Related Links:
How To Measure
the Nodal Point (WJ Markerink)
From http://www.vanwalree.com/optics/misconceptions.html#6
6. With panoramic stitching, where the camera and lens are rotated as a unit, the rotation axis should pass
through the lens nodal point.
Many sources that write this don't specify whether it is the front or rear nodal point, but either way it is incorrect.
To preserve the perspective during rotation, which prevents a displacement of nearby objects relative to the
background, the axis of rotation should pass through the entrance pupil, which is the center of perspective of a lens.
No doubt instructions to find the position of the rotation axis will lead to the correct point, it's just that the designation
nodal point is wrong. The good news is that, unlike a nodal point, the entrance pupil can be found by visual inspection.
When you look into a lens from the front, the entrance pupil is the image of the diaphragm opening.
It won't be a simple task: Refer to this excerpt from a FAQ written by David
Jacobson on the subject:
Subject: Photographic Lenses Tutorial
Organization: Hewlett-Packard Laboratories
by David M. Jacobsonjacobson@hpl.hp.comRevised October 26, 1996
[...]
"Nodal points" are the two points such that a light ray entering the
front of the lens and headed straight toward the front nodal point
will emerge going straight away from the rear nodal point at exactly
the same angle to the lens's axis as the entering ray had. The nodal
points are identical to the principal points when the front and rear
media are the same, e.g. air, so for most practical purposes the terms
can be used interchangeably.
In simple double convex lenses the two principal points are somewhere
inside the lens (actually 1/n-th the way from the surface to the
center, where n is the index of refraction), but in a complex lens
they can be almost anywhere, including outside the lens, or with the
rear principal point in front of the front principal point. In a lens
with elements that are fixed relative to each other, the principal
points are fixed relative to the glass. In zoom or internal focusing
lenses the principal points generally move relative to the glass and each
other when zooming or focusing.
When a camera lens is focused at infinity, the rear principal point is
exactly one focal length in front of the film. To find the front
principal point, take the lens off the camera and let light from a
distant object pass through it "backwards". Find the point where the
image is formed, and measure toward the lens one focal length. With
some lenses, particularly ultra wides, you can't do this, since the
image is not formed in front of the front element. (This all assumes
that you know the focal length. I suppose you can trust the
manufacturer's numbers enough for educational purposes.)
[...]
From: davem@zeppo.cs.ubc.ca (Dave Martindale)
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Determining a Lens' Nodal Point
Date: 17 Jan 1999
dbh@ix.netcom.com (dbh) writes:
>To do it properly, however, I >think the camera needs to rotate around the nodal point of the lens >(some tripod heads make this easy). However, nodal point information >(front or rear) is not widely disseminated. Does anyone know how to >get the information? Will the manufacturers respond to inquiries? >Are nodal points measurable by laymen? Thanks.
Nodal points can be measured by a device called a "nodal slide".
The lens is mounted on an apparatus that holds the lens and allows
it to rotate around an axis that is perpendicular to the optical axis.
If the optical bench was horizontal as it usually is, then the
optical axis is horizontal and the axis of rotation would likely
be vertical. The mechanical mounting allows moving the lens relative
to the pivot point, so the pivot point can be anywhere from well
behind the lens to well in front of it.
Then you aim the lens at an object at infinity - either a collimator
target, or a real object a long ways away. Focus the image onto
a groundglass or CCD sensor or something. Then rotate the lens
around the pivot point. If the *rear* nodal point of the lens is
exactly aligned with the pivot point of the mechanism, then rotating
the lens causes *no* movement of the image. If the image *does*
move left or right as you rotate the lens, the nodal point is
either in front of or behind the pivot point, and you can tell
which direction the offset is in by whether the image moves with
or against the direction of lens rotation.
So, mount the lens, focus it, and rotate it. If the image moves one
way or the other, you need to move the lens in a known direction
relative to the pivot point. Adjust the apparatus to change the
pivot point location, and measure again. This process converges,
and when you've found the pivot point which gives no image motion,
you've found the location of the rear nodal point.
To find the *front* nodal point, just turn the lens around.
Dave
From: lubjek@jump.net (Eric Lubjek)
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Determining a Lens' Nodal Point
Date: 18 Jan 1999
>With the availability of sophisticated software to perform photo >stitching (e.g., QuickStitch, Panavue, etc.), I am considering using >my film camera to create panoramas. To do it properly, however, I >think the camera needs to rotate around the nodal point of the lens >(some tripod heads make this easy). However, nodal point information >(front or rear) is not widely disseminated. Does anyone know how to >get the information? Will the manufacturers respond to inquiries? >Are nodal points measurable by laymen? Thanks.
Finding the nodal point is relatively easy with a "VR tripod head", once you
understand the method. The method can be applied to any lens, and no
special information from the lens manufacturer is required.
Figure 1 below shows what parallax looks like if the nodal point is
slightly in front of the center of rotation. Object A is much closer
to you than object B. Note that A & B move in relation to each other.
+---------------------+ +---------------------+ +---------------------+
| | | | | |
| +------+ | | +------+ | | +------+ |
| | B | | | | B | | | | B | |
| | +------+ | | | +------+ | | +------+ | |
| | | A | | | | | A | | | | A | | |
| +-----| | | | +-| | | | | |-+ |
| | | | | | | | | | | |
| +------+ | | +------+ | | +------+ |
| lens rotated left | | lens rotation = 0 | | lens rotated right |
+---------------------+ +---------------------+ +---------------------+
Figure 1
Figure 2 below shows what you want to see when the nodal point is
directly above the center of rotation. Note that the two objects
still move, but not in relation to each other.
