I am thinking of doing some panoramas by stitching 2 or more frames together. Any helpful hints? I have a bogen 3047 head and a cambo sc. the focal length is 150 all on 4x5 film.I was planning on just connecting the negs and making contact prints. Any info is helpful? Josh

-- josh (devil_music@usa.net), May 31, 2000

Answers

Josh: That is a technique that has been used for many years. It works best if you can get the lens directly over the pivot point of the tripod head. You may need to rig a mount board for the camera if the tripod socket does not extend far enough forward. It might not be too difficult with a monorail view camera. I think Linhof Techs have a tripod socket on the drop bed. In any case, get the lens as near the center of the pivot as possible. You will also need to get the tripod head perfectly level. I would set it up with a small level and make sure it is level side to side and front to back. Hope this helps, Doug.

-- Doug Paramore (dougmary@alanet.com), May 31, 2000.

If you have front and rear shift you don't have to pan the camera.

-- Sean yates (yatescats@yahoo.com), May 31, 2000.

Sean, I respect your opinion much, but serious parallax problems may arise if you shift the front standard. You are free to shift the rear standard only. Even then, any transformations done by the regular stitching programs have to account for the shift in "lens- center".

-- Carlos Co (co@che.udel.edu), May 31, 2000.

My biggest problem has always been the light falloff at the edges of each frame, which causes an unnatural, conspicuous light "line" at each seam. Levelling the head is of paramount importance, and isn't helped much by having a Bogen 3047, the head I also use. A head with a single bubble level, as opposed to the two axis configuration the 3047 has would be better. One of those panorama attachments (who makes the one recall reading about?) would help as well.

I've done a few of these, and my alignment always seems right on. The method I use is to find the center of my scene, so I have a basic reference as to what I want to have at the far left and what I want at the far right. Once I've gotten as close to level across the horizon (for all frames), assuming you have one, I try to determine where my individual frames will come together. Believe me, this is much wiser to do in advance than it is to try to find a good branch, blade of grass or rock to piece together on the fly; the more frames you have, the more complicated this will be.

As you take each photo (in whichever direction you prefer) you'll have to relevel and recompose, otherwise the horizon will make pretty wild moves across your panorama. This is where those seam-sealing things you were groping for come in: use them as reference points for your frame joinery and to note where on the ground glass (precisely) the horizon is, so that when you swing your camera, you'll know where the horizon line and seam meet.

As long as you are consistent across each frame, you should have minimal trimming to achieve perfect alignment of the frames. The largest I've ever done was a three panel, 8x10 using a 300mm lens. Each sheet trimmed down 7.5x9.5, not bad. Like I said, the only "problem" I have is the light line at each seam, but I kinda like the effect.

Sorry for the lack of brevity.

-- Chad Jarvis (cjarvis@nas.edu), May 31, 2000.

paralax? Isn't that when your wife thinks you're going 80 m.p.h. when from your perspective it's apparent you're doing 65 m.p.h.?

-- Sean yates (yatescats@yahoo.com), May 31, 2000.

One more suggestion: Don't try to mount the panorama segments up against each other in an attempt to make it look like a "real" panoramic photograph. I've found that when you mount them with a bit of space between (quarter inch or less), so that it's obvious you've got segments rather than a single photograph, it actually looks better because the brain and eye do not seem to concentrate on the small, inevitable discontinuities between the segments.

-- Chris Patti (cmpatti@aol.com), May 31, 2000.

Regarding the prints, I can tell you how the U.S. military trained us to put aerial photographs together. Overlap your views about 30% when taking your shots. Don't mess with the resulting negatives. Let's say that you end up with five equal-sized prints. Number them left to right on the back of each print. Then lay out the center print ,# three and it's neighbor, number #2 on a light box. The center photo serves as the reference point for your Photo Mosaic, as the military and probably geologists called it. Overlap the left photo onto the center print and find that point, roughly 1/3 from the edge of each photo, where the images are in register. Mark this point top and bottom on the left-of-center print using a soft pencil. Now take a razor blade and cut the print. Turn it over so the image is down and the paper side is up. Hold the razor blade almost parallel to the paper at the newly cut edge at the top of the print and slice a section of paper to feather the edge. Carefully tear the paper from top to bottom. Apply rubber cement to the feathered edge and the back of the print that will contact the center print. Place it in register and roll to make good contact. Continue until the mosaic is complete. The print edges will not line up perfectly. You can either lay a long straight edge over the mosaic and trim or keep the constructed look. This is easier to say than to do. Another easier option would be to scan the film, Photo CD or otherwise, and let a computer program do the stitching for you. Good luck.

-- Steve Singleton (singleton1@bigfoot.com), May 31, 2000.

Just to save a lot of trouble.... first off, centering the nodal point over the pivot axis is only important if there is close subjects... there is no good formula I have seen to define close, its dependent of fl and distance. Second, don't try to stitch a bunch of 4x5's together in software on a Windows Machine, it does not work... Windows has a ceiling on file size, and when its reached it crahses the machine. I am not sure what the Macs limitations are. I got so frustrated with this, I bought a rotational camera instead. Good software stitching for large fine art prints is a few years away, but for web and smaller prints, its amazing how good the stuff is... do a search for Panoguide, this guys website is the best for Pan stitching. Also, Jaspar engineering has some pix on his web site of what you want to do... also he sells an inexpensive pan head to easily accomplish this. BTW, my 3047 has two levels on it? But after careful checking, I have learned one was way off from the factory... good luck....

-- Bill Glickman (bglick@pclv.com), June 01, 2000.

With all due respect to Bill, your "Windows" machine must be poorly configured. I mosaic air photo and LAndSat imagery daily using WindowsNT and file sizes that range into the 10's of GB!. While Photoshop can do this with moderate sized files (100's MB), it doesn't really have dedicated tools for mosaicing. There are standard algorithms available to remove the "light falloff" by regression fitting cos^4 falloff masks. The upshot is, depending on what you want and how much processing you want to do, virtually perfect mosaics can be obtained.

As others have pointed out, the best thing you can do at exposure time is to try to get the first nodal point centered on the rotation axis. In practice, if you don't have foreground closer than a few feet, and given the relatively normal focal length of your lens, centering the shutter body over the rotation point will do fine.

-- Glenn C. Kroeger (gkroeger@trinity.edu), June 02, 2000.

The information of Windows 98 limitations was provided to me by one of the largest stitching software companies today... Enroute imaging. I sent them 8 LF files to stitch, and at first they felt their software had a bug, because half way through the process, the PC would crash. Further investigation lead to the problem, it was Win 98 file size limitations, not in actual file size, but I beleive it was temp. files or something similar that Win 98 uses. Enroute imaging informed me, until Win 98 is upgraded, large files can not be stitched on Win 98. I had the same problem using stitchers from other vendors. I can not comment on NT, I have never tried it, either has Enroute. If anyone would like to confirm this information, you can email David Rowe at Enroute Imaging, and he will gladly explain the limitation of Windows 98 for stitching. Drowe@enroute.com

I am very careful about posting information.... I try my hardest to make sure what I write, accurately reflects the information I am trying to share. I apologize if I fell short this time.

-- Bill Glickman (bglick@pclv.com), June 04, 2000.