I'm implementing a path-tracer to simulate the earths atmosphere using Rayleigh and Mie scattering. Is the following correct? To properly render the rainbow I need to use Mie scatterig theory applied to water drops of some diameter. The reason we see a rainbow is that the phase function for raindrops has two backwards pointing peaks at approx 42 deg and 50 deg (or something like that) AND that these peaks changes a little for different wavelengths. So if I got this right I can't use for example the Henyey Greenstein phase function because it does not have these peaks and is not wavelength dependant. Correct? So, I will need a more accurate phase function to properly simulate the rainbow and possibly other atmospheric effects such as halos. I found a program called MiePlot that can be used to calculate alot of stuff on Mie scattering. Can this be used to create a lookup table for a more accurate phase function for different particle sizes? If so, can someone give me some directions on how to use MiePlot for this? Can I create some kind of batch file to calculate the phase function for alot of different wavelenghts and particle sizes and export this to a file? Thanks in advance, Patrik

I don't have much background to give any guidance to your question, however I believe I remember seeing a rainbow demo as part of the nvidia sdk. Parameters addressing key phrases in your question.