The New and The Old, or R11 vs R10.5 GI test by ThePinkus
Open full image in new tab Members remain the original copyright holder in all their materials here at Renderosity. Use of any of their material inconsistent with the terms and conditions set forth is prohibited and is considered an infringement of the copyrights of the respective holders unless specially stated otherwise.
Description
Hi folks,
this is just a test that confronts the new AR3 GI engine in R11 against the previous one, at almost out-of-the-box settings.
I am using Stonemason's "The Ark" scene imported into C4D by InterPoserPro.
Lighting is just from materials via the luminance channel (the light blue lights have GI generation rised at 400%)
In both case GI's Diffuse Depth is set to just 2.
The R10.5 has accuracy at 70%, the R11 has settings lowered to low.
Considerations: yep!, R11 takes longer but is flawless, R10 at this settings is faster and just useless...
Bye!
PS I know where some of my € will end... ;^)
Comments (4)
evs69
Yeah R10 looks pretty bad, whereas R11 looks awesome!
Magick_Lady
i have to agree with evs69
Becco_UK
Assuming you have correctly captioned the images - the light is defining objects much better in the earlier C4D10 render engine. The C4D11 image looks a bit flat with detail being washed out - an accurate test would have been to have had comparible quality settings.
ThePinkus
Hi Becco_UK! Thanks for the comment! The reason I have chosen those settigs is twofold: first they are almost "out of the box", second (being such) they show the "easyness" of setting up effective GI in the two releases. Maybe the R11 render does look flat (I think the scene can be improved, e.g. by working on materials, but no one prevents us from adding some appropriate lighting to enhance the visual effect... were this not a test render), but I do not concord on the fact that the R10 version defines objects better, I think most of those "details" are in fact artefacts from the GI computation. I think they clearly shows near the lamps in the upper part of the picture (especially near the one in the top right-centre, note the circular halo, the shape of which do not appear justified by the shape of the emitter). Also look at the door in the centre of the picture-- what is it that should physically yield those details? You can actually recognize on it the same circular spots of light that we found near the laps. These kind of artefacts are typical of the solutions computed by the R10's AR. Though I don't know the internal computation, I guess this is due on the way lighting from GI was mapped on surfaces (the way it was interpolated, I think). My impression is that this problem is essentilly resolved with R11's AR. Keep in mind that this particular set-up, i.e. light just from the luminance channel, thus all deriving from GI and no "helping" lights, was known to be critical for the previous AR (reading some forums around the web, this is not just my experience), and the confront of the new engine with the previous one is what shows here. I do think Maxon has done a great job addressing some ot the critical issues of the previous release. As I said above, these are almost "out of the box" settings. Therefore here also shows the starting point on which one can work to better the result. The very significative difference, IMO, is that in the R11 it is a good point with improvement margins, while in the R10 it is a mess that one has to work on to reduce and hide those artefacts... which is not at all so obvious to achieve (both in finding the right settings and in render time/resources)! O course just my impression ;^) !!! Cheers, Stefano