+---------------------+ +---------------------+ +---------------------+
| | | | | |
| +------+ | | +------+ | | +------+ |
| | B | | | | B | | | | B | |
| | +------+ | | | +------+ | | | +------+ |
| | | A | | | | | A | | | | | A | |
| +-| | | | +-| | | | +-| | |
| | | | | | | | | | | |
| +------+ | | +------+ | | +------+ |
| lens rotated left | | lens rotation = 0 | | lens rotated right |
+---------------------+ +---------------------+ +---------------------+
Figure 2
The way to find the nodal point with a VR tripod head is easy:
- Just mount the camera on the VR head and make a good guess
about aligning the lens in the "left-to-right" direction
on the VR head. You want the geometrical center of the lens
directly above the tripod's center of rotation.
- Slide the camera all the way "forward" on the VR head.
- Set the tripod up in a place where there are both near and far
objects in front of the camera.
- Keep sliding the camera "back" on the VR head until what you see
looks like Figure 2 above.
There is more info about how to create panoramas at www.placebase.com.
There is a "How To" page, and a "Links" page. You can find links to
the major VR head vendors, etc. Some of them have very good info
about how to find the nodal point on their web sites.
Once you have created a panorama, you can put it on PlaceBase for the
world to see.
Good luck!
--
PlaceBase | Looking for a place to work, stay, or
LIVE?
the Visual Database of Places | Find it at PlaceBase! No plug-ins
required.
http://www.placebase.com | Panoramic (360-degree) views of the Real
World.
From: "melcrose" melcrose@ptcnet.net
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Determining a Lens' Nodal Point
Date: Mon, 18 Jan 1999
FYI, quickstitch (EXCELLENT ! ! ! PROGRAM ! ! ! ! !)
doesn't worry about any of this..
stand
shoot.
dump to computer..
stitch.
PERFECT everytime.. the program is just way too great!
From: Joe McCary - Photo Response mccary@erols.com
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Determining a Lens' Nodal Point
Date: Wed, 20 Jan 1999
dbh wrote:
> With the availability of sophisticated software to perform photo > stitching (e.g., QuickStitch, Panavue, etc.), I am considering using > my film camera to create panoramas. To do it properly, however, I > think the camera needs to rotate around the nodal point of the lens > (some tripod heads make this easy). However, nodal point information > (front or rear) is not widely disseminated. Does anyone know how to > get the information? Will the manufacturers respond to inquiries? > Are nodal points measurable by laymen? Thanks. > >Dale
To manually determine the Nodal point of a lens (on an SLR) is fairly
easy. First place the lens directly over the pan point. Next select 2
objects witgh vertical lines, one directly behind the other a vertical
half wall works well with a building behind (one close to camera and one
distant). Rotate the camera (pan) so these lines move from one side of
the image to the other and watch the lines. If they "move" apart, move
the camera's position back or forward and repeat the above. If it still
moves reposition again. repeate this until the 2 lines remain inline
with each other while you have panned the camera from one side to
another. Mark this position and you have found the nodal point. Each
lens will be different.
An old fashioned Circuit camera had the film in a curved cylinder around
the rotating lens. This is the end result you are seeking with the
nodeal point and a paning camera and then stitching the images
together. Using one of the new tripod heads from KiWi makes the pan
process easier. They also have a new leveling attachment for the heads
as well. this is 3 small legs between 2 platforms. You screw the "legs"
up or down and watch the levels and can easily acheive a level platform
for shooting your pans.
A good resource for panorama infoprmation is a website called
placebase. www.placebase.com they have several links and suggestions.
They are also about to release software for showing stitched panoramas
on the web without having to download any plugins first. I have an
example on my website of the placebase java applet if you want to see
it.
http://www.erols.com/mccary Follow the VIRTUAL TOUR link to see some
stitched panoramas done for a job.
Joe McCary
From: nospam@imagina.com (Jan Steinman -- jan AT bytesmiths DOT com)
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Determining a Lens' Nodal Point
Date: Wed, 20 Jan 1999
dbh@ix.netcom.com (dbh) wrote:
> With the availability of sophisticated software to perform photo > stitching (e.g., QuickStitch, Panavue, etc.), I am considering using > my film camera to create panoramas. To do it properly, however, I > think the camera needs to rotate around the nodal point of the lens...
Don't get so carried away with the "how" that you forget about the "what!"
There are some very good explanations of "nodal point" in this thread, but
in reality, you can do some wonderful stuff without worrying too much
about nodal points.
In particular, the longer the lens, the less important nodal alignment
is. I've done some telephoto partial-pans with a nodal error of 100mm or
more,
with NO apparent affects! If you're starting out with a normal or
medium-wide lens (say, 28mm and longer), the affects are not very
noticable. As you get wider, the effect is more severe.
I've done some nice stitched 360 pans with a 16mm full-frame fisheye on a
Bogen 3275 gear-head tripod, with little attention to nodal alignment. As
you pan, the edges of the frame seem to be moving faster than the center.
It is a bit disorienting, but not a fatal flaw for amateur work.
My combination (Olympus OM-4T, Zuiko 16mm f3.5 fisheye, Bogen 3275), has a
left offset of about 25mm, and a rear offset of about 16mm -- in other
words, the rotation axis is at the film plane. I believe most of the
distortion I see is due to the 25mm lateral offset, not the front/back
error.
Bottom line: "eyeball" your camera so the pivot point is as close to
below the lens laterally, and as close to the focal length in front of
the film
plane as possible with the equipment you have, and have fun!
--
: Jan Steinman -- Jan AT Bytesmiths DOT com
Date: Sat, 27 Mar 1999
From: WDAVID WDAVID@prodigy.net
To: panorama-l@sci.monash.edu.au
Subject: Re: nodal points
> Any advice I can get > on setting it and on a mechanism for helping me mount the camera > accordingly, would be much appreciated.
The nodal point is where incoming parallel light rays cross and invert
as a result of lens refraction. Put a vertical stick in 2 feet in front
of your camera and, through the lens, note something behind the stick,
"far" away. Rotate the camera and observe through the lens. If the
"something" moves out from behind the stick as you rotate, you are not
rotating over the nodal point. Shift your camera slightly forward or
back and repeat on its mount. If the "something" now moves faster, you
moved your camera in the wrong direction. Repeat until you can rotate
your camera and things behind the stick remain behind it.
