HELP! Finding the nodal point on the A7II

[etrigan63]

OK panoramas experts, I have been driving myself nuts trying to find the nodal point (no parallax point) on my A7II. I have looked at all of the videos, read all of the articles and nothing works. Using the Two Pole Method shows no movement when I pan the camera (indicating parallax) regardless of where the lens is over the center of rotation.:wtf:

Is there some special trick to getting this to work on the A7 II? I don't own a laptop so I can't look at a live view on a big screen.

HELP!

 

[Vivek]

Not sure if it would help- turn the sensor steady shot (aka IBIS) off?

 

[hiepphotog]

It helps to state which lens you have tried. The two stick method works really well with anything from 50 and up, while it would not work so well with WA and UWA.

One method that you can deal with these WA and UWA is by using a laser pointer: http://michel.thoby.free.fr/Laser_LPP/New_way_to_set_LPP.html

Good luck!

 

[mjm6]

The nodal point is a function of the lens, not the sensor (for the most part, the filter stack probably has a nearly imperceptible impact on the nodal point), so you must really be asking this about the lens you are using, not the camera.

Here is a page that has a few methods to do this, depending on how precise you want to be:

http://www.johnhpanos.com/epcalib.htm


---Michael

 

[etrigan63]

I was testing with a Canon FD 35mm f/2 lens @ f8. The laser method won't work for me because I don't have the vertical part. I will look at the John's panos site.

 

[Jack]

Carlos, assuming you are focusing near infinity, start at about the 1.5x the lens focal (52mm or 2-inches in this case) from the sensor on the camera, and that should be pretty close. Have the close reference point fairly close to the camera, like 5 feet.

 

[mjm6]

Carlos,

Look at the example they have using a piece of tape on a window to test the nodal point. I think this approach make the most sense for normal stitching uses. It is at the bottom if I recall.

---Michael

 

[etrigan63]

Carlos, assuming you are focusing near infinity, start at about the 1.5x the lens focal (52mm or 2-inches in this case) from the sensor on the camera, and that should be pretty close. Have the close reference point fairly close to the camera, like 5 feet.

Wouldn't I have to compensate for the adapter? Also, what projection should I use when stitching it together? No matter what I do, I keep getting the middle bulging towards me.

 

[jerome_m]

The nodal point of a lens is the position of the entrance pupil. If you look into the lens from the front, you will see a diaphragm at some distance inside. Estimate how far that is into the lens as seen through the front glass and you will have a good start for your nodal point.

Also, what projection should I use when stitching it together? No matter what I do, I keep getting the middle bulging towards me.

The most used projetions are:
-rectilinear, which keeps straight lines straight but is not very good for very large angles of view
-cylindrical, which is usually good for large angle panoramas.

Here a small discussion on projections: Projections - PanoTools.org Wiki

 

[pegelli]

Wouldn't I have to compensate for the adapter? Also, what projection should I use when stitching it together? No matter what I do, I keep getting the middle bulging towards me.

Carlos, no need to compensate for the adapter, you also don't compensate for the "built in" mirror box when using a DSLR. It's pure a function of how far the lens is from the sensor at the appropriate focusing distance, and this is always the same.

The middle bulging towards you is not happening because of not turning around the nodal point, but is set by your viewpoint, amount of turn and projection method. I'm not a specialist, but this tutorial on Cambridge in Color is an excellent starting point.

 

[mbroomfield]

Perhaps your front "pole" is not close enough. Set up a test scene with the front pole only a couple of feet away, with the distant pole 50+ and you should easily see the difference as you slide the camera back and forth.

If you want to eliminate the front bulge you might be better off using a longer FL lens, standing further back, and more stitches with a cylindrical projection.

 

[mbroomfield]

Can you post some scenes you've been trying?

 

[mbroomfield]

Carlos, try 63mm (or at least start there). This happens to be right over the green distance scale.
I used my A7r and 35/2 FD to take the following set, in order;

• 
  [*]Full image, I lined up the edge of the vase with the tree leaving a little light between them as the vase was a little blurred. Still pretty easy to see the gap between them.
  [*]100% view of center
  [*]100% view when pointed all the way to the right
  [*]100% view when pointed all the way to the left. Perhaps not perfect but pretty close
  [*]I then slid the camera forward by 20mm to exaggerate the misalignment
  [*]100% view when pointed all the way to the right
  [*]100% view when pointed all the way to the left

 

[Jack]

Wouldn't I have to compensate for the adapter? Also, what projection should I use when stitching it together? No matter what I do, I keep getting the middle bulging towards me.

Ah. In a "spherical" or pan-captured panorama, even when perfectly nodal-correct, you will get the center bulge -- that's optical physics. So a couple ways to deal with that. Most stitching software has a "flat projection" option of some kind -- they may call it something else depending on the software -- and you need to use that to render a regular flat looking capture from a panned capture. BUT! Be aware that the software will now squeeze the center bulge significantly to accomplish that, so your image will need to be cropped back to rectangular, and this in turn will truncate image material at the outside edges. Hence, when you want a flat rendering from a panned capture, you need to shoot a LOT looser than you think, meaning use a significantly wider lens, like 1/2 whatever your visually correct focal length is for a 180 degree pan.

 

[mjm6]

Jack, I think he means the bulge in the pano resulting image from cylindrical projection...

But you are right about the optical distortion associated with retrofocal lenses causing an issue as well, albeit a different issue to contend with.


---Michael

 

[dmward]

Here are the two items I purchased from Amazon for doing nodal point panos with my A7II

Amazon.com : SUNWAYFOTO DDH-03i Fluid Panning Tripod Clamp Arca Compatible DDH03i Sunway : Tripod Heads : Camera & Photo

Amazon.com : SUNWAYFOTO MP 140mm Rail Nodal Slide Arca Compatible Clamp DMP-140 Sunway : Tripod Camera Mounts : Camera & Photo

The rail has its zero point near the middle of its travel with 45 away from camera (+) and 35 toward camera (-).

I did a test with the 16-35, 24- 240, 24-40 and 28m. The test with the 24-240 was at the 24 and 50mm zoom markings on the barrel. The other two zooms were at each extreme.
All three zooms needed up at +30 and the 28mm at 0.

I did the test with a fiberglas garden stake about 3 paces in front of the tripod and the vertical of a sliding glass door about 10 paces past the stake.

I then imported and merged the frames in Lightroom CC using the merge to pano feature.

There may be a more precise approach but the results look fine to me.

 

[etrigan63]

Here's a test with the Canon FD 50mm f/3.5 Macro lens:


Panorama Test by Carlos Echenique, on Flickr

7-shot stitch using Photoshop CC. I chose this lens because Jack recommended a longer focal length and because of Jerome's suggestion regarding the exit pupil. The iris on the FD 50mm f/3.5 Macro is right at the aperture ring.