This is the final of a four-part series discussing the technical side of photography.
I enjoy the technique of taking panoramas, as I like the way I can capture the scale of sweeping landscapes with dramatically increased resolution. I also enjoy getting more complex with panos, mixing it up a bit by stitching multiple long exposure, HDR or art filtered shots for unique effects. It increases the number of steps, but yields results that are not often commonly seen.
The software I use to stitch panos is called Hugin — a free software for photo stitching that, once one gets passed the confusing interface, does an admirable job. Olympus Viewer 3 will stitch panos, although the Panorama Scene Mode must have been used in the camera when taking the pictures for it to do so, and it is has some limits in application (see below) that have steered me away from it.
While I like taking panos, I don’t generally like displaying them. On a computer or mobile device, the limited screen space makes it easy for the viewer to lose sight of the details, while at the same time zooming in loses the scale. I have also found that it’s easy for me to get carried away when taking panos, resulting in ridiculously long chains that only make them more awkward to digitally display and oftentimes to the detriment of my composition. To put these points in another way, too-large panos can sometimes lack a point of focus. Its a possibility brought out by both the odd aspect ratio and the thought process that goes into to choosing to do a pano in the first place. The takeaway is not to do a panorama just for the sake of doing one. Visualize the shot, then use the technique just enough to make it reality.
One way to overcome the problem of sizing a panorama on a digital monitor (besides printing) is to shoot both vertical and horizontal panoramas. A simple way to do this is to frame in portrait orientation. Shooting panoramas in portrait orientation makes for a more proportional, taller image, which also provides more leeway above and below. When processing panoramas, the upper and lower edges of the image will be generally uneven after the software corrects for distortion, and the top and bottom will have to be cropped. It’s important to allow as much leeway as possible so that important elements in the scene don’t have to be cropped out. After learning this lesson, I now always shoot panos using portrait orientation.A much more demanding method is to stitch above and/or below the frames, rather than just on the sides. Laying out frames both horizontally and vertically (like quadrants in a grid) can make the shape of the overall panorama much more monitor-friendly, and the end-result is a super high-resolution image. Of course, the number of alignments that could go wrong dramatically increases and special panorama gear is also required. Also, the sides now may also come out uneven, so extra space on the sides must also be left for cropping. I’ve never tried this method, as I don’t own the proper gear.
Olympus only allows for landscape orientation as well as only horizontal stitching with their Panorama Scene Mode interface with Olympus Viewer 3, which is the main reason why I don’t use it, though they ought to be able fix the issue with a simple firmware update. Since I use Hugin instead (and I suppose any other stitching software will do as well), there is no reason for me to use the camera’s Panorama Scene Mode at all.
I went with RRS after reading from reviewers that it’s products are among the top in the business, and I liked that they were made in the USA. After discussing my needs by phone with the helpful staff, I chose a TVC-24 Tripod and BH-40 ball head — a medium weight option optimized for traveling. While one can go smaller and lighter than the Two-Series for µ4/3 gear, I also needed my tripod to handle digiscoping with a spotting scope, and in any case a slightly heavier tripod would be more versatile in accepting heavier kits, should I ever want one. I chose a four-leg tripod instead of three because of airplane carry-on restrictions. Indeed, this has saved me multiple times from having to check in my tripod.I’m not going to insist that RRS is the only high-end tripod to get, or even that photographers need to spend that much to get a tripod that will suit their needs. The only other tripods I’ve ever used were of the plastic $50 variety (the ones I said were next to useless, and I didn’t pay money for them), so I’m in no position to compare RRS to other brands. Thus while I am sure there are less expensive options that will do just fine, I don’t know enough about them to recommend any. What I will say is that I do know that most photographers loathe their tripods and dread carrying them around. So probably the best endorsement I can give for RRS is that I love my tripod and actually want to carry it with me. Such lightweight and durable, precision-machined parts and well-designed gear for amazing photographic capabilities is simply a pure joy to use. Plus, RRS has such an amazing product list that anything I could possibly need, they make.
Finding and Keeping the No-Parallax Point:One of the more complicated factors with panoramas is the parallax point. Parallax is the slight variation in angle of view caused by the movement of the viewer. It can be easily demonstrated by closing one eye and then switching to the other. Notice how what you see shifts depending on which eye is open. When the camera angle is turned to take another frame for a panorama, a similar shift in perspective can occur, so that when two frames are stitched together, parts of the image may not align properly.
The way to the correct this is to ensure the no-parallax point is occupying the exact same spot for each image, and the way to do that is to make sure that it is directly over the axis of the tripod mount. I use a simple rail (with measurement markings on the side) to slide the camera back a bit. The exact location of the no-parallax point can be determined with a simple home test. I learned the concept from the Really Right Stuff website, and while I followed the instructions the best I could, I’m not sure that my my entire process went exactly like theirs. Just as important as finding the exact point is keeping it. Your gear needs to be marked in a way that you can easily recreate the exact position in the field. How you do this will vary depending on the type of gear you are using and how it is marked. I chose a method that I could easily recreate and remember the positioning of the gear. It might not have been exactly what the tutorial at RRS was instructing, but it made sense to me, and the important part is the consistency.
All of my mounting equipment is from Really Right Stuff, which has an entire line of gear specifically for shooting panoramas. My current setup is of the simple and flexible variety — the 192 Precision Plus Package and BOEM1-L camera L-plate set designed for the OM-D E-M1. RRS makes camera plates specifically designed for many of µ4/3 cameras, and they fit like a glove. Aside from the precision quality of the parts, RRS etches measurement markings on all their gear to assist in exercises just like this. The 192 Precision Plus Package consists of a rail and two clamps that can be attached back-to-back either at either the same or 90-degree orientations. This makes it usable for both camera plates and lens feet. I don’t only use it for panoramas. By using it to position heavier lenses (as well as my spotting scope) back a bit from the tripod, it also serves to balance them for ease of use on my BH-40 ball head.
The test setup is simple. Align the camera tripod with two upright objects in a well-lit area so that the camera is looking straight at the two objects, and the rear one is immediately behind and completely obscured by the other. The thinner the two upright objects are, the easier it is to see them for the purposes of the test, the more precise your results will be, and the higher focal lengths you can test. They might even be easier to see if they are two different bright colors. I didn’t have any particular distances between the objects or from the tripod. RRS only says that one should be near and the other far, but it seemed to me that they had to both be close enough to the camera that they could be seen clearly through wide angle lenses.
- The camera has to be completely level on the tripod.
- For the sake of simplicity and consistency, I chose to slide my clamps all the way back on the rail. Throughout the test, the camera stayed in that position, and I only moved the rail forwards or backwards on the tripod mount.
- RRS suggests lining up the camera’s sensor plane over the center of the tripod as a starting point, and moving from there. But don’t worry about getting this precisely; just eyeballing it will be enough. You’ll be moving away from that point soon enough.
- RRS then says to pan the camera to the left. If the rear object shifts to the left, move the rail forward a bit. If it shifts to the right, move the rail back a bit. Repeat until the rear object doesn’t shift at all, and then you have found the no-parallax point. For my native µ4/3 lenses (which is all I own), the no-parallax points were all located well in front of the sensor plane. So, I invariably just had to slide the rail back incrementally until I found it. I can’t say that this will always be the case with every native µ4/3 lens, but it seems likely.
- I recorded the measurement on the rail where it met the center marker on the clamp. However, because my rail’s numbers started from the front of the tripod instead of the back, where I started my camera (its just the way I preferred to set it up so that I could see the leveling bubble), the number I recorded was not the distance that I moved the camera. In this way I think I may have deviated from RRS’s tutorial. This, however, is of no consequence, as long as I can recreate the exact position in the field.
- For zoom lenses, the no-parallax point will differ at different focal lengths, and my go-to landscape lens is my 12-40mm f/2.8 Pro lens. I found the no parallax point at 12mm focal length to be when the rail was positioned at the 45mm mark, with the rail bubble facing forward (where the camera would not cover it) and the camera slid all the way to the back of the rail. At 25mm, the no-parallax point was at the 10mm position.
- For shooting horizontal panos in portrait orientation, simply turn the camera to portrait orientation on the tripod. Just like with landscape orientation, Really Right Stuff has a marking on the plate to make sure the camera is lined up at the center when in portrait orientation. Note that this will also provide you with the no-parallax measurements for shooting vertically-stitched panoramas, but actually getting perfectly level multi-row panos requires pano gear that I don’t own. (Edit: the more I think about it, the more I think it may be possible with my rail and ball head combo, but I haven’t tried it yet.) While the RRS version of this gear (PGE-2) is expensive, it serves as both a gimbal head and a full panorama kit, so it may become worthwhile for me to buy it if I ever get a large-enough super-telephoto lens for birds and wildlife. (Edit 2: I just broke down and ordered one for my birthday.)
Once understood, overcoming parallax in a pinch is possible without complex gear. I have even gotten by shooting panos hand-held, stepping back slightly when I do my re-framing in a way that my body moves with an imaginary pivot point slightly in front of the camera, so that the no-parallax point (hopefully equating to my imaginary pivot point) stays stationary as best as I can manage it. This can work well enough if the scene does not have any foreground objects crossing stitching points (the farther away they are, the better), or straight lines running diagonally across stitches from the foreground to the background. That said, I am still just guesstimating the proper position and while it may help, probably won’t be perfect. Also, hand-holding panos is not ideal, and means that I have to crop more than normal, so using portrait orientation to get a little more vertical leeway is even more important. One of the panos in this post did not line up perfectly, and I used some crude tricks in my post processing software to hide the misalignments.Movement can also be difficult or even impossible to capture in panos if the movement crosses a stitching point. This is where using a rectilinear ultra-wide angle (UWA) lens can be a better option, either instead of a panorama, or for taking panoramas with less stitching required. UWAs also have a greater ability to do astrophotography with tack-sharp stars. A fast aperture and wide focal length is needed to get exposures short enough to both keep noise down and get the shot before the stars move (otherwise making star trails, for which an UWA is not needed). Another advantage to rectilinear UWAs is there ability to shoot wide in confined spaces, such as indoors. Other than in travel photography, however, I don’t tend to shoot indoors. Also on the downside, UWAs tend to be soft in the corners, and almost all of them don’t accept filters.
If the movement in question is water, such as waterfalls or ocean waves, long exposure photography can overcome this issue. Smooth out the water long enough, and any minor imperfections will be blurred away. This is much easier to do in rivers. In rivers, the waves are caused by stationary rocks beneath the surface, so they never shift location. One can get by with relatively short exposures to just smooth out the water a little. All of the long exposure panoramas of Great Falls, including the winter scene at the top of this post and the two featured in The Technical Side of Art: Part 3 – Waterfalls and Fireworks, are examples of this technique. Ocean waves, however, come in pulses and are never exactly the same twice, so much longer exposures are needed until the water is so blurred that it appears like fog.