I agree. It's only for flare that the reflectance doesn't matter. If maximizing the illumination is important, then 50% is right.ChrisR wrote:for axial illumination of a subject you DO want 50% transmission / reflection to maximise the amount of light getting from the subject into your lens.
That said, it's probably good to keep the magnitudes in mind. If uncoated glass is 10% reflective, then the net is 9% (0.10*0.9). Swap in a 50/50 beamsplitter, and it rises to 25% (0.5*0.5). In the grand scheme of things you have to balance the 2.8X gain in efficiency against the cost & time of acquiring a suitable 50/50 material. As I mentioned, Kodak's book used a coverslip.
That sounds right with respect to aberrations. The only wrinkle is that then all the light that illuminates the subject also has to go through the aperture of the lens. So instead of getting the dreaded "black hole" effect, I think we get a "white hole" instead. But I haven't tried it either.Doing things on the cheap though, I think the better place to put a semi silvered mirror, in a simple finite system such as when using a reversed enlarger lens, is between the lens and the camera. The "NA" that side is a lot smaller, so I'm guessing the aberrations introduced would also be much less.
--Rik