When taking photos meant to be stitched together into a panoramic picture, it is important, for the best result, that the photos are all taken from the same point. When shooting a panoramic view where most of things are at a large distance, this is not as important. A normal tripod will do, or even free hand shooting. For the optimal result though, the camera should ideally be rotated around a certain optical spot inside the objective. To make this possible, I decided to make my own home made panoramic head, an adapter to mount on a normal tripod, which allignes the camera so that the correct point is placed in the tripod's turning axle.
What happens then, when the camera is not aligned correctly? We then observe so called parallax errors, where objects at different distances are moving in respect to each other. This means that the pictures will be fundamentally different, and won't be able to stitch correctly no matter what. You can observe parallax errors yourself just by looking at two things at different distances from you, and turning your head in different directions. Close one of your eyes, and try to make the objects appear fixed in respect to each other, while turning your head! You will find that you have to move your neck to the right when turning the head to the left and vice versa. That way you hold your open eye fixed, just as the panohead holds the important optical point of the objective fixed.
Above I have only talked about "a certain optical spot", and not specified what point this is. (From a practical point of view, this is not important, since any callibrating methods will find the correct spot anyways.) Most sources either simply refer to this point as "the nodal point", or state that the nodal point (which is a technical optical term) is the correct point. However, there are two nodal points in an objective, and some say the rear one is the correct one. This fact can evidently be traced back to many photographic encyclopedias, but those sources refer to a special type of panoramic camera, with objective and film moving in respect to each other. When we are dealing with regular cameras, this is not applicable. Others state that the front nodal point is correct. Those in the know however, tell that the entrance pupil is what we are looking for.
The following series of pictures shows the making of my first panohead:
In the first picture we see the hardware I bought for this project. These are all very regular stuff that can be found everywhere, I imagine. (What caused the greatest trouble for me was the screw – here in Sweden the standard dimensions are, quite sensibly, in millimetres, while a 1/4' screw was needed for the camera mount.) Total cost: approx. 4€. Not bad compared to commercial panoheads, which may cost almost fifty times as much.
The second picture shows the parts in their modified forms. These slots will be used to slide the different parts in respect to each other and the tripod, to enable the panohead to adapt to different cameras.
That was the main work. In the third picture the panohead is assembled. The cloth was a nice touch I think. Finally in the fourth picture we see the panohead in use, mounted on a tripod, and with a camera in place. (An APS-camera, for demonstrating purposes only.) Note the scale, which will give some reference when fine tuning the head for a particular camera. The best value thus recieved will then be used to create my second panohead.
The main purpose of the first panohead was to be able to adapt to different cameras. It has one big flaw, and that is it can't correct parallax errors when tilting the camera up and down. That panohead was for one row panoramas only. The second panohead will correct this at the expense of portability.
As before, in the first picture we see the original part. In the second they have been modified. This panohead needed much less tweakening of the part – the only thing I've made is drilling two holes, and cutting off one of the disks. Then it was easily put together, as depicted in picture three and four.