Page 2 of 3
Re: Octane VS Corona
Posted: Thu Oct 02, 2014 10:10 pm
by Karba
RobSteady wrote:@ glimpse
Sure, you can tweak it a little bit more but a simple slider for highlight compression would be great.
Is this so complicated to implement?
added, wait for a next release.
Re: Octane VS Corona
Posted: Thu Oct 02, 2014 10:39 pm
by aoktar
I'm agree with juanjgon,
I'm a old vray,etc.. guy also. So many days has been spent for fixing flickering issues of GI calculations.
Biased renderers does minimal calculation, not every pixel as brute-force methods. Not too fair compare 1 pixel to 1000 pixels. Also you can't produce flicker-free animation in a short time with biased Renderers. Only static scenes is good for them.
Re: Octane VS Corona
Posted: Fri Oct 03, 2014 11:13 am
by RobSteady
Karba wrote:added, wait for a next release.
Great!
Thanks Karba.
Re: Octane VS Corona
Posted: Fri Oct 03, 2014 12:54 pm
by riggles
What would be cool would be tone mapping, like Rienhard or Filmic mapping, instead of just preset camera response curves and highlight compression (although I'd still happily take highlight compression).
Re: Octane VS Corona
Posted: Fri Oct 03, 2014 5:36 pm
by dsyee
Highlight compression - isn't that what the new "GI clamp" in Octane does? Or does "highlight compression" mean something different?
Re: Octane VS Corona
Posted: Fri Oct 03, 2014 7:55 pm
by RobSteady
dsyee wrote:Or does "highlight compression" mean something different?
Yes.
From Corona Forum: "Highlight compression is similar to Reinhard tone mapping in Vray, with only difference that it goes in opposite direction. Increasing HL compression value will reduce burn areas on your image, while in Vray, reducing burn value in reinhard does the same thing."
Without Highlight Compression:
With Highlight Compression:
Re: Octane VS Corona
Posted: Tue Oct 07, 2014 7:22 pm
by MrFurious
juanjgon wrote:The question is ... what happens if you animate something inside the Corona scene, characters with deformations, or soft body simulations, for example? ... and if you animate the sun light, or light sources inside the room?
Of course, a biased solution can be rendered a lot faster than an unbiased one, but you need to fight with the biased render settings to get images without artifacts, and animations without artifacts are sometimes even impossible (or only possible with much higher render times). This is the magic of unbiased renderers like Octane or Arnold, you don't need to spend hours with render settings, and the final animations are always good, without artifacts.
********Anyway a biased kernel for Octane could be a game changer for interior visualization and the archviz market.*******
-Juanjo
+10!
Re: Octane VS Corona
Posted: Tue Oct 07, 2014 9:24 pm
by glimpse
Now I get curious, why people are using these unbiased engines, so limited in features (compared to CPU counterparts) when all they want is biased solution, why not just use other engine then?
Isn't all the beauty of Octane Render that it's UNBIASED GPU render engine? =)
if You want to paint out an elephant in Your Zoo just to have a giraffe..wouldn't it be better to buy a giraffe in a first place?
I might sound weird but why on earth You would like to run Octane in biased mode? Run any CPU in biased mode & it will do circles around Octane if You take interiors..biased mode on GPU will give the same problems existing on CPU..& why would anyone want these problems? all those sliders, complexity, headache with animations..- the beauty of compute power available on GPU is that it could simplify the work..- & work hours cost more than any hardware..
sorry, I don't get it.. =)
Re: Octane VS Corona
Posted: Wed Oct 08, 2014 1:28 am
by gristle
Hmm, is it unbiased if I can change samples per light? That's bias right?
Re: Octane VS Corona
Posted: Wed Oct 08, 2014 2:46 am
by aoktar
gristle wrote:Hmm, is it unbiased if I can change samples per light? That's bias right?
No, it's a factor to help calculation will be done more on some sources.
Biased is similar to estimation of illumination with less points. It takes a few rays except all rays and try to fill uncalculated area with some interpolations.