[eric nixon]

Well its taken a few years to get this right!, mostly because we were expecting an update to thin-sss (re. performance at extreme assymetries, general crashiness, and the emitter bug).

But it is possible with the current thin-sss to get a fairly convincing balloon, its just very sensitive, and a classic maxwell conundrum, in the sense that it renders slow, until you get the material close to being realistic, at which point it renders kind of normal speed for sss.



No additives used, 1 layer, 2 thin-sss bsdfs, one rough , one smooth. both with assym at -.65 (ideally you would want to use -.99 on one of the layers to get sharp highlights on the inside surface but that renders very slow, with the current thin-sss code, hint, hint)

I would recommend waiting for a more practical version of thin-sss before experimenting with balloons it was hard work to get this working (maxwell frequently refused to render, with instant softcrashes!), but i wanted to show this image to confirm visually that the current code does in fact work!

And if anyone urgently needs balloons for a commmercial project then they will be available in any colour.

[NicoR44]

Gorgeous looking baloons!!

[eric nixon]

Thanks, balloons are cool , AFAIK only maxwell can render this, heres the red one with some white text, on the simball, sl21 takes 1hr on dual xeon 6core 3.8..

[gadzooks]

Stunning work Eric! Is it the glossy reflections or the SSS itself that takes a while to render?

[eric nixon]

I think its just the light bouncing around inside the balloon, esp when there are many balloons bunched up. It is often the case that the things we find beautiful involve complex light interaction. (and are therefore slow to render).

I will try an image with the balloons lifting something (not sure what yet) and render that with motion blur to make it really slow

[JDHill]

[eric nixon]

I thought it might be interesting to show the progression renders which were saved along the way. Also these versions were crashing my machine constantly -- thin-sss being the culprit, 99% certain of that.

wip 1


wip2

[maotto]

The best I've seen so far, really

[sampson]

yup, ...very nice!

[bograt]

Did you see this? http://www.maxwellrender.com/index.php/ ... eos/4/6/14

[astanelle]

well done.

i personally miss a little transparency.
and i wonder if the little "edge" on top of each ballon is a need or just a "model- issue"

for example. http://www.ronenbekerman.com/vray-ballo ... nd-benoit/

[eric nixon]

Well thanks everyone.

Bograt, those are real balloons.. the cgi ones are the many many small ones across the sky. (blame the old NL marketing guy for decieving you)

Astanelle, theres certainly room for improvement (several limiting factors at the moment esp re. thin-sss behaviour) and there are many types of balloon, and also a deflating balloon changes the look.
In the picture you uploaded its the lighting which gives them that appearance more than the material.

EDIT; I had a look at that link, I like the appearance of the v-ray material but it doesnt give the right shadows / light distribution needed for a bunch of ballons to look real.

And finally yes, there is a 'model issue' with the tops, i should fix it, actually i already forgot to fix it, so thanks for the reminder.

[eric nixon]

I was just browsing the arnold render support site out of curiousity, and saw these rendered in-house, maybe its unfair because they can prob do better, but i suspect no better than the v-ray ones.



new WIP!;

[bograt]

I like the new WIP! looks spot on. I figured the balloons in the really close shots were real but wondered if he had done any making of or posted anything in this forum.
What parameters did you change for the new balloon material?

[photomg1]

Wow on that new wip Eric . Love the material as well : )

(where can i get it : ) )