Cant get it right

[Juz H]

OK, so Ive been to Forums, check out the set up methods, done the
flower head and wire rack etc etc but I cant get it right. I always
end up with my horizontals having vertical misalignments in them ie a
wall will have a misaligned join in it. Trying the wire rack set up I
couldnt get it perfect. I was down to changing adjustments by 1 mm.
Always the same prob with misaligned joins.
Im using my old Nikon D70 with a Sigma 10-20 on it, always set to
10mm. Im lost as to what to do. I have a Nodal Ninja 5 with a R16
head. Oh and Ive tried different rotator degrees.

Am I missing something basic here??

Looking forward to any help. Its driving me nuts. Ive been at this for
many many months and now im looking for some help please.

Cheers

Justin

[Mick Crane]

Hans has suggested using this lens at 11 rather than 10 I have yet to
try that.
You might try it.
Mick

Sent from my iPhone

> --
> You received this message because you are subscribed to the Google
> Groups "PTGui" group.
> To post to this group, send email to pt...@googlegroups.com
> To unsubscribe from this group, send email to ptgui+un...@googlegroups.com
> Please do not add attachments to your posts; instead you may upload
> files at
> http://groups.google.com/group/ptgui/files
> For more options, visit this group at http://groups.google.com/group/ptgui

[תמיר- סטודיו הוד‎]

the RD16 is not the right head for this lens
RD16 is for lens from 40 till 200 mm
you should work with the RD8.

Tamir

 

--
You received this message because you are subscribed to the Google Groups "PTGui" group.
To post to this group, send email to pt...@googlegroups.com
To unsubscribe from this group, send email to ptgui+un...@googlegroups.com
Please do not add attachments to your posts; instead you may upload files at
http://groups.google.com/group/ptgui/files
For more options, visit this group at http://groups.google.com/group/ptgui

--
בברכה
תמיר אורבאום
050-3278232
http://www.pbase.com/tamiror

[Keith Martin]

Sometime around 14/12/09 (at 17:42 +0200) ™Óȯ‚ÝÒË“ÈÂݔ“ said:

>the RD16 is not the right head for this lens
>RD16 is for lens from 40 till 200 mm
>you should work with the RD8.

Strictly speaking, the RD16 and RD8 are rotators,
not heads. Justin is using a Nodal Ninja 5 head
on the RD16 rotator, which is a very flexible
combination of kit and something that should work
well for him.

FYI, the RD16 covers all the stop-increments that
the RD8 does (4 to 20) AND it also allows stop
increments from 24 to 96. That is what makes it
well suited for use with longer focal length
lenses. But it is just as good as the RD8 at
working with shorter lenses.

k

[Rene Beaumier]

Hy Justin, I will finish today ou tomorow my own maner, I will put
that on the web and it will probably give you a good way to work.
Patience, Rene Beaumier
Le 09-12-14 à 09:52, Juz H a écrit :

[Rene Beaumier]

That is my own way.
It is not the absolute way, OK.
Material Canon, 1.6 sensor, lens Sigma 10/20 zoom, Pano head Nynja
basic, Rotation control Manfroto 300N, Mac pro on 10.5.6, Ptgui
version 8.3.4, no extra pano tool.
The nodal point are find and chect 1 time by 6 mounth just in case if
something can be broken in the lens.
Dont forget: No lens are absolutly perfect, It is always possible a
glass lens inside the zoom have his glue broken or other funny thing.

First: check if my lens are absolut clean, my captor also.
Put the head on tripod.
Set my 300N at 12 angles ,
Put the Nynja in place and LOCK IT.
Put my camera in place on the Nynja, add the lens hood, set the
adjustment for the nodal point.
Put the remote cable in place and the camera at the level.

PT Gui dont work with out of focus object.
Adjust your zoom at 10mm, put your focus the nearest object and the
farest object in your view.
Put your diaphragme in the center of the barrel.
By example your D. are f/4.5 to f/ 22, f/11 are the point who usaly
the lens are the more sharp as possible.
With my manual light meter, I set my exposition to fit with f/11.

That it for the mecanic.

For a 360 degres I will take 2 rows with 12 angles in RAW
With PSD, Camera Raw, I process with output in Prophoto 16 bits.
Drop that in PTGui, and Voila.
For a partial pano I take the quantity of pictures to covers the space
with an confortable over, just in case.
One or two row, but I prefer keeping as simple as possible to avoid
the possible distortion or the un-normal looking.
Patience and tryals with the old Egyptian computer material:
Paper and pencil.
Bye, René Beaumier

Le 09-12-14 à 10:58, Rene Beaumier a écrit :

[John Houghton]

Justin, I don't know how experienced you are with using PTGui, but
since stitching errors are often caused by less than adequate
processing rather than faulty pano head setup, it might help if you
can make a set of jpeg images available for testing. Often, the cause
of the problems can be quickly identified.

John

[Eric O'Brien]

This is a good point.

Justin should also tell us the process he currently goes through in
PTGUI. There are NUMEROUS paths than can be taken. Not all will
create pleasing results.

eo

[PTGui Support]

Hi Justin,

Please see these two FAQ articles:

4.3:
http://www.ptgui.com/support.html#4_3

5.4:
http://www.ptgui.com/support.html#5_4

Joost

[Juz H]

Wow, you guys are awesome! Thank you all for attending to my
questions.