Simon will not tell you to remove the stick before exposing the film.
(Its one of his secrets).
--
Best regards,
W. David Schwaderer
Date: Sun, 28 Mar 1999
From: Craig Woods cwoods@iafrica.com
To: panorama-l@sci.monash.edu.au
Subject: Re: nodal points
Hi Brian,
I use a new panoramic head from Manfrotto in Italy for my pans. Has two
slides to position centre through nodal point. Old lenses had a bar or line
across nodal point but as safe rule of assumption is that the nodal point is
very close to the diaghpram...damn..iris!
Craig Woods
Date: Thu, 29 Apr 1999
From: Dan Irvin DIrvin@centuryoptics.com
Subject: RE: measuring the focal length of a lens
I use a nodal slide mounted on an optical bench, with a collimator on one
end. I mount the lens and rock the lens back and forth, once the lens is
over the rear nodal point the image no longer moves when the lens is
rocked.
Dan
----------
From: Jook Leung[SMTP:jook@hili.com]
Sent: Thursday, April 29, 1999
To: panorama-l@sci.monash.edu.au
Subject: measuring the focal length of a lens
That is good to know that Nikon markings are reliable.
How do you measure the focal length of a lens?
Is there any way for the layperson to do it without an optical bench?
Thanks,
Jook
>46.5mm is the correct flange focal distance for Nikon Lenses. Make sure >that in your mounting that the lens mount is parallel to the focal plane. >We use tolerances of +.0000" -.0005" for our systems, and require >parallelism to be within .0002". > >I have measured focal length on a great many Nikon Lenses and find that >the focal length marked is usually pretty accurate. > >Dan Irvin >Century Optics. >dirvin@centuryoptics.com > >----------
---------------------------------------------------------------
Jook Leung jook@360vr.com http://360vr.com 201.894.5881
360VR Photography - Immersive Imaging - Virtual Tours
From Nikon Digest:
Date: Fri, 18 Jun 1999
From: "Eric Grummisch" grumisch@direct.ca
Subject: [NIKON] Optical Centre
>----------------- >Date: Wed, 16 Jun 1999 >From: Gopi Sundaram gopalan@cs.sc.edu >Subject: [NIKON] OT: Optical Center > >Hi all, this is off-topic: > >For you techie-gurus out there, how does one determine the optical >center of a lens ? > >My application for it is that, for a Virtual Reality system, I need to >create a panorama by taking multiple pictures while rotating the >camera about its optical center. > >Will the optical center vary with focal length for a zoom lens ? > >Feel free to reply by email if you don't think it will be useful to >the group. > >Thanks a lot, >Gopi.
Hi Gopi,
I suggest that you try the following:
-Set the camera on a tripod with the tripod column directly
underneath the approximate centre of the lens. This is most easily done if
you have a Manfrotto or Kaidan pano head, although you could use a macro
focusing rail. Line up two vertical lines within the viewfinder; one near
and one far. If inside you might use a table leg ~1 metre away as the near
point and a doorway ~ 6 meters away as the far point. It's easiest if you
put the verticals near the edge of the viewfinder. If you've placed your
references near the left side of the viewfinder, rotate the camera to the
left while looking through the viewfinder to see if the points "move"
relative to one another. If they don't, there is no parallax and you are
pivoting around the optical centre of the lens. Note the backset reference
from the head or rail to allow for easy repeatability.
Eric in Vancouver
From Nikon Digest:
Date: Mon, 21 Jun 1999
From: Steffen Kluge kluge@fujitsu.com.au
Subject: [NIKON] Re: Optical Centre
Gopi wrote:
>For you techie-gurus out there, how does one determine the optical >center of a lens ?
Not that I'm a guru, but:
The position of the optical centre (principal point) obviously
depends on the focusing distance. If the lens is focused at
infinity the principal point is exactly one focal length from
the film plane. If the magnification is 1 the principal point is
two focal lengths from the film plane. The general formula for
distance between principal point and film plane is f(1+m), where
f is the focal length and m the magnification. The formula for
the subject distance from the principal point if f(1+1/m), and
for the film plane to subject distance is f(1+m+1/m).
Hope this helps
Steffen.
- --
Steffen Kluge kluge@fujitsu.com.au
Fujitsu Australia Ltd
Date: Mon, 24 Jan 2000
From: Ron Klein panorama@gci.net
To: panorama-l@sci.monash.edu.au
Subject: Re: Finding correct Cirkut gears by shooting.
Bob Lang's computer program, and then test the results with film.
There are other ways of finding the focal length besides using a nodal
slide, it's just that the slide seems to be the fastest way.
Ron Klein
[Ed. note: see also panoramic pages
discussion]
From Rollei Mailing List:
Date: Sun, 18 Jul 1999
From: "John A. Lind" jlind@netusa1.net
Subject: Re: [Rollei] new panoramic head info
At 18:41 7/18/99 , Andre Calciu wrote:
[snip]
>the image showed the rotating axis going straight through the middle >of the front element of the lens.
I'm not surprised after reading the in-depth discussion about the "front"
and "rear" principal point (or nodal point) locations given on Phil
Greenspun's site under the lens FAQ and tutorial (written by David
Jacobson).
In the FAQ, see Q20 which talks about the pivot point for normal lenses:
http://www.photo.net/photo/optics/lensFAQ.html
The tutorial, which is deeper technically, talks about the principal, or
nodal points of a lens. The discussion is found under the section
entitled
"Object distance, image distance, and magnification":
http://www.photo.net/photo/optics/lensTutorial.html
After reading it, my understanding is these points are not necessarily
where one might expect to find them intuitively, or even inside the lens.
