Hi,
Im needing to do a glass with hard caustics from a glass onto as table.
I have been looking at this tutorial in the docs but they state
"Note that 'hard' caustics from small but bright light sources (e.g., a spot light through a cognac glass) are not currently possible."
just wondering if this is stil the case or it this possible yet ?
Thanks
Chris
Looks like the tutorial link is empty. I'm assuming it might have been https://docs.arnoldrenderer.com/display/A5AFMUG/Refractive+Caustics+using+an+Emissive+Shader ?
That page says that zero-radius point lights and other lights with zero size will not do caustics, which is true. However, the good news is that since Arnold 5.3 most arnold lights can now be made visible by setting the light's camera and/or transmission attributes to 1, so it's now easier to get caustics. You can get essentially hard caustics by using a small, but non-zero area sized light ("hard" caustics in real-life also come from non-zero sized lights). Keep in mind that the smaller the light, the more samples you'll need to get a converged image. Adaptive sampling and the Arnold denoisers are going to help you keep render times from exploding.
Hi, any plans to improve this (hard caustics) in future ?
Or maybe making Arnold bi-directional path tracer ?
The tutorial here now as an updated scene file.
There is also this fake caustic effect that will give you quicker results.
Is there any chance to implement bidirectional mode in Arnold for caustic?
For hard caustic you need set up lower value of roughness, lower than 0.1
in render settings - set specular indirect blur to 0, and set big value for indirect samples clump
50-100 if you not use dispersion and 500+ if use dispersion.
However you need understand arnold nor bidirection path tracer, and caustics for it is the Achilles heel.
Remember, if you increase value for indirect samples clump more than 10 and specular indirect blur to 0, you render well hardly and very noisy and require very big values for diffuse semples.
I attached some examples