Hi Joost,
I would like to add something related to John's comment to the wish
list. I am one of what I'm sure is a very tiny minority of digital
panographers who sometimes use a rotating slit-scan camera. When the
lens axis points at the horizon, that generates a cylindrical pano, or
a roughly eqirectangular one with a fisheye lens, that you can load
into PTGui and correct, reformat, etc. But if you tilt the lens up
or down you get a panorama that PTGui (or any other PT-derived tool)
cannot process correctly because the center of projection no longer
coincides with the horizon, but follows some other parallel on the
panosphere. This situation requires one additional lens parameter,
call it the altitude angle, and of course code that can deal with the
conical projection in the quasi-cylindrical (rectilinear lens) case.
The quasi-equirectangular (fisheye) case is simpler. With a perfect
equal-angle fisheye, or after correcting lens errors, the lens
inclination is just a constant in the Y coordinate of an
equirectangular image, or to put it another way, the horizon is not in
the center of the image. But the only way you can get a PT-family
tool to handle that image is by padding it until the horizon is in the
middle (that is so for PanoTools and Hugin, I'm only guessing that
PTGui is the same).
So my wish is not at all difficult to implement: please make it
possible to designate the position of the horizon in an input
equirectangular image. I don't much care if you display it cropped or
not, just line it up right.
Regards, Tom