Before the actual work begins we should perform some simple tests of the SPRAY setup. This is a very important step since there is a good chance that some of the numerous parameters may be wrong (although we did our best).
No scattering particles, perfect absorber at the bottom interface
Here the rays are partly reflected at the air-resin interface. Since the resin has a constant and real refractive index of 1.5 the reflectance of the top interface is about 4%. All rays transmitted to the bottom interface are absorbed if we use a perfect absorber for that interface. Here is a side view with some test rays:
The detector should have a constant signal of about 4%. Doing the simulation in the spectral range 300 ... 1000 nm with 20 spectral points and 10000 rays/spectral point one gets the following detector signal:
This is the expected behaviour. The noise level can be estimated the following way: If 10000 rays are processed and 4% reach the detector, the average number of rays is 400. The statistical variation of the number of detected rays is the square root of 400, i.e. 20. The amplitude of the noise is then given by (420-380)/10000 = 0.4% which is roughly what you see in the graph above.
No scattering particles, perfect mirror at the bottom interface
In this case all rays are reflected at the bottom interface
and should reach the detector. We get this:
No scattering particles, gold mirror at the bottom interface
This case is used to test a case with spectral variation. The optical constants of gold are taken from the SPRAY database. The obtained spectrum (300 ... 1000 nm, 71 points, 1000 rays/point) and the corresponding color are shown below:
This looks like gold, and we can continue with more serious investigations.