[interested_observer]

Various stitching utilities will have different capabilities. PTgui (the pro version is the most capable which runs about $300) is probably the most sophisticated and capable package available. By using the point of view - the ability to move the point of view around, you can solve most of these problems. Also, by using a graphics card, you can pretty much get a real time response through the user interface.

[Ed Hurst]

It won't ruin it (and it completely possible) but it can complicate stitching, particularly if a wide angle is used (you mentioned 55mm which is much less a problem). There will be some distortion in the stitched image (beyond that of a series of leveled shots), but depending on the upward angle, lens, and subject matter, it may not be noticeable.

The main thing, if you want to be sure, is that all rotation (not just horizontal rotation across the scene but also the vertical rotation to point the camera upwards) takes place around the nodal point (i.e. the optical centre of the lens). For this, you need not just a panning pano head but a multi-row panning head, which allows you to carry out both types of rotation around the right point in the lens. This is the theoretical answer. In practice, you may find this to be unnecessary - modern stitching software (I use PTGui) has got very good at managing things when the camera movements are not carried out perfectly (the further the subject is from the lens, the less critical all this becomes, which tends to mean longer lenses, without getting in much foreground, work best - but of course, in your case, you may be forced to have relatively close objects in view). But the only way to be sure is to doing it properly - and that means performing all rotations (in any axis) around the nodal point of the lens.

Hope that helps,
Ed

[kaseki]

The main thing, if you want to be sure, is that all rotation (not just horizontal rotation across the scene but also the vertical rotation to point the camera upwards) takes place around the nodal point (i.e. the optical centre of the lens). For this, you need not just a panning pano head but a multi-row panning head, which allows you to carry out both types of rotation around the right point in the lens. This is the theoretical answer. In practice, you may find this to be unnecessary - modern stitching software (I use PTGui) has got very good at managing things when the camera movements are not carried out perfectly (the further the subject is from the lens, the less critical all this becomes, which tends to mean longer lenses, without getting in much foreground, work best - but of course, in your case, you may be forced to have relatively close objects in view). But the only way to be sure is to doing it properly - and that means performing all rotations (in any axis) around the nodal point of the lens.

Hope that helps,
Ed

I've had some bewilderment whenever I've read about this topic, given that lens assemblies have two nodal points and Pentax lens manuals don't report them. A little searching, however, reveals that the desired point of rotation is not one of the lens assembly nodal points, but the assembly entrance pupil. For a simple telescope, this is at the center of the front achromat (objective lens), but for the globs of glass we use (see lens diagrams linked near the top of the MF Forum topic page) the entrance pupil positions are deeper into the guts where the iris is. And for zoom lenses, the iris position can change.

So it seems that the widely disseminated technique of finding the entrance pupil experimentally by evaluating changes in perspective with rotation is what one has to perform, unless one has an xray machine to find the iris position. It would have been nice if Pentax had marked the position on the housing. But then, it has only been recently in optics history that digital processing suitable for efficient stitching has been available.

[Ed Hurst]

The technique you refer to for finding that point is actually very easy to do and only needs doing once for any fixed focal length lens. As you say, it does shift with zooms - so, for them, you would need to repeat the procedure for various different focal lengths.

[barondla]

It would have been nice if Pentax had marked the position on the housing. But then, it has only been recently in optics history that digital processing suitable for efficient stitching has been available.

It would be nice if the point was marked on the lens. Much like some cameras have the focal plane marked. Does any lens manufacturer mark the position on their lenses?

Thanks,
barondla

[cmohr]

I found this a while ago, Finding the nodal point in panoramic photography? is has some good info about it, and some good pics that explain the parallax issue from not rotating on the nodial point.

[TDvN57]

I found this a while ago, Finding the nodal point in panoramic photography? is has some good info about it, and some good pics that explain the parallax issue from not rotating on the nodial point.

Thank you for sharing the link. Very useful information.

[leekil]

The technique you refer to for finding that point is actually very easy to do and only needs doing once for any fixed focal length lens. As you say, it does shift with zooms - so, for them, you would need to repeat the procedure for various different focal lengths.

Is there no reference online where people have already determined this for various lenses?

[Ed Hurst]

Is there no reference online where people have already determined this for various lenses?

None that I am aware of... Interesting idea though!

[Focusrite]

+1 for a lot of the comments above. Prime lenses are better for consistency when setting up over multiple different shots. Trying to find the nodal point on a zoom can be annoying - I tried once and gave up after finding it wasn't consistent when I tried to zoom to the same focal length and expect the nodal point to be the same ( it wasn't ). Now I guesstimate with zooms, and usually I'm close enough to minimize dramas. Choose the focal length to suit your desired composition - just be aware that your depth of field and getting everything in focus becomes an increasingly uphill battle with longer lenses.

Some people use a bracket for panos.

