Tutorial: How To Find Your Lens Nodal Point

Andy

David_S85 :s85 made up this awesome tutorial on how to find your lens' nodal point, for perfectly stitchable panoramas. I just did it, and it works perfectly!

TIP: If you have no Pez dispensers, two candlesticks in candleholders will do nicely :lol3

:wave thanks David!

:clap :clap

DoctorIt

Mmmm, Pez! nice work Dave

wxwax

Nodals?

I thought this was a family site? :patch

David_S85

Thanks guys. I had lots of other PEZ at my disposal -- many with much pointier heads, but Homer and Marge just seemed so right for this.

For those replicating this test with your own lenses, cams and pano heads, you should note that you don't even need to take any pictures to perform these calibrations (with a dSLR). Just look through your viewfinder and pivot your rig left and right. Manually focus somewhere between the targets and stop down the lens a bit (use depth of field preview to get a clearer view if necessary).

I did this tute in portrait orientation (since that's the only way I shoot pano's), and the # of shots and swing degrees markings on the slide reflect portrait only (in case you shoot with identical gear).

If you shoot your pano's in landscape orientation, your same front-to-back settings will work, but the number of shots will be very different from any you do in portrait.

As for why in the world would one need to calibrate a slide for lenses anyway when you're shooting objects at near infinity away... well, you wouldn't really need to. But out in the field you might want to include a foreground object. Remember, there's always grass, leaves, stones and stuff on the ground in front of you. If you have ever attempted to eliminate stitch ghosting in post-processing then you know why calibration is so important. Also, if you shoot indoor pano's (real estate stuff), then its an absolute must-do excercise.

wxwax wrote:

Nodals? I thought this was a family site? :patch

Sid.

01af

Nice tutorial! Except it won't find the lens' nodal point ... and neither the lens' optical center. It will find the lens' entry pupil (or, to be precise, the point where the optical axis and the entry pupil intersect).

-- Olaf

Andy

01af wrote:

Nice tutorial! Except it won't find the lens' nodal point ... and neither the lens' optical center. It will find the lens' entry pupil (or, to be precise, the point where the optical axis and the entry pupil intersect).

-- Olaf

There you go ruining things, with Optical Science

ziggy53

01af wrote:

Nice tutorial! Except it won't find the lens' nodal point ... and neither the lens' optical center. It will find the lens' entry pupil (or, to be precise, the point where the optical axis and the entry pupil intersect).

-- Olaf

Would you be so kind as to explain the difference between the nodal point and the "point where the optical axis and the entry pupil intersect"?

Is it still an appropriate method for calibrating a system for stitched panoramas?

If not, what is an appropriate method and why?

Thanks,

ziggy53

TOF guy

Thank you to Dave for writing the tutorial, and for Andy for posting the info .

Happy Hanukah to all!

Thierry

David_S85

OK, why don't we just call it the pivot point? And no matter what it's called, you just keep sliding the thing forward or backward to the place where nothing in the viewfinder/picture shifts when you're swinging the rig.

DavidTO

K, so I'm going through this process, and all i have to say is NO FAIR!!!!

My L bracket has no marking for the center of the camera. So's I'm guessing!

Nice tute, Dave!

David_S85

Thanks, David.

Easy enough to find the middle and mark it with a line, no?

DavidTO

David_S85 wrote:

Thanks, David.

Easy enough to find the middle and mark it with a line, no?

phuong

ziggy53, the "entrance pupil" is often incorrectly refererred to by panorama photographers as the "nodal point". it is not a serious mistake, though. just some optical science stuffs, as Andy pointed out. so you dont really have to worry about it.

i think you can look it up on wikipedia too.

David_S85

And whatever the heck one calls it, it is only important that things get adjusted enough that they line up.

Nikolai

Now, can somebody honestly tell me, what difference all this precision gadgetry makes on the outcome?

Most of my panos were shot handheld, and I have yet to see the particular problem. I'm not talking about sky banding or similar issues. My point is: the stitching s/w is pretty smart these days, and while perfect original shots would make its life easier, it seems to handle even some rather complicated cases just fine.

I do understand that this precision was vital in the days of yore, when the best you could hope for was to blend several shots together in the chemically stenched dark room. But today, with CS3, Panorama Factory and other nice stitching tools, how big is the real difference?

