Well, I went off and followed my own procedure. Here is what I
discovered with my *istD:
* The viewfinder is not aligned with the camera base. It tilts to the
right, so it has a significant negative angle.
* I took the picture, but it was not very useful, since the viewfinder
was already mis-aligned. So I would alter the procedure so that once
you see how well the viewfinder is aligned, then proceed to align it
visually as perfectly as you can with the reference grid, then take the
picture. What this does is give you a reference with respect to the
viewfinder that is visible on the screen.
* I discovered that not only is the the viewfinder not aligned, but the
sensor is mis-aligned with respect to the viewfinder, which is bad news,
because you expect at least to get WYSIWYG functionality. But luckily
it was not a whole lot, possibly a degree or so.
* I would also add to the procedure that you need to get the flattest
lens you have, preferably a higher FL macro, like the FA 100mm 2.8
macro. I started out with a different lens, and encountered sufficient
distortion to challenge the results.
I'm very disappointed with the viewfinder/base misalignment. This means
that you have to level the camera by looking through the viewfinder at
some reference (and ignore the viewfinder/sensor mis-alignment for now),
and not depend on the tripod being level. This is unacceptable, one
should not have to take a level with them to make their camera level, it
should be enough to get a rough leveling using the levels built into the
tripod head. The nicer tripods have two levels orthogonal to each
other, which should be sufficient to get a pretty decent leveling. This
defect is a real pain and should not happen in a camera like the D,
which is currently the DSLR standard bearer from Pentax. I'm wondering
whether this is happening pretty much on all *D cameras, or I have a
defective one. If you look at the design of the *D, the top, which is
what the viewfinder is aligned with, is definitely not parallel with the
base, which leads me to believe that all *D cameras have this problem.
rg
Gonz wrote:
You have to test for both a tilted sensor and a tilted viewfinder.
first find out if your viewfinder is tilted. You have to use a
reference. Since the ultimate reference is reality, and you have to be
able to trust your tripod and the relationship of the tripod to the
camera, I would suggest the following procedure:
1. set up a vertical and horizontal grid on a wall, a large piece of
graph paper or graphical velum would work.
2. make sure that the vertical and horizontal grid is level using a
large accurate level.
3. attach the camera to the tripod which has been leveled also as
accurately as possible in orthogonal orientations. We are assuming that
as far as the camera is concerned, the base is the reference, since we
cannot adjust that, so we our making our camera reference agree with the
world reference (the grid)
4. point the camera at the grid and look through the viewfinder. If the
lines/edges in the viewfinder do not agree with the grid, then the
viewfinder is tilted.
5. take a picture and look at the result. if the picture is misaligned
with the grid, then the sensor is tilted.
6. if the viewfinder and the sensor agree with each other but not the
grid, then you may be able to compensate by the amount of the tilt with
some type of wedge on the base, but this sounds like a pain.
rg
Paul Stenquist wrote:
I guess I shouldn't be surprised. How would one test for a tilted sensor?
Paul
On Jan 29, 2006, at 9:07 AM, Bob Shell wrote:
On Jan 29, 2006, at 8:59 AM, Paul Stenquist wrote:
I haven't seen any evidence of misalignment on either of my D
cameras. What makes you think your sensor is misaligned? I would
think that an accurate test is very difficult to perform in that it
requires precise alignment of a camera and target.
Actually tilted sensors has been a very common problem in digital
cameras, all the way from the cheapest point and shoot up to the most
expensive DSLRs.
Bob
