CoolColJ's test pics thread

Kram1032 · Post by **Kram1032** » Wed Sep 19, 2007 4:41 am

that Euro/Dollar thing is an interesting idea

yeah!
That will be great, indeed! xD

Damn, between my last updates, there where 5 years and it was even hard to get my latest update

xD

I guess, with my next update, quantum computers slowly get payable, already xD

CoolColJ · Post by **CoolColJ** » Wed Sep 19, 2007 9:06 am

that's the thing guys, if I wait till the end of the year when Penryn comes out, the Q6600 will be even cheaper and so will the current video cards along with newer and better video cards.

So I can either choose between cheaper older stuff or better newer stuff

CoolColJ · Post by **CoolColJ** » Wed Sep 19, 2007 9:29 am

I decided to put back the environment in the SSS liquids scene - I originally took it out as the I thought it would render faster, especially caustics as per my underwater pool test with a ground plane a week ago.

I wanted to cut down some of the skylight which is washing out the image, much better contrast now, and better reflections on the glass.
Reduced the sky turpidty as well. Resolving better now.

But now my milk settings render really blue and more watered down

It looks like it needed all that extra skylight to scatter more...
So I used Whaat's full conentration liquid values and now it renders better, but much slower in Bidirectional MLT which seems to render it fine

230+ microseconds per sample is slow....

I tried MLT and it renders much faster, but the Milk don't look right, it goes grey... and the overall scene looks different, the stuff on the back wall I have never scene. The Sun position is slightly different due to the shift bug
MLT is faster because it doesn't calculate a lot of stuff Bidirectional MLT does...?

CoolColJ · Post by **CoolColJ** » Wed Sep 19, 2007 2:49 pm

interesting - Whaat's milk values also goes blue in colour, just like mine. I suspected it maybe due to the sun. So I rendered a cloudy day version with a uniform background lighting and sure enough the milk renders creamy white again.

So I can go back to my watered down version to speed up the render!

Now does real milk go blue under direct sunlight?

another thing, does milk thin out much when inside in a narrow container?
There doesn't seem to be much thinning out of the milk at the top on the glass object in these renders

Aribeth · Post by **Aribeth** » Wed Sep 19, 2007 7:11 pm

perhaps it reflects the blue of the sky?

CoolColJ · Post by **CoolColJ** » Thu Sep 20, 2007 10:03 am

Aribeth wrote:perhaps it reflects the blue of the sky?

maybe - I have another test scene I thought up that will remove all these factors and see what's really going on

---

scaled the scene down a touch, camera is set at head height as per Sketchup

lit with a uniform background colour of 240, 240, 240

CoolColJ · Post by **CoolColJ** » Thu Sep 20, 2007 11:22 am

Wonder why the glass bottom for the milk is black and the others clear? Is it because it's absorbing and scattering so much light? or a bug or a feature? or the angle

Kram1032 · Post by **Kram1032** » Fri Sep 21, 2007 1:24 am

it's because the milk shadows it, while the other two mats let the light more or less straight through, I'd say

looks great!
I never saw such nice looking uniform light^^

CoolColJ · Post by **CoolColJ** » Mon Sep 24, 2007 9:27 am

Ahh something funny and tragic....

remeber that pool scene I did where I show caustics under the water, and some showed in the sky?

Well it turns out I made a mistake when I resumed that render, I accidently used the wrong scene file and rendered an underwater version on top of that render...that's why those under surface water caustics showed up so quickly and why there are caustics in the sky....

And why I haven't been able to replicate the same result with the same scene...

So no bug Ono!

Ahhh but that kills any glimmer of hope Indigo can render water caustics when veiwed from above the water surface!

I have tried many times in V0.9 test 7, test 6 and v0.8 and not a single sign of water caustics, with all different kind of parameters. Even with MNCR of 500,000.

I'm hoping Ono can fix this, maybe look into some things Ian has done with radium recently

psor · Post by **psor** » Mon Sep 24, 2007 9:59 am

CoolColJ wrote:
...

Well it turns out I made a mistake when I resumed that render, I accidently used the wrong scene file and rendered an underwater version on top of that render...that's why those under surface water caustics showed up so quickly and why there are caustics in the sky.... Razz
And why I haven't been able to replicate the same result with the same scene...

CoolColJ wrote:
...

Ahhh but that kills any glimmer of hope Indigo can render water caustics when veiwed from above the water surface!
I have tried many times in V0.9 test 7, test 6 and v0.8 and not a single sign of water caustics, with all different kind of parameters. Even with MNCR of 500,000.

CoolColJ wrote:
...

I'm hoping Ono can fix this, maybe look into some things Ian has done with radium recently

*SCNR*

take care
psor

Kram1032 · Post by **Kram1032** » Tue Sep 25, 2007 2:31 am

lol!
but besides the sky caustics, it looked quite matching !? how can that be? xD

CoolColJ · Post by **CoolColJ** » Tue Sep 25, 2007 2:15 pm

yeah I think it helps that all I did with the camera was move it a few metres forward and down

CoolColJ · Post by **CoolColJ** » Tue Sep 25, 2007 10:44 pm

Well who's taking bets wether these white spots turn into caustics?

my settings for v0.9 test 6.

Bidirectional MLT
large_mutation_prob 0.2
max_change 0.005
max_num_consec_rejections 500000
max_depth 10000

I found with max change of 0.005 the white bits get more dense/smoother quicker than 0.01, but the image starts off more patchy/dirty looking.
These white areas started showing within the first minute. Currently at 500+ samples, so it's going to take a while to fill out
I haven't tried to see if test 7 can do it, but I doubt it...

They still show with large_mutation_prob of 0.5, large_mutation_prob of 0.5 and max_num_consec_rejections of 50,000 as well, but I just wanted to be on the safe side

maybe they are just surface Sun specular reflections?

Kram1032 · Post by **Kram1032** » Wed Sep 26, 2007 7:24 am

the sun idea sounds possible, but I actually don't think that...
some of the ripples simply have the wrong shape, for that...
(although, it then could be rereflected stuff....)
what about making that render twice (or even more often) simultaneously?

- once with max change 0, once with 1, and the others, sine-like (linear is fine, too, I guess^^) in between

blending that should lead to an overally balancedish smooth-sharp look

CoolColJ · Post by **CoolColJ** » Wed Sep 26, 2007 10:25 am

unfortunately after a long time and 2500+ samples, I see no caustics

They are just surface sun specular reflections, and that's not even a fraction of them. The water does have 66,000 polys so it make take a while to fully fill out. The settings I used don't seem to be allowing the wave shape to come through, larger Max change is needed.

CoolColJ's test pics thread

Who is online