Where they are located is very dependent on the lens design. The "front"
point can be in front of the lens, inside it somewhere, or even behind the
"rear" point. I don't claim to completely understand the entire tutorial,
but every time I read it a little more gets clearer. For the
diagram/picture in the tripod and head brochures, the pivot point may have
been correct for the Nikon and lens depicted. For a different lens, it
could be somewhere else entirely, which is why it is adjustable if I
understood your description correctly.
Maybe the brochure should have the disclaimer: "Your nodal points may
vary." (grin)
--- John
From: Bjorn Rorslett nikon@foto.no
Newsgroups: rec.photo.technique.nature
Subject: Re: That elusive Nodal point
Date: Sun, 20 Feb 2000
Oleg March wrote:
> Could someone help me out please on figuring out a Nodal point on an 8 > mm circular fisheye and 19 mm lenses? (35 mm camera). > Any help will be greatly appreciated. > Thanks, > Oleg March
there will be a rear nodal point 8 and 19 mm, respectively, in front of
the film plane.
Regards
Bjorn Rorslett
Visit http://www.foto.no/nikon/ for UV and IR Colour Photography and
other Adventures in Nature Photography
From: "Bob Talbot" BobTalbot@st-abbs.fsnet.co.uk
Newsgroups: rec.photo.technique.nature
Subject: Re: That elusive Nodal point
Date: Sun, 20 Feb 2000
> there will be a rear nodal point 8 and 19 mm, respectively, in front of > the film plane.
... when focused on infinity :o)
From: farrar@datasync.com (Paul Farrar)
Newsgroups: rec.photo.technique.nature
Subject: Re: That elusive Nodal point
Date: 23 Feb 2000
Oleg March olegm@concentric.net wrote:
>Could someone help me out please on figuring out a Nodal point on an 8 >mm circular fisheye and 19 mm lenses? (35 mm camera). >Any help will be greatly appreciated. >Thanks, >Oleg March
Put the camera on a focusing rail on top of a leveled pan head. Center
the lens on a distant vertical line, such as a telephone pole. Place
a vertical line close to the camera aligned with the distant object.
Now pan left as far as you can while keeping both objects on the
focusing screen. Shift the camera forward or backward to align the
objects again. Keep panning back and forth, touching up the position
of the camera to minimize relative motion of the two objects.
From Rollei Mailing List:
Date: Thu, 27 Jan 2000
From: Richard Knoppow dickburk@ix.netcom.com
Subject: Re: [Rollei] Panoramic Shooting with TLR
you wrote:
>Thanks for the detailed response! > >I'm still not sure, from the description, precisely where the front nodal >point is (ie. the axis of rotation). I do have a plate on my tripod that >will allow me to set the camera back from the tripod's axis -- do I align >the tripod's axis with the *front* of the TLR? (I assume that's what you >mean...) > >Douglas Cooper
When a lens is rotated around the _rear_ nodal point the relative
positions of objects in the image do not change. The nodal points can be
found by measurement, basically wiggling the around an axis while sliding
the point or rotation back and forth anlong the lens until the image stops
moving.
However, the focal length of a lens is defined as the distance between
the focal plane and the rear nodal point when the lens is focussed at
infinity. So, if you know the focal length of a lens one can simply
measurea distance equal to the focal length from the infinity focal plane
towards the lens. The focal plane of Rollei TLRs is marked by the chrome
stripe where the rear cover fits to the camera. Measure either 75mm or
80mm (depending of course on the lens in the camera) from this line toward
the lens. That will indicate the location of the rear nodal point. I've
found by measurement that the point of rotation of the panoram head is not
exactly at this point, nor does it come up to where the front nodal point
is, its sort of partway between. Perhaps its a compromise distance of some
sort to allow for panorams at closer distances than infinity (the inside
of a room for instance). I don't know for sure. In any case the head
works. I've shot several panoramas with both my 2.8F and with a Rolleicord
IV.
The difference in the two focal lengths is accounted for by the
difference in the camera dimensions. The panorama head is located by the
small bracket with locator pins at the front. The difference in the depth
of the focussing panel of the 80mm and 75mm cameras locates the point of
rotation at about the same point for both lenses.
For "normal" lenses like the Tessar, Xenotar, and Planar used on Rolleis
the approximate location of the nodal points can be found by dividing the
lens into three along its depth. The nodal point will be found at about
the dividing points.
By "normal" lens is meant one which is neither a telephoto or
reverse-telephoto. This rule of thumb also breaks down for completely
un-symmetrical lenses like the single elements of covertible lenses.
----
Richard Knoppow
Los Angeles,Ca.
dickburk@ix.netcom.com
From Rollei Mailing LIst;
Date: Thu, 15 Jul 1999
From: Tim Ellestad ellestad@mailbag.com
Subject: Re: [Rollei] panoramic head
Pablo -
A follow-up to my previous post.
There is often concern about just where the lens node is. Just guessing
or estimating is often incorrect, especially with retro-focus lenses.
With SLR cameras, this is fairly easily done. First - do a good job of
leveling the tripod or panning head. Second - set focus and leave it be.
Third - start panning the camera back and forth while looking in the
viewfinder, noting the displacement of objects in the foreground and
background in respect to each other. Perhaps pick a good pair of scene
elements, near and far on axis and study them as you slowly pan left and
right. Move the camera forward or back a little bit and do it again.
Did the relative displacement of the two objects increase or decrease?
If it increased move the camera the opposite direction in or out on the
focusing rail and try again. If it decreased continue moving the camera
in the same direction on the rail and try again. Keep doing this until
objects in the viewfinder all move across the frame at exactly the same
rate as you pan. You will have then found the nodal point. Shoot your
sequence.
I presume that the viewing lens and the taking lens on Rollei TLR's have
reasonably matched nodes. Does anyone know if this is true? If not, you
would have to do this with a groundglass in the open camera.
My Canon EOS 28-70 f2.8 zoom has the nodal point way out near the front
element, in spite of the fact that I may be shooting at 35mm focal length.
