CoolColJ's test pics thread

[Vanessa07]

Big Fan

Yes for ripples, and I have glass in openings

[Anthony]

Those are probably from light reflecting off the wavy surface I think?

[Vanessa07]

Anthony

I know this, I don't understand why there are no caustics inside the pool, Light is stop by water???

[Anthony]

Maybe. Share your material settings, that would be helpful. Is Bidir on?

[CoolColJ]

yeah caustics from water surface. Directly seen caustics very easy to get Indigo.

Indirect caustics like under water refracted ones, very hard
Especially from the sun as I have shown with all my experiments

Bidirectional MLT can sorta do it in V0.9 test6 with high MNCR in the 50,000 range, but not v0.9 Test7 from my tests

there are not many unbiased renderers that can do it right now in a reasonable time frame. Unreleased version of Radium looks to be able to do it nicely.

Re: specularly-viewed caustics... easily the hardest light path to evaluate in an unbiased manner (made even worse when it's the sun). The ONLY technique for reliably seeing indirect caustics is photon mapping.

Bi-directional path tracing is great for caustics in particular but only directly-viewed ones. It doesn't help at all for when the caustics are viewed via a specular surface (because it can't join up the light and eye portions of the path) which explains why Maxwell/Fry/Indigo/Radium et al will struggle, regardless of which technique they use for path generation.

MLT does help to a certain extent, once it's even found an indirect caustic (which, again, for the sun, happens once in a blue moon ) which explains why Indigo and Radium should (in theory) outperform Maxwell and Fry in that respect (it's not a confirmed fact but I don't believe either one of those uses MLT, just a hunch).

As I said above, the reason why Radium v0.9 produced such well-resolved and fast indirect sun caustics was a bug that made the sun slightly too big. See this thread in particular (underwater sun caustics in a little over an hour).

http://trackvids.co.uk/radium/forum/viewtopic.php?t=60

Part of the reason it did so well was also down to the mutation strategy which, for indirect caustics, tried to mutate the path only where it needed to, unlike unit-cube that mutates it everywhere (increasing the chances of missing the sun at the end of it all).

[Vanessa07]

<medium>
<name>water</name>
<precedence>11</precedence>
<basic>
<ior>1.5</ior>
<cauchy_b_coeff>0</cauchy_b_coeff>
<absorption_coefficient_spectrum>
<rgb>
<rgb>0.124504 0.0427243 0</rgb>
<gamma>2.2</gamma>

</rgb>
</absorption_coefficient_spectrum>
</basic>
</medium>
<material>
<name>water</name>
<specular>
<transparent>true</transparent>
<internal_medium_name>Eau</internal_medium_name>
</specular>
</material>

<medium>
<name>glass</name>
<precedence>10</precedence>
<basic>
<ior>1.33</ior>
<cauchy_b_coeff>0</cauchy_b_coeff>
<absorption_coefficient_spectrum>
<rgb>
<rgb>0.0269877 0.0269877 0.0269877</rgb>
<gamma>2.2</gamma>

</rgb>
</absorption_coefficient_spectrum>
</basic>
</medium>
<material>
<name>glass</name>
<specular>
<transparent>true</transparent>
<internal_medium_name>Verre</internal_medium_name>
</specular>
</material>

Bidir is on, I try with a meshlight

[Big Fan]

Perhaps if you remove the glass?
or maybe it works better if you make them exit portals and use bidirectional - but I am no expert on caustics
To me it looks like there are correct reflections on the ceiling and it seems there is some kind of lighting getting in the pool that is showing it dished? - there is a faint line of illumination say 300mm in from the edge..

[CoolColJ]

meshlight will work better, but it will need to be quite big and doesn't look quite right. Blackbody lights are extremely bright as well

[Vanessa07]

Big Fan

If I remove glass pb is the same, no change

[Big Fan]

perhaps sunlight is better as a point source rather than spread out meshlight and diffuse light bouncing around inside?
but I am guessing

[CoolColJ]

Vanessa07 - you might have to cheat on this one in Indigo as it won't render that kind of caustic very well unless your willing to render up to 20,000+ samples with a high MNCR value of 50,000 :

render one version with just a single layer water with no volume and hope it shows caustics and then overlay that onto the normal version in Photoshop etc

the smaller the light, the harder and longer it will take for indirectly seen caustics to show up

[Vanessa07]

CoolColJ

Thanks for reply, I continu my test a few time and if it doesn't work I will try with photoshop

Thanks everybody

[DaveC]

It appears that reflected caustics simply appear much more quickly. Which is a shame

[CoolColJ]

no problems with underwater caustics when your underwater though

[CoolColJ]

CoolColJ

Thanks for reply, I continu my test a few time and if it doesn't work I will try with photoshop

Thanks everybody

try an EXR background image for lighting, it might work better
but you have no control over the lighting direction