Page 1 of 2
Restaurant in Rome
Posted: Thu Jun 09, 2005 4:56 pm
by daros
Hi, this is the first work i completly done in Maxwell. It's a 5 hours workflow and 3 hours rendering time unsing Remote Matador.
The model was made by my client. I only modeled the glass surfaces.
Materials and Lighting are decided by my client.
The luminous Bar surface must be in theory alabaster.
David Rossmann
www.arkadin.biz
www.stack-studios.com
www.animax.it
Posted: Thu Jun 09, 2005 5:03 pm
by RedRaven
very nice
how do you mean 'using Remote Matador'?
where is the restaurant - i visit often and make the most of the food when there so would really like to try it out for real. - i'm not due in rome for another 3 months so hopefully it will be up and running by then.
Posted: Thu Jun 09, 2005 5:36 pm
by daros
Hi RedRaven,
thanks for your comment.
Remote Matador is a 3D monster application that automatically renders on a 70 cpu cluster.
In a few days we have a more exhaustive web page
www.arkadin.biz , and surly, thanks to the performance improvements of Maxwell Beta, a price-drop.
David Rossmann
Posted: Thu Jun 09, 2005 5:38 pm
by ingo
Nice pictures, but a bit cold atmosphere. Has your client a glass manufacture ?

Posted: Thu Jun 09, 2005 5:39 pm
by rivoli
good stuff. just a couple of things: there's a lot of noise in those images (for a production work made with an alpha software your results are really nice, and we all know what interiors+dielectrics+caustics mean when rendering with maxwell, still noise is a bit too much) and the restaurant looks like a scaled down model to me (but this is problably due to lack of detail modelling the objects, your client fault that is).
Posted: Thu Jun 09, 2005 5:45 pm
by daros
Yes Rivoli, it is still noise here. But the first image it's really difficult to obtain without noise. With the maxwell beta may it works fine.

Posted: Thu Jun 09, 2005 6:02 pm
by RedRaven
70 cpu's thats certainly a monster
Posted: Thu Jun 09, 2005 6:46 pm
by rivoli
i'd like to see one of those 70 images (if you used all the cpus) before maxwell merged them together to get the final one.
Posted: Thu Jun 09, 2005 6:58 pm
by DELETED
DELETED
Posted: Thu Jun 09, 2005 6:59 pm
by daros
Yes buffos,
The split rendring system allows us to distribute the rendering across how many computers we want but has some problems with Maxwell.
Maxwell at moment need's to compute the whole frame to simulate correctly caustics and emitters. We are changing one of our renderfarms to dual dual-core opterons. So we need to merge less MXI together. We made test with the MXI system on 30 dual Xeon nodes. The renderings was done in 3 minutes but the merging of the 30 mxi has needed about 20 minutes on a pure 1 GB netwok.
This is a temporary solution inside Remote Matador. We are sure that our heroes of Next Limit team will find a solution to resolve the MXI issue.
Anyway most of renderings present in this forum could be made without any problem in RemoteMatador.
David Rossmann
Posted: Thu Jun 09, 2005 7:02 pm
by daros
Thanks to to Tom, Mihai, and Rivoli and the others helping me with the sandblasted glass issue. As you see, finally i put some emitters behind the glass walls to obtain a brigter effect.

Posted: Thu Jun 09, 2005 7:26 pm
by rivoli
daros wrote:
As you see, finally i put some emitters behind the glass walls to obtain a brigter effect.
well, it looks it worked out nicely.
8etty wrote:
...and the pasta and pizza is missing..
you know, it really takes a good sss to render out a pizza.
Posted: Thu Jun 09, 2005 8:28 pm
by daros
Hi tonfarben,
no secrets. the original images are rendered in 2000x1000 pixels. So, the original noise is not so different as on the images you see.
May the samplig levels.. All images i think are in samplig level 18.
This it's my empiric opinion.
It the split system (with region render otion of maxwell) we are using if we put samplig level 8, for example, the rendering quality looks like 16... and takes the time of 18.
A bit strange but it works fine.
Posted: Thu Jun 09, 2005 8:39 pm
by Mihai
Thanks for posting the images

Some very difficult situations, with most of the light being caustic light coming from the lamps.
Are you sure you used the correct size for these because in the first one some of the lighting looks really weird:
Shadows seem very faded.
All the images could have used more bounces I think.
Posted: Thu Jun 09, 2005 10:03 pm
by daros
Hi Mihai,
may the shadows are so faded because the big emitters al arount the tables. The withe surface behind the glass is an emitter.