Re: Vertical error and leveled camera MISCONCEPTION!

[P3D John W Roberts <roberts@xxxxxxxxxxxxxxxxx>]

>Date: Tue, 12 Nov 1996 14:19:04 -0600
>From: "P3D Dr. George A. Themelis" <fj834@xxxxxxxxxxxxxxxxxxxxx>
>Subject: Vertical error and leveled camera MISCONCEPTION!

>BobH comments on the vertical misalignment error found in certain cameras
>and writes:
>>I think this is not a fixed error but is the result of not holding 
>>the camera LEVEL...

>Bob, with all due respect, I have to say that this is not true!  It is a
>common misconception that you will get an alignment error if you tilt the
>camera.  The fact is that there is nothing you can do in terms of holding 
>a camera that will give you a vertical or any alignment error.  

I think I can guess what might have caused the confusion. This was discussed
in an old issue of Stereo World, in an article describing stereo photographs
taken on the surface of the moon during the Apollo missions. Though there were
special-purpose 3D cameras on some of the missions, for general photography
the astronauts had 2D cameras. They were requested to take some stereo pairs
using the two-shot method - fortunately, there was no wind to cause the scene
to shift between shots.

The primary purpose of the stereo photos was for scientific analysis - they
didn't necessarily have to make viewable pairs, though many turned out that
way. Unfortunately, the astronauts had had to spend some of their training time
on subjects other than 3D photography, so some of the stereo pairs had problems
such as vertical misalignment.

One of the pairs printed in the Stereo World article compensated for vertical
misalignment by *rotating the images in the "mount"*. This caused the scene
to appear tilted, but corrected for the original misalignment.

So the confusion could be that rotating the images in mounting to compensate
for vertical misalignment in taking the photos was mixed up with rotating
a stereo camera (with the two images locked in the same relative position)
during the taking of the photo. You discussed the reverse to the procedure in
the Apollo article - rotating the images to vertical, and introducing vertical
misalignment.

>Even if you
>tilt the camera 90 degrees (by holding it vertically) you will still get a
>perfectly aligned pair!  
>...What will happen if you tilt the camera sideways?  The actual physical
>horizon will not be leveled but will appear tilted.  If there is a clear
>pronounced horizon in the picture then this is bothersome.  

I've taken those at 90 degrees rotation. Very interesting - there's a strong
visual cue that the world has suddenly turned vertical rather than horizontal!

John R


------------------------------