@Leon 

If you are asking about what OpticStudio does in those extreme cases, this is something you can test directly in the software.

For example, this is how I would test 3.

Go in Non-Sequential mode, create a Source Ray with a single ray (makes it more tractable) and a power of 1W. Add a Rectangular Volume along the path of the ray, adjust the front face of the Rectangular Volume so it has the corresponding reflectance and scatter fraction, and sandwich the whole thing between Detector Rectangles. Make the Detector Rectangles small so that scattered rays are not likely to be detected.

The goal here is that one detector is used to measure specular reflection only, one detector is used to measure direct transmission only, and the rest is assumed to be scattered (no absorption). Of course, there’s a small chance that a scattered ray is detected by a Detector Rectangle, but its rather unlikely and you can run it a couple of times to check.

If you run a raytrace and look at the power detected by each detector, you’ll find:

Both detector detect 0.25W. Meaning, there’s 0.25W of specular reflection, 0.25W of direct transmission, and since the Source has a power of 1W, 0.5W is scattered. In other words, the power is first split for scattering (0.5W is scattered, 0.5W is not scattered), and then for reflection (0.25W is reflected, 0.25W is transmitted).

I’m attaching my file for your reference. I hope this makes sense.

Take care,

David

Thank you David for the reply. 

It’s crystal clear to me now, that is:

The scattering takes the 1st piece of the incident optical power, the reflection/transmission share the remaining optical power.