Are there any TDs / more technical Lookdev guys here who could offer some tips on writing arnold shaders? I'm specifically working on a physically accurate (as close as i could get) atmosphere for a low orbit earth (rayleigh + mie scattering etc). I've been looking at O'neil's Accurate Atmospheric Scattering and Bruneton's Precomputed Atmospheric Scattering (which is a mouthfull) I've never written anything this mathematically complex before, anyone have any guidance or other reference? I could always do a rough approximation in color ramps but it would be rather crude and much less versatile.
>>571253What's so hard, download the shader from O'neil's Accurate Atmospheric Scattering and translate it to arnold.
>>571272because his implimentation doesn't use multiple scattering and he left out the phase function
>>571274i should note that its because most of these resources are built to run in real time, which is not a concern.also osl shaders aren't arnold specific
What kind of money are you offering for me to make it?
>>571277not asking for that
>>571253If you want to simply have a "dome-light texture" that follows the details of this paper, it is possible in OSL. If you want an actual volume shader that scatters light and casts volumetric shadows within it, you need to write a volume shader in C++, which is a herculean task.Here is an example shader written in C++ for the alSurface:https://github.com/ChaosGroup/vray_al_surface/blob/master/std_plugin/vray_brdfalsurface.cpp
>>571589OSL can do volumes though
>>571589i def wouldn't describe 350 lines of c++ as a herculean task, but i see your point. even though im not sure where the limitation in osl you're describing is, its not that big of a deal since everything in osl is part of a library with a c++ api anyway