In article <58prmh$rt7@cello.hpl.hp.com>, jacobson@cello.hpl.hp.com says...
>In article <58pa5k$4ah@newsstand.cit.cornell.edu>,
>Bob Neuman
>>With the F-series Nikons, a line placed parallel with the edge of the
>>frame (that is rendered straight by the lens) will appear straight and
>>parallel in an F finder (and if it doesn't, it is possible for the user
>>to adjust the frame placement and rotation in the F-series bodies
>>[though it is not easy to get it right...]). (With apologies to non-US,
>>non-"N"-body-designating folks...;-)
>Even if there was pincushion distortion in the viewfinder, if the edge
>of the viewfinder frame was straight, then a straight line on the film
>would appear curved but parallel to it, if you know what I mean by
>that slightly ill-defined phrase.
Yes, but this is not the case - a truly straight image is rendered
pincusion, but the frame edge is not similarly pincushioned (it must
be bowed intentionally to look straight in the finder). When I check
VF masks for rotation in cameras I buy, it is possible to place the
curved (distorted) line fairly well parallel to the mask edge for
the film check, but in picture taking, when a straight line is not symmetrically placed in the frame, but I want parallelism with the
edge of the frame, the curvature makes alignment difficult, even
with a grid screen installed.
>Now I've *heard* that not all
>viewfinders actually have rectangular frames, rather that some are
>slightly barrelled to make it *look* straight in the viewfinder.
>Does anyone know for sure if this rumor is true, and if so, which
>cameras actually have barrelled viewfinder frames.
Yes, all the ones that I have seen that have pincushion distortion,
which is most current cameras...
>One data point: a Nikon 8008 is not barrelled. I took a ruler with
>nice grid lines and held it over the lens mount (no lens on the
>camera) and caused a distant street light at night to cast shadows of
>the grid lines onto the focusing screen. I could arrange the ruler so
>one of the lines was right next to the frame edge, and the distances
>between the line and the edge appeared constant all the way across.
Umm, I would suggest trying that again (make sure that the ruler is
flat...;-) I just set up a small light in the darkroom, stepped back
about 6', placed an opaque flat object with a straight edge over the
lens hole, drew it down until a slit of light was revealed "parallel"
with the frame edge, and the familiar non-parallelism/non-coincidence
was again revealed (as it is with a non-distorting lens in place [like
any Nikkor 85mm, for instance] when a straight subject line [horizon,
building edge, etc.] is brought near the frame edge) - the frame edge
and the object edge did not have the same "curvature" (the frame edge
appears straight, while the object edge appears curved, which is the
opposite of reality...;-). When the opaque object was drawn more
across the frame, its distortion coincided with that of the lines of
the installed grid screen, and repeated the (geometrically reversed)
effect at the other side of the frame. I must conclude that the VF
has pincushion distortion, and the frame edge is barreled to make
it look straight.... (though I have heard that the mask itself
is not barreled, but may be by the optical system above it -
a happy coincidence for the camera maker...?;-)
Hope This Helps