Can anybody who onws this gear showcase this?
Like, make a pano with the rail and dialed in pivot/modal point, and then make another one just handheld, and then run them both through the same tool?

In fact, I can run it for you, I just don't have to panning gear :cry

Just to be clear: I'm not saying it's not needed, I'm simply wondering about ROI and quality improvement factor..

Andy

Nikolai wrote:

Now, can somebody honestly tell me, what difference all this precision gadgetry makes on the outcome?

Sure. I can shoot any number of combinations with my gear, and when I bring the files into photoshop, they stitch together instantly, no waste, and no distortion. I have set nodal points for my 24 TS-E, my 16-35L (at 16, 21, and 35).

Nikolai

Andy wrote:

Sure. I can shoot any number of combinations with my gear, and when I bring the files into photoshop, they stitch together instantly, no waste, and no distortion. I have set nodal points for my 24 TS-E, my 16-35L (at 16, 21, and 35).

And can you make comparisons with the handheld with the same settings? That would be awesome!

DavidTO

Nikolai wrote:

And can you make comparisons with the handheld with the same settings? That would be awesome!

Nik, you're welcome to borrow mine if you want to run that test for yourself...

Nikolai

DavidTO wrote:

Nik, you're welcome to borrow mine if you want to run that test for yourself...

Oh, really? Sweet, I'd love to! Thanks, man!

pathfinder

Nick, my take is if you are only shooting objects at effective infinity, nodal points do not matter very much at all, as when you are hand holding.

But if your shot has objects in the forground that are much closer to the lens, then the nodal point rotation axis matters quite a bit more.

Nikolai

pathfinder wrote:

Nick, my take is if you are only shooting objects at effective infinity, nodal points do not matter very mugh at all, as when you are hand holding.

But if your shot has objects in the forground that are much closer to the lens, then the nodal point rotation axis matters quite a bit more.

Jim,

You're correct, most (if not all) of my panos shot at effective infinity or hyperfocal distance at least.
And I agree, with the primary subject being on the foreground this make total sense.
I guess I never had a chance to shoot a pano in such an environment

pathfinder

Nikolai wrote:

Jim,

You're correct, most (if not all) of my panos shot at effective infinity or hyperfocal distance at least.
And I agree, with the primary subject being on the foreground this make total sense.
I guess I never had a chance to shoot a pano in such an environment

Not sure just what kind of image with a large forground object would be a good subject for a pano offhands.

But I have watched Marc work with the 24T&S and I am resolved to pursue the skills required to use that lens more effectively.

David_S85

pathfinder wrote:

But if your shot has objects in the forground that are much closer to the lens, then the nodal point rotation axis matters quite a bit more.

A situation where an aligned head is just about mandatory would be indoor real estate VR's; as the proximity to walls, doors, lamps, etc. is so very near.

The pano in this post (from the pano thread) has a hand railing that is less than a foot from the end of the lens. The resulting stitched shot would be festooned with ghosting without a calibrated pano head.

Another important reason for a pano head is to preserve vertical space. Any decent pano head will include bulls-eye levels to assure the camera and head rotates around as level as possible. One only needs to shoot a handful of handheld panos to discover that post-stitching, both the horizon line is skewed and a good deal of picture area needs to be cropped out from the top and bottom. A calibrated pano head helps tremendously to solve both those issues, as well as reducing processing time. And accuracy is almost always assured.

There will still be some issues with ghosting; wind in trees, cars and people moving, but a pano head is just so much easier to deal with than fixing a myriad of problems later.

Nikolai

David_S85 wrote:

A situation where an aligned head is just about mandatory would be indoor real estate VR's; as the proximity to walls, doors, lamps, etc. is so very near.

The pano in this post (from the pano thread) has a hand railing that is less than a foot from the end of the lens. The resulting stitched shot would be festooned with ghosting without a calibrated pano head.

Another important reason for a pano head is to preserve vertical space. Any decent pano head will include bulls-eye levels to assure the camera and head rotates around as level as possible. One only needs to shoot a handful of handheld panos to discover that post-stitching, both the horizon line is skewed and a good deal of picture area needs to be cropped out from the top and bottom. A calibrated pano head helps tremendously to solve both those issues, as well as reducing processing time. And accuracy is almost always assured.

