Hi gummidge,
The simplest answer (and physically completely wrong!) would be, render the model one time in the way you have it now, and one time with the “refraction” set to “1”. With that you should be able to set both images in, e.g., Ae as layers and in difference (Blend mode top layer). This will give you the “light pass” for the dar areas. Blur this and apply it on top of the original rendering, perhaps with a 10% or 20% transparency only. This will fake the resulut described below. (I consider a similar set up as my little example.
https://www.amazon.com/clouddrive/share/PrBp7FZ47Z3OqBL8yX-Ghe9kaU7zElQG_aJHqG73t78
==============
To dive more into it…
Do you have a reference image of what you like to achieve?
To have at least one “letter” as scene file would certainly help to search for a solution.
The drop in illumination around the edges seems natural to me. If you take a look how much light, compared to a flat area, is going to be in that inner edge, and how much more material is there… (A typical case in Construction physics, when heat is hitting a wall, vs an edge, hence why the black stuff shows up there in old building (longer story, but similar). In short normally the inside is equal in area to the outside (surface). In the edges, the tiny small area has to illuminate twice the area of the material thickness. All of that if the acrylic glass is clear and only dull on the outside.
If the dull side would be inside as well, and not eliminated by “glue”, then the diffusion effect of that dull inner surface would allow for a better distribution.
In practical light (e.g., on set) the diffusion is considered the light source, not the element that provides the light in the first place. So, if the inside would be dull and with that had diffusion qualities the edges would receive more light. Besides that, the acrylic refraction would help here as well. There is more, but minor effects, typical stuff, hence the request for a reference image, which I consider a standard to work with.
If you look at Acrylic glass, my impression is more that you look at these small side frames and get more of the inside than to look through, and the inside means here—all the way to the other small side, hence the darker areas.
All of that based on “real world” lighting. Hence why I asked about a reference image. Your ideas or targets might be crystal clear to you, but I might have a different “visual” about. So for the sake of communication, this is mandatory.
In C4D light is not refracted by default, nor diffused by default either. Even soft light - irregardless how it was produced, will not act inside of a material based on the diffusion. The Subsurface Scattering Shader might help here, but only partially.
In other words, to just build things as in reality and expect C4D just to simulate it—will not work easily—if at all. Light is simple and complex at the same time. Caustic behavior, the spreading of light in the inside and diffusion (which are both volumetric) are not really given. Even with a lot of set up it normally takes massive render power.
Normally there are way more options to fake it with a quick render, than options to get a physically correct result (light bounces indefinitely, so any “physical correct” rendering is kind of questionable anyway).
I hope that gave some more ideas why I prefer to think more about what I want to have as result, than how it would be in reality. For me 3D rendering is more a tool to get my images as I like to have them, I do not see 3D as a reality engine with an button to press to make things real. Reality and the simulation of it needs to be understood in detail, and replicated as such. Like I say since long here in this forum, just to take an HDRI and expect that anything else will follow automatically—will not work. The material is as critical as the light source and the understanding of it. Camera, Light and surfaces/materials create a triangle, and that can work in many way.
All the best
Sassi
Image left: normal/refracted, image left: not refracted and difference
