ChrisR wrote:If they put the iris in at the back as an afterthought, so it's not in the "right place", then heaven knows - you'll have to wait for Rik! I'd guess the PMR may change when the lens is wide open because the limiting aperture moves.
Correct, the PMR would change in that case, because the limiting aperture would move and thus would be treated differently by the optics (if any) on each side of the lens.
As an extreme example of this situation, consider a lens that has been modified to be telecentric, by adding a limiting aperture at the rear focus distance. In this case the lens with telecentric aperture has PMR=0 because the apparent rear diameter = physical diameter while apparent front diameter is infinite. But if you open that added aperture far enough that it does not limit the ray cones, then the lens reverts to its normal non-telecentric behavior with two finite diameter pupils.
In the last set of pictures showing the triangular aperture, I would be happier if the aperture were focused sharply instead of being fuzzy. But the computed value of 0.731 is close enough to that of the Oly 20 mm to support the claim that the new lens is a clone of that.
mjkzz wrote:...demonstration of "doughnut" shaped bokeh
Interesting -- thanks for this. Apparently a significant amount of remaining spherical aberration when used wide open.
Oh boy, three blade aperture . . . but as stacking progresses, it is getting smaller and smaller, kind of weird but final stacked image looks fine, so for single shot image, like stack #63, it is probably "dreadful" :-) But for final stacked image, it is not a problem at all
Correct. The shape of the aperture has little effect on the appearance of in-focus detail, because that appearance is determined by diffraction and lens aberrations and not by geometric blur due to spreading of the ray cones. It is the same reason that you can use a phase objective (phase ring = annular aperture) and still get good stacked results within the in-focus slab.
--Rik