Here is my last one for today. I hope I got nailed.
That looks fine, if not actually perfect. You should waste no more time and start taking some real panoramas. Any stitching errors that you get will not be due to the setup of your panohead.
Adamo, there is no offset for nadir Adaptor on NN5.
Offset on NN3 is 6mm, not 3.
to me, as for john, it looks quite good for lower rail setting. What is your URS you used?
There is always the need to place some manual CP to the "outer circles of the fisheye lenses" and to set them along the verticals and horizontals. When you have got a very good result, save this as a template. This will reduce the amount of placing CP manually. But when I remember right you normally are working with Autopano?
I will reload my pics of D300s Samyang 8mm with orange laser placed dots in horizontal and vertical to set up a template.
There is no need to get rid of the "hexagon outline", the reference for the LRS is the round circle of the panohead. Or do I misunderstand you?
My LRS is 59.5-60 and URS is 97.
I thought, looking at Smooth method pictures that there should not be any hexagon outline, so I wanted to move back URS to 98-99 and eliminate this hexagon. Also, looking closely the very bottom of tripod legs are kind of mangled out of shape.
I want to see your laser method to find NPP. Please, post it so I can downloaded it.
Have to say that I'm learning a lot from you and John.
Do you feel better?
From the stitch of your 6 images, the upper rail setting looks to be good, in so far as there is no parallax shift visible between shots. However, there is nothing really close to the camera in any of the overlaps to properly assess this. You should take some shots specifically to check for parallax in order to validate the URS (or maybe check with live view).
To get some close objects I tried grid method. The result is not bad, but not perfect. I can see some misalignment next to center button to the left of it.
Also, I uploaded images from the grid method. I wonder, if this misalignment can be fixed in PTGui. Would you be kind to send me PTS and your stitch, if you decide to try it.
I can't thank you enough for all your help. Now my panoramas start to look as they should.
Here is the link to images:
To judge the alignment of images, it is best to output the stitched image as a psd layered file. You can then readily see misalignments over whole of an overlap region between two image layers in Photoshop by switching the upper image off and on. The line where the seam will be placed by the blender is the main region of interest, as that is where any misalignments are most likely to be visible. This is generally half way across the overlap.
In taking shots for checking parallax, you should shoot with the same angular increments that will be used for your actual panorama shoots. In your case you are taking 6 shots around, incrementing yaw by 60 degrees. So take your shots of the grid at 60 degree increments, not 30 degrees as you have done. Obviously, you can omit the central image of three so that you then have two taken at the required 60 degrees.
Your setup looks to be adequate for most purposes, so I suggest you try taking some real panoramas.
Thank you for your insightfull reply, as always.
I just followed instruction doing drid method. Don't know any better yet.
You didn't mention where can I get some info how to shoot and stitch nadir.
In low light conditions and when shooting bracketed shots for HDR processing, shooting with a hand held camera is not very practical, because of camera shake and alignment problems. Methods of dealing with these are described at:
http://tinyurl.com/2we8hwq (Nodal Ninja nadir adapter)
Stitching a nadir shot presents problems because it's difficult to maintain the position of the entrance pupil exactly when shooting, so you get parallax problems and changes of perspective. PTGui can cope with this very well when the nadir area is a flat floor by using the viewpoint correction feature in the Pro version. There's a tutorial describing this process on the PTGui web site, and I have one at:
Otherwise, you can manually align and blend in a nadir shot by extracting a rectilinear view of the nadir area from the stitched equirectangular panorama image, patching over the nadir hole, and then returning the edited nadir view to the equirecangular image. There are many ways of doing this, e.g:
Convert the equirectangular image to 6 cubic tiles (e.g. with Pano2VR, Pano2QTVR, PTGui), edit the nadir tile, rebuild the equirectangular image.
Use the patch facility in Pano2VR
PTEditor from the Panorama Tools package
Super Cubic plugin by Rune Spaans: http://www.superrune.com/technical/s...eretocubic.php
PanoTools Wiki tutorials: http://wiki.panotools.org/Zenith_and...iting_overview
When using the cubic tiles methods, you need a defished nadir shot to patch in manually. You can use PTGui to do the defishing by including the nadir shot in the full panorama project. Align it as best you can with the optimizer (with all the image and lens parameters unchecked for the other images). Then output two equirectangular stitches: one containing only the nadir shot and one containing all the other images. You can then convert these to cubic tiles and merge the corresponding nadir tiles.
I'll review all referenced material you kindly provided.
Do you think that nadir adapter is something I should consider?
BTW, I prefer shooting panoramas that can be printed vs. VR format At least for now.
Since NN5 is discontinued I may not be able to buy it later.
Where can I see tutorial about CP placement for good stitch?