I have read and adjusted for Sigma Lens as per the ptgui links. ie i
already had that set up.
Thanks Rene for your detailed list.

Eric, my process is usually to load images, check the rectangular
setting is correct, if N and Z present set the 90 / -90 then hit
align. If needed i will add a few points and or horizontal or vertical
alignments. Optimize and ... thats it usually. If I want to play with
lighting etc I do that in CS4.
I have done testing with numerous sticks and windows and wire etc to
the point that I cannot see 1 pixel of difference form left to right.
Then I do the stitch and get the same old errors.

John, it would be great of you could have a look at some images for
me. How do I post them?

I love doing panoramic landscape work but have had a lot of pain in
fixing the stitches. Ive done quite a few. I want to be able to do
interiors, but, well, here I am.

[michael crane]

2009/12/15 Juz H <just...@netspace.net.au>:

> I have done testing with numerous sticks and windows and wire etc to
> the point that I cannot see 1 pixel of difference form left to right.

I would have difficulty to make this statement with this lens as the
image twists so much.
Which part of the images do you compare ?
As I understand the non parallax point thing varies according to where
the light is coming from.
I usually setup by comparing a place 1/4 in from the left with 1/4 in
from the right.

mick

[John Houghton]

On Dec 15, 12:21 pm, Juz H <justi...@netspace.net.au> wrote:
> John, it would be great of you could have a look at some images for
> me. How do I post them?

It's best if you pack jpeg copies in a zip file (if necessary, use
www.7-zip.org ). Assuming it will be a largish file, maybe use
www.filedropper.com or similar free service to upload it if you don't
have private web space, and send the link to me at the email address
in my profile, or post it here.

John

[Juz H]

Hi Michael, sorry mate I assumed that it would be taken for granted
that my testing wouldnt include the outer edges of the lens.Testing
was done to the limit of where I could see lens distortion starting to
come in. What Im getting at is that there was no parallax errors for a
majority of the lens, usable lens :)

OK John Ill get some images and send them. thanks.

Justin

On Dec 15, 10:54 pm, John Houghton <j.hough...@ntlworld.com> wrote:
> On Dec 15, 12:21 pm, Juz H <justi...@netspace.net.au> wrote:
>
> > John, it would be great of you could have a look at some images for
> > me. How do I post them?
>

> It's best if you pack jpeg copies in a zip file (if necessary, usewww.7-zip.org).  Assuming it will be a largish file, maybe usewww.filedropper.comor similar free service to upload it if you don't

[michael crane]

2009/12/16 Juz H <just...@netspace.net.au>:

> Hi Michael, sorry mate I assumed that it would be taken for granted
> that my testing wouldnt include the outer edges of the lens.Testing
> was done to the limit of where I could see lens distortion starting to
> come in. What Im getting at is that there was no parallax errors for a
> majority of the lens, usable lens :)

I don't know I follow what you mean.
What I meant is.
put some tape on a window and photograph it
once so the tape is in the left of the viewfinder about a 1/4 in
second spin the camera so the tape is on the right of the viewfinder
about a 1/4 in
compare the images with regard to shift of stuff in the distance
compared to the tape.
mick

[John Houghton]

On Dec 16, 10:30 pm, michael crane <mick.cr...@gmail.com> wrote:
> compare the images with regard to shift of stuff in the distance

Justin sent me a set of images which stitched very well, but through
some oversight only cover about 80% of the 360x180 view. There is
evidence of a small amount of parallax, which should be possible to
eliminate by a small adjustment of the camera position:

http://www.johnhpanos.com/justin.gif

John

[Eric O'Brien]

Watching the foliage, some of the position changes between frames
looks like it was a breezy time when the shots were taken. As people
often mention, if control points were generated in those moving areas,
the result can be a bad stitch. (The same problem can exist with
waves on a water surface, or fast moving clouds.)

On a tangent, I will mention my guess/feeling/idea that many bad
stitches are incorrectly blamed on a poorly place no-parallax
("nodal") point, when the *actual* cause is that the lens optimization
isn't quite good enough.

There is a HUGE amount of discussion about that no-parallax point, and
it's no wonder early users tend to work on that item. Often to great
frustration. :(

PTGui, particularly with fisheye lenses, generates control points
mostly toward the center of each image. For a full frame fisheye
lens, the corners often have NO control points at all. Without
control points in the areas of the image that need to be most severely
remapped, it is easy for those areas to be remapped not... quite...
correctly. And those areas often are where the stitching actually
occurs!

So... you need to manually add some control points in those areas
before you try to "calibrate" a lens.

eo

[John Houghton]

Another factor causing bad stitching in Justin's panoramas turned out
to be upside down nadir and zenith images, caused by unreliable data
from the camera's orientation sensor. Switching off the auto rotation
feature in the camera is a simple and effective solution.

John

> > For more options, visit this group athttp://groups.google.com/group/ptgui- Hide quoted text -
>
> - Show quoted text -

[Juz H]

Thanks everyone for your responses, in particular John.

So far I think I have set up my camera fairly well but there are
things in PTGui that I have yet to master. Ill try to do a "what Ive
learnt form this" response when I have actually worked it all out. Im
close but not yet.

> > > For more options, visit this group athttp://groups.google.com/group/ptgui-Hide quoted text -

[Kevin Wilton]

Juz

If the camera is spot on, the rest is a doddle. Go back to
your nodal point.

Kev