I am trying to model the haziness (transmitted scattering behavior) of two material types (hazy glass and a polymer) and from what I understand, the Trans property in Radiance is most suitable. From my experimental measurements, I know sample 2 is more hazy compared to sample 1. After doing the coversions to calculate A1, …A7, the values do not seem to make sense .as seen in Table 2 below, especially A3 values.
Could someone please advise on the data and if this is the right approach.
The results may not look correct, but they seem about right based on my checks. I know it seems odd that your blue channel is > 1.0, but your starting total transmission+reflection are close to 100% (94.7% and 88.8%, respectively). The way RGB colors work, it’s the Y channel that matters, which is (0.265*R + 0.670*G + 0.065*B), and sometimes R or B needs to be > 1.0 to make this work for a near-unity Y value. It’s rare that this causes problems during a calculation, since the Y channel is generally the one that governs ray propogation even with Russian-roulette going.
Thank you very much for the reply and explanations.
So based on the A1,…A7 values, is it obvious to see or say that sample 2 is more hazy than sample 1? i.e, will the simulation reflect that sample 2 is more hazy (higher scattered transmittance) than sample 1? I am asking because as mentioned before, from lab measurements, we know sample 2 is hazier than sample 1.
It depends what you mean by “haziness.” The behavior of the trans primitive is to split the transmission into “diffuse” and “specular” components. The diffuse part is purely Lambertian, meaning the same in all directions, and the specular part is either pure if the roughness is 0, or has some associated scattering. Your roughness values are very close to 0, meaning there won’t be much scattering for whatever specular there is.
Based on your values, the first sample is smoother, so there will be less scattering in the specular transmission than sample two. Sample 1 also has a much higher ratio between the specular and diffuse transmission than sample 2, so sample 2 should look “hazier” in two ways, the amount of scattering and the ratio of diffuse to specular.
I hope this helps. It’s often a good idea to set up a test scene to see if you can reproduce in simulation what the material looks like in a photo.
Following up from my questions before, is it fair to compare the haziness (light-scattering) properties of two materials as Trans if one of them is glass and the other is a transluscent plastic?
I am asking because i have tested my simulation with the two samples in table 2 (using them as Trans), and the result is not what I expected. Basically the idea was to see/show that the distribution is more uniform with sample 2 compared to sample 1 since sample 2 is hazier.
The main differences between the glass and trans material types in Radiance is that the first produces physically accurate transmission and reflection, where the second offers more flexibility with respect to the diffuse and specular parts. The glass type has no diffuse component, and the specular part is pure (no scattering).
To answer your question, you can approximate actual glass, especially if it is etched or sand-blasted, using the trans type. However, it is difficult to determine with certainty the correct parameters, which explains your struggle. The “haziness” of a surface has a lot to do with the degree of scatter in transmission, which is governed by the roughness parameter. You probably need to play with this parameter more to get a better match to the physical sample, and you may have difficulty getting very close to it as it is only a crude approximation to real-world scattering.
The best approach of course would be to measure the BSDF of your material and model it with the same-named material type in Radiance. However, that requires special equipment and a lot of additional techniques that you probably don’t want to consider at this point.