Vray caustics problem with objects of small size

[stevesideas]

Hi there,

I have a scene where a number of water droplet models are sitting on a plane. Each droplet is small in size..probably up to 7mm in diameter. I have a direct light setup and vray caustics turned on. Would you say it would be very hard to get an accurate caustic solution ?

I have tried all different types of search params etc but can only get blurry solutions..

[simmsimaging]

I run into that problem a lot too. Try really small search distances.

b

[simmsimaging]

more useful (sorry )
Last time I used it I was working with search distances around .2 cm - assuming real-world scale of the water drops.

b

[sv]

Also you could try the direct visualisation of the caustics approach that vlado showed in the thread called "just for fun" or something similar

[stevesideas]

Yes, im looking into that thread now..although I dont fully understand what Im supposed to be doing. Should I be using photomapping and lightcache plus that script to see the caustics being calculated ?

[simmsimaging]

I don't think you need photon mapping GI if that is what you mean. I pretty much only use BF/LC and it works fine.

You may need to crank up the caustic intensity on your lights to see things happen. Use the global one in the render dialog, or for individual lights through their v-ray properties. Go crazy high (like 20-30,000) and see if that makes a difference. If you are using max lights needing intensities like that is not unheard of.


b

[stevesideas]

Ok cheers. I'll give this a go.

[simmsimaging]

Not sure if you have seen this, but it is a helpful tutorial:

http://software.asgvis.com/index.php...d=28&Itemid=55

I'm still hoping for simpler realistic caustics in Vray 2.0

b

[stevesideas]

I'll check out that tutorial. The only way I can see the caustics being calculated is if I whack up the exposure and turn on srgb...then I can see them. Otherwise, its too dark to see anything

[stevesideas]

That tutorial has helped out a bit...Now im using vray lights, I dont have to have the exposure right up now..

Thanks for the help..now I just need to get the nice looking caustics

[stevesideas]

Ok. So now I have a scene setup. This is the result. I still cannot seem to get nice clean caustics.

My caustic settings are:

Multiplier: 1 (my light intensity mulitplier is 100)
Search distance: 20
Max photons: 50
Max density: 2

I am firing 5000 photons from the light.

Im not reeally sure how to get clean/smooth results ?

[simmsimaging]

Here's a quote from Vlado I keep on hand to remind me when setting up caustics:

The answer to that would depend on how you are lighting the scene. But in general, I myself would use just the normal caustics (not the directly visible ones), max. photons set to 0, and max. density locked to 1/3rd to 1/4th of the search distance. Then I would play with the Search distance to reach the amount of blurriness that can be tolerated, and then its just a matter of shooting more photons until they become relatively noise-free. Also, I would try to use relatively small light sources or point light sources (large area lights might take a very long time to clear up the caustics in this way).

Maybe this will help? If not I'm sure Vlado or a vray guru can give you some advice.

b

[stevesideas]

So normal caustics being caustics that dont use the method and script as outlined in that tutorial. I thought the script just allowed the user to view the caustic build so tweaks were easier...does it adjust the method that caustics are actually calculated ?

[simmsimaging]

I think Vlado explains it in the original thread, but my understanding is that it will not show indirectly visible caustics - so caustics seen through glass do not show up. Something like that. I usually just toggle in on long enough to get a feel for how the caustics are shaping up and then turn it off to be sure I'm getting the best results.

The main take-away from that quote was the approach to the settings, I think.



b

[stevesideas]

Cheers. I appreciate the help. I'll try that method. Already the caustics are looking better...just need to be a bit tighter now