If you reeally want to get serious with Panos, a good Indexed rotator is a great tool. You can set the degrees of index from a few presets, you just rotate it till it clicks into the next point.

This is a funny thing. I have taken panoramas hand-held that had no right whatsoever to stitch together perfectly: and they did! Saying that, I can have more dramas on a setup that is closer to ideal, so it can be ( frustratingly ) hit-and-miss. It also depends on other environmental factors, like how close you are to the foreground subjects of the photo. If you do shoot free-hand ( and why not? It's easy and free! ), try to rotate the camera around the nodal point as you shoot, rather than rotate your body and swing the camera around in an arc - it will help give the best chance of giving you a good stitch.

I have upgraded my gear over the years, and while I don't have a 645Z, I use a full 2-axis nodal mount for my K-1 and absolutely love it! One of the best purchases I have made, and not too expensive. It has made panoramas much easier to shoot with fewer errors, and in optimal conditions ( i.e, no wind, set up level, sturdy tripod ) I think it would handle the extra weight of the Z and a small lens with relative ease - I just don't think it's long enough. Even with my 24mm Samyang on the K-1 it's near the back end of its adjustment.

I think at the very least ( if you don't want to go too crazy with upgrades ) I would suggest you consider something like a longer Arca-Swiss rail. You could then at least slide the camera back so that the ball head's centre of rotation is central to the nodal point of the lens. It will help prevent parallax errors ( emphasis on help - it won't be a silver bullet ). In many of the lenses I use on the K-1 the nodal point turns out to be near the front element of the lens, and with the longer flange mount distance and larger lenses of the Z I think a longer rail would get you a lot closer to ideal.

I was going to mention Affinity for stitching, but it might be the same pano engine as in the Serif PanoramaPlus mentioned above as Serif is the publisher of Affinity. Either way, I find the Affinity stitcher finds control points that several other software programs seem to miss. Just this past week I stitched a 39 image (3 rows x 13 images) pano with Affinity.

I have used Affinity and Microsof ICE. Aside from the sometimes frustrating act of going around circles choosing your overlap points, Affinity is very good and the one that I choose out of the two. The perspective of the finished image looks ( to my eyes ) so much more natural than Microsoft ICE, and is usually a good representation of what I saw when I initially envisaged and shot the panorama. Saying that, Microsoft ICE is free download, and is still very, very good. When I first downloaded it, it would only take JPEGs. Last time I downloaded an update, it would take TIFFs, and now people are saying it can take RAW files now too? For free software, it's getting better and better. I would recommend it just so you can experiment a bit and make sure your setup is workable before commiting to further purchases or making your hike without field testing your technique. Just be aware, it is beneficial to be very deliberate when using ICE when taking multiple-row panoramas. Shooting the rows in a logical order and a set pattern will help the software place the panels correctly when stitching ( again, it may have gotten better at this by now ).

If the tripod and pano base are level can the camera be aimed up or does this ruin a pano?

It won't ruin it (and it completely possible) but it can complicate stitching, particularly if a wide angle is used (you mentioned 55mm which is much less a problem). There will be some distortion in the stitched image (beyond that of a series of leveled shots), but depending on the upward angle, lens, and subject matter, it may not be noticeable.

I agree that this shouldn't be an issue ( as I try to mentally picture what it would look like ), unless you are shooting multiple rows. You may then get some parallax between the top and bottom rows, as your setup will only take into account parallax on the horizontal axis. Saying that, natural subjects tend to be pretty forgiving for parallax, as the natural/repetitive patterns are good at hiding any errors. I also agree you will probably get some distortion. Shooting a wide angle panorama is bit like taking a spherical picture - it doesn't always translate perfectly when you flatten it out onto a surface. What I mean by that is how they translate/adapt the spherical globe of the earth onto a rectangular map - things have to get distorted and stretched to make it work without cutting it up. Again, a single or double-row pano shouldn't be too much of a drama, and many or most of the processing software should allow you to play around with perspective anyway.

I would again stongly suggest playing around before you go, if you can replicate the scenario somewhere more accessible before-hand, and see if it works. My first 3 or 4 Milky Way panoramas didn't work due to a mix of poor technique, ignorance, and not having the ideal gear, and it was so frustrating to come back from all of the driving and time spent out in the night to find the effort was all for naught ( except for the experience - that is always valuable ).

Is there no reference online where people have already determined this for various lenses?

I stumbled upon this some years back when I began looking into it. Hardly a comprehensive database, but they had the measurement for the 14mm Samyang lens that I thought would be handy, and I'm sure that there was another website somewhere that had some more measurements . . . . but in the end it seemed a bit too difficult, and I just manually worked out the nodal points of each lens by trial and error instead

Anyways, hope some of that was helpful, and welcome to the mad mad world of panoramic stitching - perfect for capturing things wider than what your camera + lens can do. Be warned, it can be an addictive rabbit-hole that will mercilessly lead you on an endless pursuit for more pixels and more detail. I absolutely love it when you finish a panorama and it works out well ( it's still a buzz for me ) - but I'm getting to the point where I'm choosing my battles. If my wider lenses can capture it, I'm more likely to be happier with just that ( + putting it through Topaz Gigapixel for more detail ) than battle with focus stacking, HDR processing on multiple panels, and dealing with any stitching errors for a photo that just may not be worth ever printing big. I once took a panorama with my 135mm lens of a field: 6 panels, with 6 shots per panel to focus stack. I started the post processing but soon gave it up when I realized that the final product would not be worth spending 2+ or so hours on to make it work, and that I had taken a wider shot of the field anyway!

Above all, have fun and good luck with the pano!

[stevebrot]

Last question...for now, what post processing program is the easiest?

If your computer is a Windows box, Microsoft ICE might be worth looking at as a free alternative.

Image Composite Editor - Microsoft Research

The big issue is positioning the camera so the "panning" motion does not introduce distortion while remaining true to the earth.


Steve

[TDvN57]

If your computer is a Windows box, Microsoft ICE might be worth looking at as a free alternative.

Image Composite Editor - Microsoft Research

The big issue is positioning the camera so the "panning" motion does not introduce distortion while remaining true to the earth.


Steve

I tried this Microsoft app a couple of weeks ago and found that the output files are greatly reduced in quality and size. Despite that it did not perform as well as ACR.

Not sure if you guys/gals have found the same but it seems there are multiple versions of stacking/panorama/hdr "engines" in Adobe. In bridge you can use Photomerge which seems to use the same interface as from within PS, in RAW Editor there are three options (1) HDR (2) Panorama and (3) HDR Panorama, plus then of course you have the options from within PS which is more elaborate.

It could be that these options all use the same underlying engine or not. In practice I found that if one method fails to do a merge or the outcome is bad, one of the others may succeed or do a better job. In a way that is a plus because you have three tools that can offer options to fix bad panning in the field. Within PS of course there is an option to correct geometric distortions which seems to fix the small sins of the field work.

[AggieDad]

Multiple row panos are not only possible, they are might be preferred. As I mentioned in an earlier post, I use longer lenses – 85mm and 100mm – to a) eliminate distortion, and b) provide greater detail. But the longer lens also minimizes (eliminates?) distortion which makes stitching multiple row panos easier and the results very pleasing.

The biggest problem for you with a multi-row pano might be stitching. Some software does not play well with multiple rows. But there is also the concern for final image size. You need to have enough RAM in your computer so that it doesn't "choke" while putting together a big pano. And you have to have some time on your hands as larger panos take some processing time. Relax and have a cup of coffee.

This pano was done during some testing and dramatic content was not a consideration. But it readily shows how a 100mm lens can give you a pano with incredible detail. The image in this instance is 75º wide (20 images) and about 21º (3 rows) in height. The finished image is 45,255 px x 12468 px which shows at only 10% of its size when full-screen on my 27" iMac.. To demonstrate the detail you can get (this is a K-3II with an old SMC Pentax-M 100mm f/4 lens), look at the little red rectangle in the pano and then the blow-up of the red box to show the detail captured.


A 3 Row Panorama (45,255 px x 12468 px) - Note the Red Box


Red Box from the Pano Above Showing Detail a Longer Lens Pano Allows

[SteveinSLC]

That's pretty good. You can read the writing on the satellite dish!

[barondla]

Multiple row panos are not only possible, they are might be preferred. As I mentioned in an earlier post, I use longer lenses – 85mm and 100mm – to a) eliminate distortion, and b) provide greater detail. But the longer lens also minimizes (eliminates?) distortion which makes stitching multiple row panos easier and the results very pleasing.

The biggest problem for you with a multi-row pano might be stitching. Some software does not play well with multiple rows. But there is also the concern for final image size. You need to have enough RAM in your computer so that it doesn't "choke" while putting together a big pano. And you have to have some time on your hands as larger panos take some processing time. Relax and have a cup of coffee.

This pano was done during some testing and dramatic content was not a consideration. But it readily shows how a 100mm lens can give you a pano with incredible detail. The image in this instance is 75º wide (20 images) and about 21º (3 rows) in height. The finished image is 45,255 px x 12468 px which shows at only 10% of its size when full-screen on my 27" iMac.. To demonstrate the detail you can get (this is a K-3II with an old SMC Pentax-M 100mm f/4 lens), look at the little red rectangle in the pano and then the blow-up of the red box to show the detail captured.


A 3 Row Panorama (45,255 px x 12468 px) - Note the Red Box


Red Box from the Pano Above Showing Detail a Longer Lens Pano Allows

Spectacular pano with amazing detail.

Thanks for sharing,
barondla