It puts the focal plane about 125mm behind the pivot. I know it's because
of all the zoom and retro-focus hocus-pocus in the lens but it's still
hard to believe. It works, though.
Tim Ellestad
ellestad@mailbag.com
From Rollei Mailing List;
Date: Thu, 15 Jul 1999
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] panoramic head
Actually lenses have two important nodal points. The one that matters for
panoramic work is the rear nodal point. The rear nodal point is the point
which light rays appear to come from. In some telephoto designs the rear
nodal point can be out in space somewhere in front of the lens, while in
retrofocus wide angle designs the rear nodal point can be behind the lens!
We used to locate nodal points with a gadget called a nodal slider. You
mount the lens on it and using a projected light beam measure the virtual
point which the light seems to come from after passing through the lens.
I'm a little rusty on this though, since it has been more than ten years
since I had access to an optical lab.
Bob
From Panoramic Mailing List:
Date: Fri, 23 Feb 2001
From: John Strait jstrait@acm.org
Subject: Re: Canon G1 nodal points
Hi Gerhard,
You wrote:
> I'm looking for the Canon G1 nodal points for the normal lens and the 0.8 > wide angle converter. > Any help would be greatly appreciated.
If you're unable to find the nodal point specifications for the Canon G1,
you can set it up using the "Lamp Post Test" described by James Rigg on
his page "Setting up a panorama head" at
http://www.panoguide.com/technique/setup_panohead.html . His site
http://www.panoguide.com is an excellent source for information about
panoramic technique and has reviews of all the major stitching
applications.
Regards,
John Strait
----
http://www.panoramafactory.com
Date: Tue, 01 May 2001
From: "John Futhey" jfuthey@sonic.net
Newsgroups: sci.optics
Subject: Re: Nodal Points of Fresnel Lenses?
Most Fresnel lenses are designed for a particular application rather than
as replacements for conventional optics. This ususally results in the
lens being corrected for spherical aberration, i.e., it more or less
mimics an asphere. But there are often other corrections made. Although
Fresnel lenses have some performance weaknesses, one advantage is that
each groove can be cut to any reasonable angle needed. Another factor
that goes unnoticed (and disregarded by most ray trace models) is the
"back surface" or "draft surface" angle of each groove. There are some
subtle design issues regarding how these angles are chosen, again,
depending on the specific application. Most manufacturers of Fresnel
lenses do not share the groove data base of any particular lens because
that constitutes their competitive edge (pun intended).
BTW, another thread mentions using Fresnel lenses designed to go in the
back windows of RV's. Although this might be fine for certain functions,
one must understand that these lenses approximate a concave lense, not
convex.
Good luck on your quest.
John Futhey
...
Date: Tue, 01 May 2001
From: "John Futhey" jfuthey@sonic.net
Newsgroups: sci.optics
Subject: Re: Nodal Points of Fresnel Lenses?
Hi Bob,
I would say that 90% of Fresnel lenses are made for one of the following
applications:
1. Overhead projectors (usually in pairs cemented together)
2. Back windows of RV's (negative lenses, i.e., concave)
3. Some computer projectors (ususally smaller, 1" to 6" diag.))
4. Motion detectors (usually much smaller and translucent)
5. Rear Projection TV's (fairly large, up to 50" or so)
With the exception of the Rear Projection TV application, almost all of
these are designed with two finite conjugates, although sometimes one of
the finite conjugates may be large enough to be considered infinite for
most purposes.
Back when we thought we had an energy crisis (as opposed to now when we
really do have one) quite a few Fresnel lenses were made as solar
concentrators, and these were made with one infinite conjugate. Some of
these were 50" or larger diagonal.
John Futhey
"Bob May" bobmay@nethere.com wrote
> I think that you are right there on the concave vs. convex surface! > I would find that the nodal point would more be a line than a point as the > lens is actually a number of lens segments moved in space so the outer > regions should be closer to the lens surface than the center segments. > Basically, this indicates that a Fresnel lens needs to be built for a > particular optics setup. I would also assume (with all that connotates!) > that most Fresnel lenses are designed for infinity or near infinity work as > one focus of the design. > > -- > Bob May
From Panoramic Mailing List;
Date: Tue, 07 Aug 2001
From: "M. Denis Hill" denis@area360.com
To: panorama-l@sci.monash.edu.au
Subject: RE: fence line nodal point determination
It works like this:
Mount the camera so it rotates around the center of the tripod, i.e., the
lens is neither to the left nor right of center. Aim the camera down a
fence line or any pair of aligned objects. This could be a stake in the
ground aligned with a tree behind it, the corner of a building, etc.
Rotate the camera back and forth.
If the two object remain aligned, you are rotating about the nodal point.
If they move relative to each other, slide the camera back or forward on a
focusing rail and try again.
It's simple and it works. Remember that the nodal point changes not just
with the lens, but also with the distance at which the lens is focused.
That's why the computer in a RoundShot Super want to know distance in
addition to lens.
Denis
...
From Panoramic Mailing List;
Date: Tue, 01 May 2001
From: Robert Bilsland rbilsland@yahoo.com
To: panorama-l@sci.monash.edu.au
Subject: Re: Q: Nodal pt for Coolpix990
After a quick look through my Panorama bookmarks try
these.....
http://www.edb.utexas.edu/teachnet/qtvr/NodalPoint.htm
This link explains how to find a nodal point of a
lens, fisheye or not.
http://philohome.free.fr/lbracket/lbracket.htm
When it comes to Nikon 950/990 and fisheye lens's,
Philo site is one to keep bookmarked.
http://www.360texas.com/tips/15.htm
If after all that you want to know specifications then
look here it tells you where the nodal point is.
Hope these help.
Bob Bilsland,
Worcestershire,
England.
