Minimum subdvis for caustics Photon Mapped caustics?

V-Ray V-Ray for Maya General
1

Hi all,
I know that you need a higher min. subdivs for progressive caustics. Should I also boost it for the photon mapping method?

2

Also, is Photon Mapping method is dead at this point or Chaos is planning to support it in the future?

The reason I chose Photon Mapping to study caustics now is that I find Progressive method cpu usage awfuly low, even after boosting the ray bundle size to something like 392.

3

I totally understand you. did you try to lower ray bundle size rather than raising it ? I had some good results once lowering it to 64.
As for the caustics subdivs, I would say that it totally depends on your scene, if it’s one light producing caustics or several lights. I have one scene here, one light producing caustics on a large object (15m), with 16000 subdivs.

4

Interesting, I will give it a try, thanks!

When it comes to subdivs, I had some rather nice results with a light of similar size and 8000 subdivs, as you said, I felt I should go for 16000.

Photon Map calculations are sometimes riddiculously slow, on the top of that they tend to freeze the entire Maya…
Last caustics map I saved weighted around 15GB, what if I would go for animation? Would it make a map for every single frame? :smile:

regards,
Karol

5

You have to cheat with caustics :slight_smile:
Think… what exactly should produce caustics… what should receive caustics. Exclude everything you can ! Off course it’s biased, but that’s what we do with Vray anyway.
I worked on a scene yesterday, and went from 8min caustics calculation time to 3min… only by excluding/including objects.

You water produces caustics, but doesn’t need to receive caustics… your wall receive caustics but doesn’t need to produce caustics

6

You are totally right,
In this case I light linked the caustics light to affect the water surface only but you are right there there so much more to play around with.

A hassle nonetheless, I am looking forward to some progress with progressive solution.
I am afraid the photon mapping solution is a bit ancient at this point. I remember we studied it at Uni from my 65yo prof. :smile: