ChrisR wrote:I'm not happy that I understand what's going on here.
I thought the focal length of the reflector would be significant - presumed that was why Oly have two different reflectors. But you're getting much the same results with a white diffuser (that's like a paper cone in place of the concave mirror, right??) which suggests to me it's basically just scattering the light.
Oly's drawing is a bit rough but suggests to me that the reflected rays would come to a focus (bounced off the subject) at the front of the lens somewhere - could it even be the entrance pupil? Would that remove the "black hole"?
But Oly's reflectors are "semi-matte" anyway
Adding some detail to Andrew's answer...
First, about the black hole. Nothing will remove the black hole except light that comes from the lens itself, or appears to as seen by the subject. That requires a partially reflective mirror someplace in the system. No matter how cleverly you shine light around the lens, a mirrored surface in front of the lens will stubbornly persist in reflecting the lens back to itself. If there's no light coming
from that direction, then there's no light going back
to it either, hence the black hole.
Now, about the reflector.
Imagine an idealized system in which the illumination comes from behind the subject in perfectly parallel rays and bounces off a perfectly shiny and perfectly parabolic reflector. The illumination pattern at the focus of the reflector would then be a tiny point, no bigger than the Airy disk of the diffraction pattern implied by the f-number of the reflector. This would not be very useful! So what to do, what to do? One option is to simply defocus the reflector, essentially putting the subject in front of or behind that useless focus point. That's better, but still each small area on the subject sees light coming from only one direction, not ideal for avoiding spurious effects such as those shown
HERE.
So, maybe it's better to mix things up a little. There are two ways to do that. One way is to mix up the light rays that hit the reflector. Andrew's system does that, courtesy that diffuse reflector that the flashes shine on. The other way is to make the reflector at least a bit diffusive. If each little chunk of reflector scatters rays across the whole subject, then each little chunk of the subject sees light coming from the whole reflector. That's what the Olympus reflector does.
There's kind of a threshold effect. Once you have enough diffusion in the reflector (or the light source) so that each part of the subject sees light coming from the whole reflector, then adding more diffusion doesn't change the quality of the light, it just cuts the efficiency. If you have enough light to make up for the lost efficiency, there's nothing at all wrong with a paper cone.
--Rik
