messiah render

[chikega]

Thanks Mike for the invite. :) Here's the project file if anyone cares to take a stab at this. The sphere objects are separate for Metanurbing. The HDR probe image is also contained in the project file. All parameters are fairly close ... but I didn't include any reflections for the material. I did a quick test render, without using GI Noise reduction, at the specified resolution and it took 13min without reflections. Some of you that are more accomplished with ArchViz type stuff may be able tweak it a whole lot. Have fun! :)

http://www.3ddmd.com/images/Messiah/LightingChallenge.jpg (http://www.3ddmd.com/download/LightingChallenge1.mpj)

[dobermunk]

Thanks chikega! I tweaked your settings mildly and set to render while being marched through the landscape by my wife. It was a a great walk - 3 castles, a bunch of deer and one downpour! Anyway - the render's up at the link.

[Wegg]

http://www.eggington.net/%7Ewegg/RenderTests/MikeTest.jpg

Took almost two hours. I think the results from Messiah are quite good quality wise but its tough to really get it super fast and not introduce a lot of grain.

[PaulNewman]

Fantastic render Wegg! :thumbsup:

Pity about the polys on the ball.

[Wegg]

You can see that in MikeRB's render too. No SubD or anything. Couldn't go changing the models.

[PaulNewman]

You can see that in MikeRB's render too. No SubD or anything. Couldn't go changing the models.However, didn't chikega say The sphere objects are separate for Metanurbing.? Just wondering since the scene is very interesting and the messiah render rather nice. I'm sure a metaNURBS level or two on the one ball won't affect render times.

Anyway, nice going!

[PaulNewman]

I just did quick 1280x720 limited region renders on the big ball with the other geometry off...

0:50 = polys, no subD
1:05 = subD level 1
1:03 = subD level 2

1:04 = subD level 3
Seems like subD adds some overhead regardless of level (at least at low levels). Don't know what the difference would be on the whole scene, but on 2 hours it may not be much unless you're trying to shave off minutes and seconds to out-render someone else? :)

[Wegg]

I think its important not to mess with anything but surface properties and render settings in these tests. MikeRB's renders have those same artifacts so you just have to roll with it. As soon as you go changing stuff like that. . . it becomes a whole new test.

Here (http://www.eggington.net/%7Ewegg/RenderTests/MikesLighting.mpj) is my .mpj if you want to play.

[chikega]

That's a nice render, Wegg! I posted it over on Mike's thread if you don't mind too terribly. I think the shadows are more accurate even though it took longer to render. I wonder how much longer it would take Modo to reduce the splotchiness that you see in the shadows and corners of his render? Mike, would you have the modo scene to download?

As a side note, I think it's interesting how the different renderers interpret the HDR Sphere Probe ... messiah and blender both seem to cast a greenish/gray shadow on the left wall as where others are fairly even in illumination.

[Wegg]

I'm pretty sure Modo is doing it wrong. Its warping it. . . and flipping it.

Modo's shading noise reduction technique is far superior to Messiah's currently. They will always have the advantage when using that. But yes. . . Messiah can hold its own when accuracy and quality are priorities over blazing fast render times.

[chikega]

I think you may be right Wegg. I believe I saw something about a bug and modo's HDR probe handling on the Luxology forum. Also, what is the spec on the machine you used to render the image above. Mike has a pretty nice rig. It would be good to factor this into the equation. :)

EDIT: Yep, it's a modo HDR bug:

http://forums.luxology.com/discussion/topic.aspx?id=7580&show=hdr%20bug

[Wegg]

I have an X2 same as Mike. I don't think mine is clocked quite as high as his is but its pretty close.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.