From: dickburk@ix.netcom.com (Richard Knoppow) Newsgroups: rec.photo.equipment.large-format Subject: Re: H, H' (optics question) Date: Wed, 31 Jan 2001 jjs john@stafford.net wrote: >(asking in this forum because of good information found here.) > >I've been asked to make one, and possibly more, tripod plate adapters >to move the pivot point of the tripod screw to the 'center' of the lens >for multiple-image panoramic photographs. By 'center' I mean where the >image converges then inverts. Sorry for the nontechnical description. > >Now, I do have the specs given by Schneider, but I still don't >understand. It appears that there are two possible points: H and H'. >Am I close? Which would I use as the point in question? Or would I take >the difference between the two? (Or something else?) > >Specfically, we are beginning with the Biogon 38mm (for Hasselblad), >and that one looks (?) easy. But when moving on to the longer lenses >I'm not sure close will be good enough. > >Thank you for any help you can lend. > >(We don't want to use something like the Kaidan platforms because they >are too flimsy for this particular project.) > What you want is the rear principle point. This is where the image appears to the film to be coming from. It is defined as being one focal length from the focal plane toward the lens at infinity focus. If you know the actual (not marked) focal length of the lens you just measure that distance from the film toward the lens and there you are. Otherwise you must measure the actual focal length of the lens. With modern lenses the marked and actual focal lengths are close enough. For a symmetrical lens, like the Dagor, the rear principle point will be about one third of the length of the lens from apex to apex, measured from the rear. The Symmar is near enough to being symmetrical for this rule of thumb to hold. I get into arguments about whether a panoramic camera should rotate around the rear or front principle points. All the literature I can find suggests the rear point is the right one. --- Richard Knoppow Los Angeles, Ca. dickburk@ix.netcom.com
From rollei mailing list: Date: Fri, 7 Jun 2002 From: bigler@ens2m.fr To: rollei@mejac.palo-alto.ca.us Subject: Re: [Rollei] Panorama head question > I'm wondering if using a Rollei panoramic head Joe. There has been at least two models of the R-TLR panorama head. Basically what you need **for a R-TLR** is to rotate around a point which is located somewhere in the middle of the lens and not the bottom screw. So any modern rail like Manfrotto has on catalog will do the job to adjust to the **proper** rotating point. What is this proper point of rotation is explained below, I hope in detail (pardon me if it is too long). You can jump to the conclusions directly also. Moreover for the R-TLR there is a mechanical advantage since most of the weight is under the lens and not under the bottom screw. So the camera, mostly Planar and Xenotar (heavier than Tessars and Xenars) R-TLR cameras are much better balanced when the fixing point is in fact under the lens. > generally gives images that appear to line up properly when put > together afterwards (I'd be scanning and then joining them in > Photoshop), or if particular conditions are required, such as > avoiding certain subject matter (like regular geometric patterns, > squares etc). I'm thinking primarily of making images from two or > three frames, shooting landscapes, cityscapes and interiors. Any > hints appreciated. Joe B. The question you raise is the question of 'panoramic stitching'. There are several issues there. First, even if you do not take into account the question of parallax alignments, which are solved if you set the right rotating point, you'll not be able to render as a straight line on the image a long-long straight line of the object. In architecture, this is certainly not acceptable to have straight lines rendered as polygonal lines. In landscape this is less a problem. Now we come to the delicate and potentially controversial question of "what is the proper rotation point". For a long time I had mis-conceptions about that and what I am saying here is the result of careful examinations and discussions with some French experts so I'm pretty sure about the theory behind even if it is hard to find a detailed reference in a textbook. First, the right rotation point for the kind of panoramic stitching you need is *not* the *rear* nodal point of the lens. Yes a photographic lens is always a compound so you have to take into account the fact that the front and read nodal points are separate. The rear nodal point is used to rotate a lens barrel in another kind of panoramic cameras, where the *film* is fixed w/respect to the landscape and the *lens* is rotating. This is another story. Here with a regular camera and a "panorama" attachment, the film is fixed w/respect to the lens and both (lens + film) rotate together. The problem is different, but what is the problem ??? You want that aligned objects in the object space be rendered as aligned images on the film for **all successive images** you want to stitch. In other words, what you **do not want** is to see that somebody initially well hidden behind a tree on image #1, could be visible on image #2 because the lens has slightly shifted sideways (a parallax effect) due to an improper choice of the rotation point. In this case even the best stitching software will be helpless. So what you want is that aligned points in the object space are rendered as aligned points... well not in the image space in general but **on the detector**, a film or a digital sensor. Usually this detector is plane. And the key point is that except the only object plane which is, strictly speaking, the optical conjugate of your film (and yes you can also consider a pair of Scheimpflug slanted planes), **most objects are rendered as out-of-focus images**. Hopefully depth-of-field of course allows to a certain extent a whole volume of the object space to rendered "approximately sharp" on film. This is the core of the story. Out of focus images have no reason to obey optical conjugation rules like true optically conjugate planes (parallel or slanted in the Scheimplfulg case). So **out of focus images do not care for the position of nodal points**. The right rotation point **is not the front nodal point of the lens**, in the most general case. There is an exception for symmetric or quasi-symmetric thick lenses, see below, and *there* is a source of confusion. The right rotation point is in fact **THE ENTRANCE PUPIL OF THE LENS**. Why. Because out of focus images are built by the projection of the *exit* pupil on the film, for optical reasons difficult to detail here, you should rotate the camera around the centre of the **entrance pupil** to keep the alignments identical on successive images to be stitched. Difficult to explain without a ray tracing on a diagram, but here is the idea. Plot an entrance ray going through the centre of the entrance pupil, and consider all alignemnts