Vray light / lightmaterial quality questions

[simmsimaging]

Hi - I have been playing with lighting scenes using both Vray lights with images mapped to them, as well as image planes (just single poly's) with images mapped to them using VrayLight material.

What I'm trying to figure out is what is the *best* way to do it - and by that I mean the highest image quality/most realistic light distribution? I am not as worried about fast - just "good" (My current project involves glass and liquids and I want the caustic effects and refractions to be as nice and real looking as I can get them.)

I'm assuming using a Vray light is better, and from something Vlado said I would assume using an .exr as the mapped image versus an 8bit (png etc.) would be the better - but is that the case? I'm generally using the Universal setting approach, and I am also using photon based caustics quite a bit as well.

On the same sort of line: if I am using IBL, and I'm using GI with the Universal settings approach, and want it to be "good" rather than fast is it best to using the Vray dome light and just leave the environment slots empty/off? (both in Max and Vray rollouts) I have tried it any number of ways but I can't really be sure which is the right way as the difference is often pretty subtle.

Any advice/help would be much appreciated.

Thanks,
b

[rivoli]

I'd say vray light over lightmat. when vray lights couldn't be mapped, I used to place a light (set as portal) just in front of planes with vray lightmats just to help vray doing the sampling. now this is just an extra step no longer needed (unless the geometric object in question can't be replaced with a flat square light), vray lights should do a faster, cleaner and more efficient job. with much more control too: affect only diffuse, or speculars, reflections, being visible or not and so on..
as you say, hdr images are better than ldr, both for illumination and reflections. should be easy to test though, I guess you cna easily find out what work best for you in any given situation.
being a dome light an environment in itself, I'd say that yes, you don't need the environment on (which it might be blocked by the light anyway, even it checked invisible). unless of course you don't need different reflections than the dome ones, or different background for refractions, or anything else specific to your scene.
as long as caustics go, you may have seen this already, but just in case:

http://www.chaosgroup.com/forums/vbu...ad.php?t=35579

and, well, hope at least something from above makes some sense but I'm sure loads of people on this board can provide actual insight.

[simmsimaging]

Thanks Rivoli - that does help a lot.

I know what you mean when you say it can easily be tested, but I really find I always seem to spend 10 hours blindly messing around for every productive hour - so sometimes it's good to just go to the experts for help

I was not really sure about dome lights and gi because it seemed to work fine for me, but I read in a few threads that image mapped Vray lights didn't produce very clean GI, or so I took them to mean anyway, so I wasn't sure if the best quality needed some other combination.

I have read through that caustic thread many times but I still have trouble fine-tuning them to get a nice result. Sometimes I get lucky, others it's just hours of playing around. My hope was that finding the "optimum" lighting solution would shorten up that time. Someone (I think it was Paul Oblamov?) also suggested just turning off photon mapped caustics entirely when using Brute Force GI because it could do the job on its own, but I wasn't sure if that produced the same look or not. Anyone know for sure?

Thanks again!
b

[rivoli]

Someone (I think it was Paul Oblamov?) also suggested just turning off photon mapped caustics entirely when using Brute Force GI because it could do the job on its own, but I wasn't sure if that produced the same look or not. Anyone know for sure?

it can be done for sure, in the universal settings thread there are a few examples from vlado that look really good, but I've never been able to get anything as good in reasonable times. might be just me though.
on the other hand I find the direct caustics trick gives quite good results most of times. even though it's not easy to get a totally clean solution either, and you need to shoot a BUNCH of photons.
here's a little something I did a while back, doesn't look bad but still far from being clean. and already took its time as it is

[simmsimaging]

it can be done for sure, in the universal settings thread there are a few examples from vlado that look really good, but I've never been able to get anything as good in reasonable times. might be just me though.
on the other hand I find the direct caustics trick gives quite good results most of times. even though it's not easy to get a totally clean solution either, and you need to shoot a BUNCH of photons.
here's a little something I did a while back, doesn't look bad but still far from being clean. and already took its time as it is

That looks really good Rivoli. I'll have to try that method out again. I was testing it for a job but I was having a lot of problems with the reflection/refraction being different and just gave up on it. When you say a "bunch" of photons what do you mean; like 10K-15K or more?

I forgot about the refraction tests in Vlado's thread - that would have answered the GI caustics question pretty quickly

Thanks
b

[rivoli]

if I remember correctly I stopped at about 15k caustics subd, it couldn't be really less than that. and it still needed more, which was doable of course. but I don't remember if the size for the caustic map was getting out of hand or the time to calculate it, or both

[simmsimaging]

if I remember correctly I stopped at about 15k caustics subd, it couldn't be really less than that. and it still needed more, which was doable of course. but I don't remember if the size for the caustic map was getting out of hand or the time to calculate it, or both

Thanks. I will play with it again.

b

[rivoli]

hope you'll manage to get something useful out of it, and hope we'll get the chance to see something from your new work. if you're allowed to post it that is.

[simmsimaging]

We'll see how it goes - but I will post the image/images when I can. Mostly been doing 2D stuff with only small 3D elements, so nothing much to post lately.

b