Lower rail setting?
I'm not sure if I'm getting how to find the LRS. Some people say that looking through the viewfinder and aligning with the centre mark is the way to go. Where would these stitching errors would occur if not set correctly ? I know that the upper rail setting sorts the adjoining images (left to right)....is the LRS to align the nadir with the images above it. Anyone know of a test for the LRS setting please.
That is not an accurate way of setting the LRS, but it's a quick and useful way of establishing a good starting point for one of the other methods. Many people would regard it as "good enough". I'll suggest three other methods (in no particular order of preference):
Originally Posted by Johnytuono
1. Direct visual adjustment of the entrance pupil position. Parallax arises from the movement of the entrance pupil when the camera is rotated. You can view the entrance pupil (the bright spot in the middle of the lens) through an improvised sight such that you can readily detect any change in its position as you rotate the panohead like this:
You want no change in the horizontal and vertical position of the entrance pupil between the starting and final settings of the head shown. (Shine a small flashlight into the DSLR viewfinder eyepiece to illuminate the entrance pupil, or position the camera in front of a bright window).
2. The "sawtooth" method. Tilt the camera down just sufficient to include the the nadir point and shoot n shots around as you would for a 360 degree panorama. The camera should be focused on the floor where there needs to be suitable features for easy control point assignment. Stitch the images with PTGui with control points only on the floor. Check for an irregular sawtooth pattern in the edge of a circular knob or rotator at the base of the panohead. See Smooth's tutorial:
3. Take two downward pointing shots with the head panned through 180 degrees between them. The camera should be focused on the ground. Stitch the images in PTGui with control points only on the ground and output a layered PSD file. View the file in Photoshop with the opacity of the top layer set to 50%. You will see something like this:
Any parallax resulting from a change in position of the entrance pupil will cause a misalignment of the centre screw. If there are scale markings on the lower rail, you can use the ruler tool to estimate the adjustment of the LRS needed to correct the misalignment (half the measured lateral shift). Note that vertical shift in the above image might be corrected by twisting the camera slightly on its mount to the upper rail.
Thank you for that informed reply John.
When I did this test I came to 156mm (just like Heinz), but noticed that the image didn't align, as if the camera wasn't centred on the camera plate (I'm using U3), like it wanted moving to the right slightly. Would that be critical, and if so how do I move it as it has a plastic block on that's supposed to auto-align the camera to the clamp?
The plastic automatic align plate on U3 centers the camera in the middle of the clamp. So in the vertical on upper rail. You can not change this system.
I always test my LRS with 2. Smooth's Sawtooth method. The result will correct slightly misplacements of the sensor during production as well as lens problems because the pics are worked out by PTGui, which corrects these problems. Same as John's 3.
Believe me, the sawtooth method is correct. I have tested it with several bodies.
Nothing should be wrong with your camera plate U3.
At least PTGui will correct a lot.