in object space defined by this ray. It will exit from the lens at the centre of the exit pupil, and all out-o-focus circles originating from various aligned objects (on the entrance ray) projected by the exit pupil on film will be centered on the same film point. In other words, out of focus images will be 'aligned' (i.e. their centres will be the same) on the same film point, even if there is no optically "true sharp" image on film. So nodal points and conjugation formulae do not play a direct role in this parallax and alignment issue. This was something I refused to admit for a while. This entrance pupil is very close to the entrance (front) nodal point in a quasi-symmetric lens like a purely symmetrical repro lens of view cameras. Generally most view camera lenses are quasi-symmetric and the entrance pupil is located very close to the front nodal point. This applies to R-TLR lenses but **not** retrofocus (distagon) or telephoto lenses. For example there is a 180mm Nikon telephoto lens for 35mm cameras, where the entrance pupil is located **in the film plane *!!!! so there, as incredible as it may appear, the screw under the camera body is perfectly OK as the proper rotation-and-stitching point !!! Well several authors including the most respected ones like Kingslake duly mention the fact that the "centre of perspectibe is the centre of the entrance pupil", but unfortunateley they often take as an example a single element lens, which is a simple example of a perfectly symmetrical lens, where both nodal points are the same and are located in the pupil planes, both pupil planes being identical. I've hardly ever seen described the case of a a thick, asymmetrical compound lens on this delicate issue. Conclusions : they are very simple. - panoramic stitching with a regular camera is not appropriate for certain subjects where you do not want a long straight line rendered as a polygonal line. - to avoid as much as possible unwanted parallax effects, you should rotate around the centre of the entrance pupil of the lens. For a quasi-symmetrical lens like a R-TLR lens this is located a few millimetres in front of the diaphragm. For modern lenses precisely documented by the manufacturer like MF SLR Zeiss lenses, the position of the entrance pupil is written in the datasheet and I suggest that you browse through those Zeiss datasheets to make your own idea about the difference between a quasi-symmetrical lens design and a retro-focus or a telephoto as far as the position of the pupils is concerned. - once you know approximately where the entrance pupil is located, you can refine the position of the proper rotation point if you have an adjustable rail. There the ground glass attachment of the R-TLR may be helpful although unwanted parallax effects will be very hard to see on a 56x56 ground glass. I hope this is understandable, and above all, not controversial ;-);-) -- Emmanuel BIGLER bigler@ens2m.fr
From rollei mailing list: Date: Fri, 7 Jun 2002 From: bigler@ens2m.fr Subject: Re: [Rollei] Panorama head question, distorsion on edges From Joe B: > >I'm wondering if using a Rollei panoramic head... >From Richard K: > The images won't merge evenly at the edges due to the the lens being > rectilinear. Yes Richard ; this is an additional effect I did not mention in my long post. The proper choice of the rotation point as hopefully provided by one of the R-TRL panorama head models, unfortunately, cannot do anything for this either. For those who can read French there is an excellent article on the subject of panoramic stitching with a regular camera by Ronan Loaec in "Chasseurs d'images" No 232, april 2001, pp60-73. BTW on pp 50-53 of the same issue there is a nice comparison of used and new medium format cameras. Including of course Rollei MF cameras. All the article on panoramic stitching is excellent and without maths. Nothing to object to the contents in practice. The author clearly shows on pictures what can be done. And a lot of excellent pictures can be achieved even if some professionals do object against polygonal lines and edge distorsion. The parallax issue is very well demonstrated experimentally, very convincing. Since the author --a very minor point-- was fuzzy about the definition of the proper rotation point, I tried to find what was behind and after several errors and wrong ideas on my side (like the entrance nodal point) I was eventually taught the entrance pupil story, thanks to a passionate e-mail exchange on a French Internet photo-forum. -- Emmanuel BIGLER bigler@ens2m.fr
From: dickburk@ix.netcom.com (Richard Knoppow) Newsgroups: rec.photo.equipment.large-format Subject: Re: Focus calculation for lens and bellows length? Date: 10 Sep 2002 "Matt Clara" no.email@thisguy's.expense wrote... > "Richard Knoppow" dickburk@ix.netcom.com wrote > > "Ed Margiewicz" Edmarg1@comcast.net wrote... > > > Hello, > > > Is there a formula to calculate how close a certain lens will focus at a > > > certain bellows length? For example, is there a way to calculate how > close > > > a 300mm lens will focus at 38cm from the film plane? > > > Thanks, > > > Ed > > > > > > > > The basic equation is: > > > > 1/F = 1/u + 1/v (1) > > > > Where: > > F = Focal length > > u = subject distance > > v = image distance > > > > Lens to image distance = v = Fu / u-F (2) > Just to be clear, when you say "image distance", do you mean the film plane? Approximately. The image appears to come from what is called the second principle plane of the lens. The actual location of this point depends on the type of lens. For most "normal", i.e., not telephoto or retrofocus, lenses it will be somewhere between the iris diaphragm and the back end of the lens. In fact, its distance is exactly one focal length from the image to the lens, so, it you know the focal length fairly accurately, the location of the rear or second principle point can be easily found. For most purposes, where only approximate calculation is needed, the principle point can be assumed to be at the diaphragm or the middle of the lens barrel. -- Richard Knoppow Los Angeles, CA, USA dickburk@ix.netcom.com
