Page 8 of 50

Re: Nvidia Fermi Octane Competition

Posted: Sat Mar 20, 2010 9:05 pm
by enricocerica
ribrahomedesign:Really nice results !

Pozek:Nice scene, curious to see how you'll manage the square columns to create the fx :?

Re: Nvidia Fermi Octane Competition

Posted: Sat Mar 20, 2010 10:09 pm
by pedrojafet
Pozek, Escher rules... - and u did it too... Congrats! Try to light up a bit more...
Cheers!

Re: Nvidia Fermi Octane Competition

Posted: Sat Mar 20, 2010 11:03 pm
by ribrahomedesign
Hy there
i have to place a additional entry ,i found a old design of a swiss challet i did about one year ago ,octane is
a fantastic tool ,i am looking forward to the beta 2. thanks Radiance for this software .

Rico

Re: Nvidia Fermi Octane Competition

Posted: Sun Mar 21, 2010 1:12 pm
by ROUBAL
Hello,

I like action scenes, and so I have decided to build a new scene with my "Red Goose", but as I get a black viewport each time I trespass around 60% of my 512 MB Video RAM (I suppose that the remaining RAM amount is used to manage my dual screen display), I have had to rebuild almost everything from scratch with low poly models. I have had to remove the very detailed Wright Cyclone 9 motor and to build a low poly version with textures instead. The model is less detailed than I would have wished, but as I'd want to build a scene on the ground with more objects in the scene, I have to keep a memory margin for the additional geometry and textures.

Currently there is no motion blur in Octane and I have had some difficulties to fake the rotation of the propeller.

In my previous submarine scene, the propeller was in background and out of focus, but for an airplane it is an other story ! So, I have modelled a kind of "fresnel lens" with a low IOR, and I have kept only fiew parts of this waved disk to simulate the sun reflection on the rotating propeller. This specular object is placed just in front of the propeller.

fireflies removed with denoiser filter.

Pathtracing kernel used for every renders. I would have prefered direct lighting in some cases, but I couldn't get a correct transparency on Plexiglass materials.

448 607 Triangles. The amount of samples per pixel is different for each image. As I had few time, I stopped the render when the image looked enough good. More infos on the proof screens.

Unfortunately, for the images Airplane_08.jpg and Airplane_12.jpg, the proof screens including rendering time and s/px data have been lost (pasted images deleted by accident in photoshop after moving in Octane scene). This said, all the images are made from the same scene with different IBL images used for background and lighting and the memory usage was around 50% of the 512 MB.

Re: Nvidia Fermi Octane Competition

Posted: Sun Mar 21, 2010 1:32 pm
by enricocerica
Really cool images Roubal ;)
One with some dof would be great.

Re: Nvidia Fermi Octane Competition

Posted: Sun Mar 21, 2010 1:41 pm
by ROUBAL
Thank you Enrico ! I will try do make one more, but I have already used Dof on all of them, with focus on the pilot or between the pilote and the motor. I always have the fear of exagerating this effect, because on images taken from distance, everything is generally in focus. There is an exception for the side view, taken with a wide angle lens (Wing tip blurred close to the camera).
In fact, in most of the image the most blurred thing should be the sky, due to DOF and moving as well.

Re: Nvidia Fermi Octane Competition

Posted: Sun Mar 21, 2010 8:26 pm
by ribrahomedesign
very , very nice Roubal.
this images are fantastic , imagine how it looks with
motion blur. good work

Rico

Re: Nvidia Fermi Octane Competition

Posted: Mon Mar 22, 2010 12:03 am
by ROUBAL
Thank you Rico ! I'm currently working on a low poly UV textured hangar. The one I modelled some years ago was one of my first works and I used mostly procedural textures and really too many polygons for my video RAM. And yes, motion blur would be welcome !o)

Re: Nvidia Fermi Octane Competition

Posted: Mon Mar 22, 2010 6:02 am
by pozek
enricocerica wrote:ribrahomedesign:Really nice results !

Pozek:Nice scene, curious to see how you'll manage the square columns to create the fx :?
Thank you Enricocerica for your words,
the problem is to align corectly the camera based on initial camera setup in sketchup, the most important thing is to keep correct FOV from the modelling software. It's a trick that simulate this "paradox" perspective. There are some models on warehouse that reffer to this problem but only this treat this model on perspective, the others made model are based on "axonometry" or two point perspective. I try to render in octane but seems that Octane not handle correctly the model when FOV is set to 1 (simulating "axonometry").
The explanation of this model is here :) : http://www.youtube.com/watch?v=OHxiYM62 ... re=related
Keep up good job.

Re: Nvidia Fermi Octane Competition

Posted: Mon Mar 22, 2010 6:05 am
by pozek
pedrojafet wrote:Pozek, Escher rules... - and u did it too... Congrats! Try to light up a bit more...
Cheers!
Thank you very much for your words Pedro
Indeed Escher rules :)
If i have time I will try to light up the scene