There will still be some issues with ghosting; wind in trees, cars and people moving, but a pano head is just so much easier to deal with than fixing a myriad of problems later.

David,
thank you, those are all valid points:-)
Now, how much did you say that stuff cost?

EDIT: nm, I went to RRS site... What can I say. Package (base + rail) $360 + dovetail $30 + s/h... So, total gonna be about $400..$420.
I guess I'll pass for now, I don't do this stuff often enough to justify it...

David_S85

It is pricey. No doubt about it. But I wanted their ball head and PCL clamp anyway, and an L-bracket for the body, so all I needed "extra" was one of their MPR-CLII slide dealies. QR on my bicycle wheels is fine, but I like the manual control of the clamp. That's how I convinced myself it wasn't that much more $. I was fooling myself of course, but the good sales job I did convinced myself that it was worth it.

Nik, you're a creative sort of guy. You could make a pano head on the cheap I bet. Plenty sites on the web for those. My first head was a home-made job and it got me by for a few years.

Nikolai

David_S85 wrote:

It is pricey. No doubt about it. But I wanted their ball head and PCL clamp anyway, and an L-bracket for the body, so all I needed "extra" was one of their MPR-CLII slide dealies. QR on my bicycle wheels is fine, but I like the manual control of the clamp. That's how I convinced myself it wasn't that much more $. I was fooling myself of course, but the good sales job I did convinced myself that it was worth it.

Nik, you're a creative sort of guy. You could make a pano head on the cheap I bet. Plenty sites on the web for those. My first head was a home-made job and it got me by for a few years.

You know, each time I'm stumbling across somthing like this I wonder if there is an easy/cheap way to get an access to the metal-processing devices (milling machines? I don't know the proper English terms... ). I know how to work most of them, and I'd make one probably in less than a day from a piece of scrap aluminum.

Nikolai

David_S85 wrote:

Nik, you're a creative sort of guy. You could make a pano head on the cheap I bet. Plenty sites on the web for those. My first head was a home-made job and it got me by for a few years.

Yep. Googled "DIY pano head" and came with several interesting threads.
Like this one: http://www.tawbaware.com/forum2/viewtopic.php?t=278

dbd

Free panhead on heavy zoom lenses at one focal length
Heavy zoom lenses that have a tripod adapter may provide a panorama head for one focal length.

When I was determining the nodal points on my Olympus 50-200mm zoom, I noticed that the offset from the tripod adapter mounting point went thru zero at one focal length within the zoom range. So, using the tripod adapter mounting point as a panoramic adapter works for single row panoramas at (for my lens) about 56mm. Using a normal pan head thus allows single row horizontal or vertical panoramas without additional hardware, at one focal length. Also, the tripod adapter allows the lens to be rotated to align either the native horizontal or vertical resolution of the camera with the 'short' side of the panorama.

Free panhead, only one number to remember: the focal length where the nodal point is at the tripod adapter mounting position.

On other zooms, YMMV.

Dale B. Dalrymple
http://dbdimages.com

Nikolai

dbd wrote:

Heavy zoom lenses that have a tripod adapter may provide a panorama head for one focal length.

Good point, Dale!

beautythief

Here's a question: How sensitive is the "Nodal Point" or "Entrance "Pupil" determination to the positions of the objects used to set the axis of rotation for the camera - the two objects on the table. Does the determination of the correct axis of rotation depend on the distance the two objects are placed from the lens? Has anyone checked this sensitivity to the positioning of the target objects? If so, what did you find?

David_S85

beautythief wrote: »

Here's a question: How sensitive is the "Nodal Point" or "Entrance "Pupil" determination to the positions of the objects used to set the axis of rotation for the camera - the two objects on the table. Does the determination of the correct axis of rotation depend on the distance the two objects are placed from the lens? Has anyone checked this sensitivity to the positioning of the target objects? If so, what did you find?

Dunno. I'm sure that some pano specific sites have used math for stuff like this, but all I can suggest is to set up your own test for each lens and focals you might be using in the field. Its the only sure fire way I know of of being accurate. And note, unless objects are near, it isn't too necessary to set up the markings or know the nodal slide numbers since things at infinity are way less able to mess up pano shot alignments in post processing.

Wow. Old thread. Still. Alive.