Date: Sun, 27 Oct 2002 From: Peter Rosenthal petroffski@mac.com To: hasselblad@kelvin.net Subject: Re: [HUG] Notal Point of a Lens. > Does anyone know how to determine the notal point of a > lens? Also does anyone know of a tripod head that you could adjust the > rotation of the camera/lens around a notal point of a lens. > > Ken Martin Hey Ken- The nodal point of the lens is relatively easy to find. I should say first tho' that if you will be shooting at infinity (everything in your pano at infinity), you don't need to find the nodal point. The nodal point of any lens is that point at which all of the rays cross. In most lenses this is somewhere in the front 1/3 of the optical barrel. It's not where you think it would be. The whole point being that when you rotate the camera the foreground must not shift relative to the background. If it does, NO pano software can compensate. Each pano section photo will be different than the next. To get started I'd put my camera and lens assembly on a small lazy-susan with the axis of the lens directly over the center of rotation (mark the center with a pencil or some-such). A LOT of light will help. Set up a pencil vertically, about 18" or so in front of the camera. While looking through the VF, rotate the lazy-susan. You'll notice that the pencil will shift relative to the background as you rotate it. Move the camera back or forward until the pencil remains stationary relative to the background as you rotate the lazy-susan. The center-of-rotation is now directly under the nodal point. Measure the distance from the center point to the front of the filter ring. That will help you figure out how far the tripod socket is from the center of rotation. Do this for each lens you figure you will use. There is no randomly chosen tripod head that will compensate for this foreground/background shift. I've made my own adapter to compensate for my 28mm (dinky 35mm format) vertical format lens. I've built in a slot to move the camera forward 1/4 inch or so for my 21mm for indoor panos. There is a lot of 35mm and digital pano work being done but am unaware of too much MF pano work. The files will be huge! Not to mention the processing power required to stitch these big suckers together. It will make your computer think hard!! Be sure to overlap each photo 50% or so. Less than that is dicey. Trust me. With my vertical 28mm, I have to shoot 14 frames to overlap each 51% (I've indexed the ring to let me know how far to rotate). Maintain the same exact exposure for each. Here are links for pano stuff... These people are THE authority on such things. It's endless. http://www.kaidan.com/products/pano-prods.html http://www.kaidan.com/industry.html Have fun, go nuts, spend money! Peter Peter Rosenthal PR Camera Repair 111 E. Aspen #1 Flagstaff, AZ 86001 (928) 779-5263
From: westin*nospam@graphics.cornell.edu (Stephen H. Westin) Newsgroups: sci.optics Subject: Re: Front nodal point of camera lens Date: 24 Feb 2003 "jriegle" jriegle@att.net writes: > I have a 50mm camera lens of 6 elements. How can I find the front nodal > point (or estimate). I understand I have to rotate around this point for > indoor panoramic shots (stitching several together) to avoid the shifting of > the image of closer objects when rotating the camera. > > Thanks, John The most ingenious method I've seen for this was in a computer graphics paper by Paul Debevec; I have no idea if it was an original idea of his, or if he copied it from someone else. He set the camera on a flat surface so that the lens axis was parallel to the plane of the surface. He then drew lines on the surface that appeared exactly vertical in the image. The lines converge at the fornt nodal point of the lens. Of course the accuracy of this technique depends on how accurately you can measure verticality in the image, but it requires a minimum of equipment. And if you are capturing a panorama of a distant environment, small errors won't matter much. -- -Stephen H. Westin
From: George giocar3@libero.it Newsgroups: sci.optics Subject: Re: Front nodal point of camera lens Date: Mon, 03 Mar 2003 "jriegle" jriegle@att.net wrote:t >I have a 50mm camera lens of 6 elements. How can I find the front nodal >point (or estimate). I understand I have to rotate around this point for >indoor panoramic shots (stitching several together) to avoid the shifting of >the image of closer objects when rotating the camera. > >Thanks, John In this article, there is a pratical method to find this point: http://www.funsci.com/fun3_en/panoram2/pan2_en.htm The method is based on the rotation of the objective on a vertical axis and to shift the objective until the image formed is stationary on the film. According to R. Kingslake: "This point coincide with the second nodal point,"... (Applied Optics and Optical Engineering, 1965, Vol 1, p 220). Are we sure this point is not a nodal point but "the center of perspective of the lens - which is the entrance pupil"? George
From: M. Denis Hill [denis.hill@verizon.net] Sent: Tue 5/13/2003 To: panorama-l@sci.monash.edu.au Subject: RE: Nodal points: was Fuji TX-1 It's late, so here's the short version. I'll put together something with illustrations on my website when I get a chance. Whatever the camera type you have to look through the lens. For an SLR, look through the prism. For a view camera, look through the ground glass. For other cameras, e.g., XPan, Fuji G617, put a ground glass at the film plane and use a dark cloth. Focus at the distance at which you plan to shoot. As a starting point, move the camera back so the film plane is the focal length behind the rotation point. So, if you are using a 20mm lens, slide back so the film plane is 20mm from the center column of the tripod. Find a "picket fence" or equivalent to look down. I often use two light stands. I place them so, when I look through the lens, one is hidden by the other. I usually place them about 10 and 20 feet from the camera. I suspect that you'll get the best results if you place the targets at the nearest distance of interest in your planned shots and the greatest distance. I've not proven this. If you can find a picket fence, look along it so all picket tops are in a straight line going away from the camera. Now rotate the camera left and right. If the position of the pickets (or light stands) relative to each other stays the same, you are rotating around the nodal point. If the rear light stand peeks out to the right or left of the front light stand as you rotate, you're off and need to slide the camera forward or back relative to the center of the panning stage. That's it. Once you've done this, you can visualize why it's important to do. Take it a step further and picture a rotational camera smearing the image if it is not rotating around the nodal point; as the camera moves, objects move relative to other objects at different distances in the image. M. Denis Hill Qualified Panoramic Photographer
From: Stefano Rumi [stefrumi@tin.it] Sent: Thu 2/19/2004 To: Lenses@topica.com Subject: RE: [LENSES] [despammed] RE: Cosine to the 4th explained The rear nodal point should be a focal length away from the film plane! As far as I understand, the front and rear nodal points are two fictitious points into which (front nodal point) light from the scene converges and from which (rear nodal point) it diverges to reach the film plane. (Quite badly stated, since a lens makes light CONVERGE, hopefully, on the film plane; but I hope I was clear enough). What is not clear to me is this: if light behaves AS IF it came from the rear nodal point, fall-off would not depend on design but on focal length only; on the other hand, the last moment light is refracted is when it leaves the rearmost element of the lens, and that's further away from the film plane in a retro focus design, justifying the hypothesis of a better illumination evenness. Doubtfully yours, Stefano
End Of Page