but not in a
single rectangular photo.