Panoramas and nodal points

Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

Bob Thomas wrote:

Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

That depends on the stitching software that you use. Most of them can
handle minor alignment problems in all 3 axis. Some of them can deal
with rather significant problems. A few can't handle much at all.

I would suggest you try it and see.

I have a Manfrotto pano head that does a great job. I use Hugin and
Photoshop CS for stitching. I will also shoot hand held panos, when the
tripod isn't there or practical. Hugin usually handles it without much
problem. PS CS isn't too bad either - well, for less than 360 degrees.

Clyde

Try www.PanoramaFactory.com. Infinitely better than PS.

Graham


"Clyde" wrote in message
news:TBdEc.129378$HG.73390@attbi_s53...
Bob Thomas wrote:

Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

That depends on the stitching software that you use. Most of them can
handle minor alignment problems in all 3 axis. Some of them can deal
with rather significant problems. A few can't handle much at all.

I would suggest you try it and see.

I have a Manfrotto pano head that does a great job. I use Hugin and
Photoshop CS for stitching. I will also shoot hand held panos, when the
tripod isn't there or practical. Hugin usually handles it without much
problem. PS CS isn't too bad either - well, for less than 360 degrees.

Clyde

In article ,
says...

Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

You don't need to be too accurate if you are taking pictures at a
distance. The only time you might run into problems is if the subject
were fairly close (say less than 3 feet).
You don't really need a pan head if you tripod head has a pan motion.
Just get a slotted bar like those used on flash brackets and use it
to adjust the point at which the camera is positioned over the axis
of the tripod.
Make yourself a little protractor so you can turn the camera in even
increments. Narrow field of view lenses make pictures that are easier
to stitch than ones with wide angle since there is less distortion
at the edges. I find Panorama Factory does a good job fixing all the
problems and I use an ultrawide angle lens with an angle of view of
120 degrees.

--
Robert D Feinman
Landscapes, Cityscapes and Panoramic Photographs
http://robertdfeinman.com
mail:

Bob Thomas wrote:
Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

I have lost track of it, but somewhere recently I read a cogent argument
that for DigiCams the normal-to-nodal-point distance has shrunk to
insignificance, and that even in ordinary camera scale, the requirement
that pivot and nodal points coincide was always overrated, anyway.

So, who was it said that? Speak up, please.


Frank ess

Yes the closer the better - but even with perfect there will still be some
work to do.


--
http://www.chapelhillnoir.com
home of The Camera-ist's Manifesto
The Improved Links Pages are at
http://www.chapelhillnoir.com/links/mlinks00.html
A sample chapter from my novel "Haight-Ashbury" is at
http://www.chapelhillnoir.com/writ/hait/hatitl.html
"Bob Thomas" wrote in message
...

Hi,

I'm building my own pano head for taking panoramas - and it might not
be an engineering masterpiece.

Would I be right in saying that the accuracy and effectiveness of the
pano head will be in a direct ratio to my accuracy in building it?

I.e. Would I be right in saying that if my pano head allows rotation
*nearly* around the nodal point, the image will be *nearly* as good as
if I were 100% accurate? Or is it all or nothing ?

Bob T.

Clyde wrote in news:TBdEc.129378$HG.73390
@attbi_s53:

That depends on the stitching software that you use. Most of them can
handle minor alignment problems in all 3 axis. Some of them can deal
with rather significant problems. A few can't handle much at all.

Absolutely not! No stitching software can handle pictures
better than any other if there are probblems with not
rotating around the nodal point. Tilting and rotation, yes,
then you want to use Panorama Tools - it is the best one
for dealing with those kinds of problems. But - none can
handle nodal points mismatch.


/Roland

"Tony Spadaro" wrote in
. com:

Yes the closer the better - but even with perfect there will still be
some work to do.


I would rather phrase it like this

* For panoramas with only far away objects - don't bother with
any nodal points at all - it is not important. You can, if
you want - try to rotate around the camera at least.
* For panoramas where the nodal point starts to become important,
any crude device that approximately rotates around the nodal
point is of extremely high value.
* I would be surprised if you would find any situation where the
crude device is not enough. Objects must very near to the camera
indeed.


/Roland

"Tony Spadaro" wrote in message
. com...
Yes the closer the better - but even with perfect there will still be
some
work to do.

If you use software like PanoTools it is unlikely to be an issue. That
will do all the "work" for you.

Cheers,
David

"David J Taylor"
wrote in :

"Tony Spadaro" wrote in message
. com...
Yes the closer the better - but even with perfect there will still be
some
work to do.

If you use software like PanoTools it is unlikely to be an issue. That
will do all the "work" for you.


Yepp - why would it be so extremely important to find
the exact nodal point?

Let us say that you miss the nodal point with 3 mm (thats
a lot of you make a crude bracket). Let us further say
that you rotate the camera 20 degrees between photos. Then
the (virtual) nodal point will move 3 mm*2*tan(20/2) = 1 mm
sideways.

This means that nearby things will be 1 mm wrong with regard
to infinity. Using a 50 mm lens (35 mm equiv) and a spacing
of 10 um (reasonable for a full frame 35 mm digital camera),
then 1 mm will correspond to one pixel wrong at 5 meters.

So - if you have objects nearer than 5 meter you must find
the nodal point with an accurace of 3 mm to assure better
than one pixel accuracy.


/Roland